diff --git "a/data/validation-00000.jsonl" "b/data/validation-00000.jsonl"
new file mode 100644--- /dev/null
+++ "b/data/validation-00000.jsonl"
@@ -0,0 +1,1000 @@
+{"id": "demo_0000", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Tool-calling traces and CI integration in Go.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "01ecda677e27b8918134d21a0d1cd85ef04617f9"}
+{"id": "demo_0001", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Python.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "ce6f758b1bc428101ae12ad43bff7a4e88c0cfc7"}
+{"id": "demo_0002", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for MoE expert routing for code in Go.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "724a661b7365a7ceedaa2c6eb1ec2248b6dc7be0"}
+{"id": "demo_0003", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Governance, auditability, and safety gates in Python.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "e026f2a318ea6750793d9529645190cb6103242d"}
+{"id": "demo_0004", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in TypeScript.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "aa5e688888ba819372064089899cd28221f9fa41"}
+{"id": "demo_0005", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for MoE expert routing for code in Go.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "55f016e0f5b3b2fae7261a93255883cdb7e2e829"}
+{"id": "demo_0006", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for MoE expert routing for code in Go.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "55f016e0f5b3b2fae7261a93255883cdb7e2e829"}
+{"id": "demo_0007", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Diff-first patching and SWE-bench style evaluation in TypeScript.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "b08c20a651510f12ed82346dc8a929711f3fdae9"}
+{"id": "demo_0008", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Go.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "f4a8b6d07ae6c10d7d031fd3e71902189cfbece4"}
+{"id": "demo_0009", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Tool-calling traces and CI integration in Python.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "3901f67d39afc0bfe5397c2c27f8ead5b2b472e7"}
+{"id": "demo_0010", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Self-grading and automated critique in TypeScript.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "f31708cf2dc2bf12377b51649cb192e41c358dac"}
+{"id": "demo_0011", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Tool-calling traces and CI integration in Rust.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "bfa0ece85ebe9acb845d6bef2bf266c580695c9b"}
+{"id": "demo_0012", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Tool-calling traces and CI integration in Go.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "da7d498e0aa830fb9d9079a323908d7d3e164681"}
+{"id": "demo_0013", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for MoE expert routing for code in TypeScript.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "5d8081fa4e2b4d2c429db2955e8f9dd90080d145"}
+{"id": "demo_0014", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Diff-first patching and SWE-bench style evaluation in Rust.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "0de742ee682add3fabe10b41b63c699058d167da"}
+{"id": "demo_0015", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for MoE expert routing for code in Python.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "a7ccacbb4f991696e10c4df1b3a721ca21aaccfa"}
+{"id": "demo_0016", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Self-grading and automated critique in Go.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "6ddfb66b8ab85ff28b1c9d587350b2c6c891a905"}
+{"id": "demo_0017", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Self-grading and automated critique in Go.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "85d5a5bc6f30d51ffc33cee542d2c5f2d1d15657"}
+{"id": "demo_0018", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Self-grading and automated critique in Rust.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "d2815ca3d0e438822e466acc43474ba2874f5aad"}
+{"id": "demo_0019", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for MoE expert routing for code in TypeScript.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "a4322f9593d84f6099aa35b3061f2aa8a7adab8e"}
+{"id": "demo_0020", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Diff-first patching and SWE-bench style evaluation in Python.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "ed6687b23f76fa0808d028f46af13a368a5e5155"}
+{"id": "demo_0021", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Diff-first patching and SWE-bench style evaluation in Rust.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "882626e1c64172100c8201be60a636d13fc4cd1d"}
+{"id": "demo_0022", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Rust.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "1a0fe8867460b6a6a80121de51265743a509981b"}
+{"id": "demo_0023", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for MoE expert routing for code in Python.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "a7ccacbb4f991696e10c4df1b3a721ca21aaccfa"}
+{"id": "demo_0024", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for MoE expert routing for code in TypeScript.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "84c33edd4bbd46f9b7ed6f724fa0c0c0891f9673"}
+{"id": "demo_0025", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in TypeScript.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "97e3218662e2f00746256643faef6885ea18cd7b"}
+{"id": "demo_0026", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Governance, auditability, and safety gates in Python.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "2cd0a884e47168413d32f61dc500ce673c8e8ce6"}
+{"id": "demo_0027", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Governance, auditability, and safety gates in Python.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "2cd0a884e47168413d32f61dc500ce673c8e8ce6"}
+{"id": "demo_0028", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Diff-first patching and SWE-bench style evaluation in Python.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "52094c608f86848d71b36f8b28490eeab830b40f"}
+{"id": "demo_0029", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Governance, auditability, and safety gates in Rust.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "29b4334a491410a737f41c7f169c9fcc2e01bf01"}
+{"id": "demo_0030", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for MoE expert routing for code in Go.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "c77684ed94a4148f8449f9b073cd70d3bc577a1d"}
+{"id": "demo_0031", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Diff-first patching and SWE-bench style evaluation in TypeScript.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "da67f5c566686221ddbda9c967c5c084e482c0de"}
+{"id": "demo_0032", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for MoE expert routing for code in Go.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "724a661b7365a7ceedaa2c6eb1ec2248b6dc7be0"}
+{"id": "demo_0033", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Self-grading and automated critique in TypeScript.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "8c70093964bcfcfa32b3ab7b473fcf5233227529"}
+{"id": "demo_0034", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Diff-first patching and SWE-bench style evaluation in Rust.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "0de742ee682add3fabe10b41b63c699058d167da"}
+{"id": "demo_0035", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Diff-first patching and SWE-bench style evaluation in Go.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "8e9b6a48836548d0f9202dc0384ed0d98e123f41"}
+{"id": "demo_0036", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for MoE expert routing for code in Rust.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "a757555ae2d72f4f9a6c937f6ef6303650b71604"}
+{"id": "demo_0037", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Diff-first patching and SWE-bench style evaluation in Go.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "8e9b6a48836548d0f9202dc0384ed0d98e123f41"}
+{"id": "demo_0038", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Python.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "ce6f758b1bc428101ae12ad43bff7a4e88c0cfc7"}
+{"id": "demo_0039", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in TypeScript.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "fab772571976b248bba6e42050c283d2dcacdc46"}
+{"id": "demo_0040", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Go.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "f4a8b6d07ae6c10d7d031fd3e71902189cfbece4"}
+{"id": "demo_0041", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Self-grading and automated critique in Python.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "54a082fe1d81e29f0620aed5d29e8a89b05a7741"}
+{"id": "demo_0042", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Self-grading and automated critique in Python.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "cf868e52e7b626f6e1445e520642918b3a2a1910"}
+{"id": "demo_0043", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Governance, auditability, and safety gates in TypeScript.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "740d9e6139f5f9bbc02e9d9236e6d6023f4948ff"}
+{"id": "demo_0044", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Diff-first patching and SWE-bench style evaluation in TypeScript.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "7105a1d4b1d96645cc6cc8bb9a9c50b8b512bf53"}
+{"id": "demo_0045", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in TypeScript.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "3f20b9c5d6121cdd67b0ae805ccb9f1ee72869a7"}
+{"id": "demo_0046", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Governance, auditability, and safety gates in TypeScript.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "740d9e6139f5f9bbc02e9d9236e6d6023f4948ff"}
+{"id": "demo_0047", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Governance, auditability, and safety gates in Rust.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "407f5e4eeddb9a55f35db2f65cf2cc7ffcb2996e"}
+{"id": "demo_0048", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Tool-calling traces and CI integration in Python.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "9d80a224ffbc2b265069c9355557fa2452ab1b46"}
+{"id": "demo_0049", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Rust.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "8cd98daa7ed94eceb808aace257ad77a00c093c1"}
+{"id": "demo_0050", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for MoE expert routing for code in TypeScript.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "04cf27cd6c7cc60a6ad34383c06c00220df1d7d6"}
+{"id": "demo_0051", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for MoE expert routing for code in Rust.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "a757555ae2d72f4f9a6c937f6ef6303650b71604"}
+{"id": "demo_0052", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Diff-first patching and SWE-bench style evaluation in Python.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "ed6687b23f76fa0808d028f46af13a368a5e5155"}
+{"id": "demo_0053", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Self-grading and automated critique in Rust.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "204fb7aa6df38d2d8077eb752afceef4537dabc4"}
+{"id": "demo_0054", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Tool-calling traces and CI integration in Rust.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "6b80883d024afe8123d755ffa41e5b3cc0dcfb74"}
+{"id": "demo_0055", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Tool-calling traces and CI integration in Rust.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "2ad46a119bbbd13e9f59ca46d56fc6ceffd8c0ad"}
+{"id": "demo_0056", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Governance, auditability, and safety gates in Go.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "6818e4c49db4ba63a4ff27bae5ee7d4d3383255f"}
+{"id": "demo_0057", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Tool-calling traces and CI integration in Python.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "b12d3c546b09fdd7d4abbaa5a72e2e41deabef1d"}
+{"id": "demo_0058", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Tool-calling traces and CI integration in Go.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "01ecda677e27b8918134d21a0d1cd85ef04617f9"}
+{"id": "demo_0059", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Diff-first patching and SWE-bench style evaluation in Go.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "9ac511806d0fd80846c9ee66e870bc422000ca5d"}
+{"id": "demo_0060", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for Self-grading and automated critique in Rust.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "0fcd85479910d25d7597ad10e547762ff04885d2"}
+{"id": "demo_0061", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Diff-first patching and SWE-bench style evaluation in Rust.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "0de742ee682add3fabe10b41b63c699058d167da"}
+{"id": "demo_0062", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in TypeScript.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "aa5e688888ba819372064089899cd28221f9fa41"}
+{"id": "demo_0063", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Go.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "d165292165be3a4c543c7636d89fefd10b078e7c"}
+{"id": "demo_0064", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for Tool-calling traces and CI integration in Go.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "b1f3c1daa457ede0e06e5e190d15d6e50635238b"}
+{"id": "demo_0065", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for MoE expert routing for code in Go.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "c77684ed94a4148f8449f9b073cd70d3bc577a1d"}
+{"id": "demo_0066", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Governance, auditability, and safety gates in Python.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "2cd0a884e47168413d32f61dc500ce673c8e8ce6"}
+{"id": "demo_0067", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Governance, auditability, and safety gates in Go.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "b0ee387d89f5f635c1f9cc32f9773b4068014798"}
+{"id": "demo_0068", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Go.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "d165292165be3a4c543c7636d89fefd10b078e7c"}
+{"id": "demo_0069", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Self-grading and automated critique in TypeScript.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "5ecb56d92db1e83dd4aeacdea1ea7ed21c892010"}
+{"id": "demo_0070", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Self-grading and automated critique in TypeScript.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "d02ca6f3f580c87df094e25ecb3de78fc576555b"}
+{"id": "demo_0071", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Self-grading and automated critique in TypeScript.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "d02ca6f3f580c87df094e25ecb3de78fc576555b"}
+{"id": "demo_0072", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Python.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "ce6f758b1bc428101ae12ad43bff7a4e88c0cfc7"}
+{"id": "demo_0073", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Diff-first patching and SWE-bench style evaluation in Rust.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "8dacf6e08ac4a42cf38a0c147636cfdd66f96f97"}
+{"id": "demo_0074", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in TypeScript.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "3f20b9c5d6121cdd67b0ae805ccb9f1ee72869a7"}
+{"id": "demo_0075", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Tool-calling traces and CI integration in Python.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "b12d3c546b09fdd7d4abbaa5a72e2e41deabef1d"}
+{"id": "demo_0076", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for MoE expert routing for code in Go.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "55f016e0f5b3b2fae7261a93255883cdb7e2e829"}
+{"id": "demo_0077", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Rust.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "57884ed4c3d00bb495a4382eb0ea01a961e42e16"}
+{"id": "demo_0078", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Governance, auditability, and safety gates in TypeScript.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "fe883ef5866a8ab03c83ebb26b3251f8679aca88"}
+{"id": "demo_0079", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for MoE expert routing for code in Go.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "441e2206b28951463f35934572603bd6fa2398d7"}
+{"id": "demo_0080", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Tool-calling traces and CI integration in Rust.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "256c808b21b4248d2f7c4c8c03d35d40bff23d5c"}
+{"id": "demo_0081", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for MoE expert routing for code in Go.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "724a661b7365a7ceedaa2c6eb1ec2248b6dc7be0"}
+{"id": "demo_0082", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Diff-first patching and SWE-bench style evaluation in Rust.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "8dacf6e08ac4a42cf38a0c147636cfdd66f96f97"}
+{"id": "demo_0083", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Self-grading and automated critique in Go.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "be168e7e68cd992bda3c031199a780770e3d7833"}
+{"id": "demo_0084", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Python.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "79a6d964f0af5a0c080ec8cbfafd1114d574e578"}
+{"id": "demo_0085", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for MoE expert routing for code in TypeScript.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "84c33edd4bbd46f9b7ed6f724fa0c0c0891f9673"}
+{"id": "demo_0086", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Self-grading and automated critique in Go.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "c012733817f6f8ceeff47e5dc6c8bdcfb44433fe"}
+{"id": "demo_0087", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for MoE expert routing for code in Rust.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "a757555ae2d72f4f9a6c937f6ef6303650b71604"}
+{"id": "demo_0088", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Governance, auditability, and safety gates in Rust.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "d2ee9ea5723a120aeb73ff995a456cc5d8f41428"}
+{"id": "demo_0089", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Diff-first patching and SWE-bench style evaluation in Python.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "214c8aa1c81b49da6e98d286de14957000ecb489"}
+{"id": "demo_0090", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Self-grading and automated critique in Rust.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "0cb9a5e8d76f6bfd1e0d104cb5af19b52e4f9fae"}
+{"id": "demo_0091", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Tool-calling traces and CI integration in Python.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "9d80a224ffbc2b265069c9355557fa2452ab1b46"}
+{"id": "demo_0092", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Diff-first patching and SWE-bench style evaluation in Go.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "8e9b6a48836548d0f9202dc0384ed0d98e123f41"}
+{"id": "demo_0093", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for MoE expert routing for code in Python.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "d10ebbd0771e561742de9157381a089d64c68539"}
+{"id": "demo_0094", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for MoE expert routing for code in Go.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "441e2206b28951463f35934572603bd6fa2398d7"}
+{"id": "demo_0095", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Self-grading and automated critique in Python.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "3adfb9c8e53d1b7119af2ca07c587ce30d29dcb5"}
+{"id": "demo_0096", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Tool-calling traces and CI integration in Go.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "ed2736226ef8b9ff056b4302095ccecddcf966bd"}
+{"id": "demo_0097", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Self-grading and automated critique in Go.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "85d5a5bc6f30d51ffc33cee542d2c5f2d1d15657"}
+{"id": "demo_0098", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Diff-first patching and SWE-bench style evaluation in TypeScript.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "7b045c7e00592cb0936703c6efbe07d24fed1640"}
+{"id": "demo_0099", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for MoE expert routing for code in TypeScript.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "c6a3c7c940db60ab9c0aad92f2280b426d27d44d"}
+{"id": "demo_0100", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Diff-first patching and SWE-bench style evaluation in Python.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "399b590cd03414ba736ffb6888734b51cacb8ca9"}
+{"id": "demo_0101", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for Diff-first patching and SWE-bench style evaluation in Rust.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "b1045ddb05b3c862b41b79b8eba5e07fecfd99b9"}
+{"id": "demo_0102", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Tool-calling traces and CI integration in Rust.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "6771342ab90a99fa5a829bbba8c24379d3ed4b2d"}
+{"id": "demo_0103", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Governance, auditability, and safety gates in Go.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "b0ee387d89f5f635c1f9cc32f9773b4068014798"}
+{"id": "demo_0104", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for Self-grading and automated critique in Python.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "fed8c1215ad6b2320879502cbbc51dc2e7766e7f"}
+{"id": "demo_0105", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Governance, auditability, and safety gates in Python.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "2cd0a884e47168413d32f61dc500ce673c8e8ce6"}
+{"id": "demo_0106", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Self-grading and automated critique in Python.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "fed8c1215ad6b2320879502cbbc51dc2e7766e7f"}
+{"id": "demo_0107", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Governance, auditability, and safety gates in Go.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "6818e4c49db4ba63a4ff27bae5ee7d4d3383255f"}
+{"id": "demo_0108", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Governance, auditability, and safety gates in Python.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "e026f2a318ea6750793d9529645190cb6103242d"}
+{"id": "demo_0109", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Self-grading and automated critique in TypeScript.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "5ecb56d92db1e83dd4aeacdea1ea7ed21c892010"}
+{"id": "demo_0110", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for Self-grading and automated critique in Go.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "5effe6fde62e48aa6e9be1d6e57385f23da7e153"}
+{"id": "demo_0111", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Tool-calling traces and CI integration in TypeScript.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "b70cf5f1b56ffca136b21a28a0f03cb2993dfc03"}
+{"id": "demo_0112", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Diff-first patching and SWE-bench style evaluation in Rust.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "882626e1c64172100c8201be60a636d13fc4cd1d"}
+{"id": "demo_0113", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Self-grading and automated critique in Rust.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "204fb7aa6df38d2d8077eb752afceef4537dabc4"}
+{"id": "demo_0114", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Tool-calling traces and CI integration in Go.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "da7d498e0aa830fb9d9079a323908d7d3e164681"}
+{"id": "demo_0115", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Self-grading and automated critique in Go.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "be168e7e68cd992bda3c031199a780770e3d7833"}
+{"id": "demo_0116", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in TypeScript.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "3f20b9c5d6121cdd67b0ae805ccb9f1ee72869a7"}
+{"id": "demo_0117", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Governance, auditability, and safety gates in Go.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "b0ee387d89f5f635c1f9cc32f9773b4068014798"}
+{"id": "demo_0118", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Governance, auditability, and safety gates in Go.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "6818e4c49db4ba63a4ff27bae5ee7d4d3383255f"}
+{"id": "demo_0119", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Governance, auditability, and safety gates in TypeScript.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "00ebe4210e2b4c255de24fd7a88ee68a0837ec64"}
+{"id": "demo_0120", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Self-grading and automated critique in Python.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "dabe0c58e324c9c8c2bdbfe80d61536a1063518d"}
+{"id": "demo_0121", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Self-grading and automated critique in TypeScript.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "4fc6dade0f8d7671f7cbe49b182676215b21cf70"}
+{"id": "demo_0122", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Tool-calling traces and CI integration in Rust.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "6771342ab90a99fa5a829bbba8c24379d3ed4b2d"}
+{"id": "demo_0123", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Governance, auditability, and safety gates in TypeScript.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "fe883ef5866a8ab03c83ebb26b3251f8679aca88"}
+{"id": "demo_0124", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for Diff-first patching and SWE-bench style evaluation in Rust.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "b1045ddb05b3c862b41b79b8eba5e07fecfd99b9"}
+{"id": "demo_0125", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Tool-calling traces and CI integration in Go.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "01ecda677e27b8918134d21a0d1cd85ef04617f9"}
+{"id": "demo_0126", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Tool-calling traces and CI integration in Rust.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "bfa0ece85ebe9acb845d6bef2bf266c580695c9b"}
+{"id": "demo_0127", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Tool-calling traces and CI integration in Python.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "9d80a224ffbc2b265069c9355557fa2452ab1b46"}
+{"id": "demo_0128", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for MoE expert routing for code in Rust.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "2f0099d7d55722348ce0b94bc5863892091131f8"}
+{"id": "demo_0129", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for MoE expert routing for code in Rust.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "6a015cfda615e97b9a9735cff067ff27b811115e"}
+{"id": "demo_0130", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Tool-calling traces and CI integration in TypeScript.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "b70cf5f1b56ffca136b21a28a0f03cb2993dfc03"}
+{"id": "demo_0131", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Diff-first patching and SWE-bench style evaluation in Rust.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "6aef16a5ce69387a6f198bdb7d380162bb772ca2"}
+{"id": "demo_0132", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for MoE expert routing for code in Go.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "c77684ed94a4148f8449f9b073cd70d3bc577a1d"}
+{"id": "demo_0133", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Governance, auditability, and safety gates in Python.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "2cd0a884e47168413d32f61dc500ce673c8e8ce6"}
+{"id": "demo_0134", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Rust.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "57884ed4c3d00bb495a4382eb0ea01a961e42e16"}
+{"id": "demo_0135", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Rust.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "57884ed4c3d00bb495a4382eb0ea01a961e42e16"}
+{"id": "demo_0136", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Tool-calling traces and CI integration in Rust.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "6771342ab90a99fa5a829bbba8c24379d3ed4b2d"}
+{"id": "demo_0137", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for Diff-first patching and SWE-bench style evaluation in Python.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "399b590cd03414ba736ffb6888734b51cacb8ca9"}
+{"id": "demo_0138", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Go.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "f4a8b6d07ae6c10d7d031fd3e71902189cfbece4"}
+{"id": "demo_0139", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Tool-calling traces and CI integration in TypeScript.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "7a36c2ca2b689bcc99d6e92fcc5a9d64f4e84ece"}
+{"id": "demo_0140", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Governance, auditability, and safety gates in Go.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "f58a9b02ebf152bc1c7f48fa9d5cda2b687e07c8"}
+{"id": "demo_0141", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Self-grading and automated critique in Python.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "cf868e52e7b626f6e1445e520642918b3a2a1910"}
+{"id": "demo_0142", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Diff-first patching and SWE-bench style evaluation in TypeScript.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "1971ecc0a5017fee37a2751e068efa0ba9bdb6b7"}
+{"id": "demo_0143", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Tool-calling traces and CI integration in Python.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "b12d3c546b09fdd7d4abbaa5a72e2e41deabef1d"}
+{"id": "demo_0144", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Tool-calling traces and CI integration in Rust.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "256c808b21b4248d2f7c4c8c03d35d40bff23d5c"}
+{"id": "demo_0145", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Tool-calling traces and CI integration in Python.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "3901f67d39afc0bfe5397c2c27f8ead5b2b472e7"}
+{"id": "demo_0146", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Governance, auditability, and safety gates in Rust.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "d2ee9ea5723a120aeb73ff995a456cc5d8f41428"}
+{"id": "demo_0147", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for Diff-first patching and SWE-bench style evaluation in Python.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "399b590cd03414ba736ffb6888734b51cacb8ca9"}
+{"id": "demo_0148", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for MoE expert routing for code in Go.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "724a661b7365a7ceedaa2c6eb1ec2248b6dc7be0"}
+{"id": "demo_0149", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Diff-first patching and SWE-bench style evaluation in Go.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "9ac511806d0fd80846c9ee66e870bc422000ca5d"}
+{"id": "demo_0150", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Go.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "71257b72d4d5b4e500f00ddada2361456c6ab673"}
+{"id": "demo_0151", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Self-grading and automated critique in Python.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "3adfb9c8e53d1b7119af2ca07c587ce30d29dcb5"}
+{"id": "demo_0152", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Self-grading and automated critique in Python.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "3adfb9c8e53d1b7119af2ca07c587ce30d29dcb5"}
+{"id": "demo_0153", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Go.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "ff4427b1aeb8acedadc80320a21a8044fc27397e"}
+{"id": "demo_0154", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in TypeScript.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "fab772571976b248bba6e42050c283d2dcacdc46"}
+{"id": "demo_0155", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Diff-first patching and SWE-bench style evaluation in Python.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "214c8aa1c81b49da6e98d286de14957000ecb489"}
+{"id": "demo_0156", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Diff-first patching and SWE-bench style evaluation in Rust.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "6aef16a5ce69387a6f198bdb7d380162bb772ca2"}
+{"id": "demo_0157", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Governance, auditability, and safety gates in Rust.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "29b4334a491410a737f41c7f169c9fcc2e01bf01"}
+{"id": "demo_0158", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in TypeScript.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "aa5e688888ba819372064089899cd28221f9fa41"}
+{"id": "demo_0159", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for MoE expert routing for code in Python.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "fd55fdc9ecf12df2b0b5a36e7f1f0f93d303dbb3"}
+{"id": "demo_0160", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Self-grading and automated critique in Rust.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "0cb9a5e8d76f6bfd1e0d104cb5af19b52e4f9fae"}
+{"id": "demo_0161", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for MoE expert routing for code in Go.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "441e2206b28951463f35934572603bd6fa2398d7"}
+{"id": "demo_0162", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for MoE expert routing for code in Go.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "441e2206b28951463f35934572603bd6fa2398d7"}
+{"id": "demo_0163", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Tool-calling traces and CI integration in Rust.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "256c808b21b4248d2f7c4c8c03d35d40bff23d5c"}
+{"id": "demo_0164", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in TypeScript.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "fab772571976b248bba6e42050c283d2dcacdc46"}
+{"id": "demo_0165", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Self-grading and automated critique in Rust.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "d2815ca3d0e438822e466acc43474ba2874f5aad"}
+{"id": "demo_0166", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Governance, auditability, and safety gates in Go.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "b0ee387d89f5f635c1f9cc32f9773b4068014798"}
+{"id": "demo_0167", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for MoE expert routing for code in Go.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "c77684ed94a4148f8449f9b073cd70d3bc577a1d"}
+{"id": "demo_0168", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Go.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "532dd88780c2362d1ef34dea626ff108239e845b"}
+{"id": "demo_0169", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Self-grading and automated critique in Go.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "85d5a5bc6f30d51ffc33cee542d2c5f2d1d15657"}
+{"id": "demo_0170", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Diff-first patching and SWE-bench style evaluation in Go.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "9ac511806d0fd80846c9ee66e870bc422000ca5d"}
+{"id": "demo_0171", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Tool-calling traces and CI integration in TypeScript.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "f3b404c1695427e8871ad8a5906b0fc4c088966a"}
+{"id": "demo_0172", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Go.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "ff4427b1aeb8acedadc80320a21a8044fc27397e"}
+{"id": "demo_0173", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Governance, auditability, and safety gates in Go.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "b0ee387d89f5f635c1f9cc32f9773b4068014798"}
+{"id": "demo_0174", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Go.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "ff4427b1aeb8acedadc80320a21a8044fc27397e"}
+{"id": "demo_0175", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Self-grading and automated critique in Rust.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "a43c2fa4a3ea9a8fc5ec43f2a91fd9d185e88885"}
+{"id": "demo_0176", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Rust.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "370dec51cc86af4647aa0881c7e9b9f3a6bbcfce"}
+{"id": "demo_0177", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Diff-first patching and SWE-bench style evaluation in Go.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "1ede0fbd7add911639c09c0d992869248542401b"}
+{"id": "demo_0178", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Rust.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "1a0fe8867460b6a6a80121de51265743a509981b"}
+{"id": "demo_0179", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for MoE expert routing for code in TypeScript.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "84c33edd4bbd46f9b7ed6f724fa0c0c0891f9673"}
+{"id": "demo_0180", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Diff-first patching and SWE-bench style evaluation in Rust.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "b1045ddb05b3c862b41b79b8eba5e07fecfd99b9"}
+{"id": "demo_0181", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Tool-calling traces and CI integration in TypeScript.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "47ed1415b5f3bf9882a334b6c73ccdeb3ada9c95"}
+{"id": "demo_0182", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Rust.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "370dec51cc86af4647aa0881c7e9b9f3a6bbcfce"}
+{"id": "demo_0183", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Self-grading and automated critique in Go.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "5effe6fde62e48aa6e9be1d6e57385f23da7e153"}
+{"id": "demo_0184", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for MoE expert routing for code in Python.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "d431a0a2bc3c7de101a6f702ba1a56260d522625"}
+{"id": "demo_0185", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for MoE expert routing for code in Rust.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "2f0099d7d55722348ce0b94bc5863892091131f8"}
+{"id": "demo_0186", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Python.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "cabf427b92e83fe95318ff3a5c2d46bd31d67e6e"}
+{"id": "demo_0187", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Diff-first patching and SWE-bench style evaluation in TypeScript.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "1971ecc0a5017fee37a2751e068efa0ba9bdb6b7"}
+{"id": "demo_0188", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for MoE expert routing for code in TypeScript.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "84c33edd4bbd46f9b7ed6f724fa0c0c0891f9673"}
+{"id": "demo_0189", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for MoE expert routing for code in Go.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "441e2206b28951463f35934572603bd6fa2398d7"}
+{"id": "demo_0190", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for MoE expert routing for code in TypeScript.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "a4322f9593d84f6099aa35b3061f2aa8a7adab8e"}
+{"id": "demo_0191", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Tool-calling traces and CI integration in TypeScript.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "2fd96b170ecf4c762b8fa01b03e48f8dc581d55a"}
+{"id": "demo_0192", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in TypeScript.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "bb6f07fec726ce3ac5b3ef6b3d61c76abe83fd3d"}
+{"id": "demo_0193", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Python.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "79a6d964f0af5a0c080ec8cbfafd1114d574e578"}
+{"id": "demo_0194", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Governance, auditability, and safety gates in TypeScript.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "740d9e6139f5f9bbc02e9d9236e6d6023f4948ff"}
+{"id": "demo_0195", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Python.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "cabf427b92e83fe95318ff3a5c2d46bd31d67e6e"}
+{"id": "demo_0196", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for MoE expert routing for code in TypeScript.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "84c33edd4bbd46f9b7ed6f724fa0c0c0891f9673"}
+{"id": "demo_0197", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for Governance, auditability, and safety gates in TypeScript.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "e55625866d9d72afa0ca219100fd584e58fd1a20"}
+{"id": "demo_0198", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Diff-first patching and SWE-bench style evaluation in Rust.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "8dacf6e08ac4a42cf38a0c147636cfdd66f96f97"}
+{"id": "demo_0199", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Governance, auditability, and safety gates in Rust.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "d2ee9ea5723a120aeb73ff995a456cc5d8f41428"}
+{"id": "demo_0200", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Governance, auditability, and safety gates in Go.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "6818e4c49db4ba63a4ff27bae5ee7d4d3383255f"}
+{"id": "demo_0201", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Diff-first patching and SWE-bench style evaluation in Go.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "1ede0fbd7add911639c09c0d992869248542401b"}
+{"id": "demo_0202", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Self-grading and automated critique in Go.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "85d5a5bc6f30d51ffc33cee542d2c5f2d1d15657"}
+{"id": "demo_0203", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Governance, auditability, and safety gates in Python.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "2cd0a884e47168413d32f61dc500ce673c8e8ce6"}
+{"id": "demo_0204", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for MoE expert routing for code in Go.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "55f016e0f5b3b2fae7261a93255883cdb7e2e829"}
+{"id": "demo_0205", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Tool-calling traces and CI integration in Go.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "3b6175b68f17af9ab68a3e4275a82a74736d9b6d"}
+{"id": "demo_0206", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Go.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "ff4427b1aeb8acedadc80320a21a8044fc27397e"}
+{"id": "demo_0207", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for MoE expert routing for code in TypeScript.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "5d8081fa4e2b4d2c429db2955e8f9dd90080d145"}
+{"id": "demo_0208", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Tool-calling traces and CI integration in TypeScript.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "b70cf5f1b56ffca136b21a28a0f03cb2993dfc03"}
+{"id": "demo_0209", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Diff-first patching and SWE-bench style evaluation in TypeScript.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "b08c20a651510f12ed82346dc8a929711f3fdae9"}
+{"id": "demo_0210", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Governance, auditability, and safety gates in Go.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "03e3880ec5391fcf3ecef131ad5aa0e0b53ad63e"}
+{"id": "demo_0211", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for Self-grading and automated critique in Go.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "5effe6fde62e48aa6e9be1d6e57385f23da7e153"}
+{"id": "demo_0212", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Diff-first patching and SWE-bench style evaluation in TypeScript.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "7b045c7e00592cb0936703c6efbe07d24fed1640"}
+{"id": "demo_0213", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for Governance, auditability, and safety gates in Rust.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "57ef0242e37ea02626901d906d595b09bf7d9d85"}
+{"id": "demo_0214", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for Governance, auditability, and safety gates in Go.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "f58a9b02ebf152bc1c7f48fa9d5cda2b687e07c8"}
+{"id": "demo_0215", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for MoE expert routing for code in Go.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "441e2206b28951463f35934572603bd6fa2398d7"}
+{"id": "demo_0216", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Tool-calling traces and CI integration in Go.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "da7d498e0aa830fb9d9079a323908d7d3e164681"}
+{"id": "demo_0217", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Tool-calling traces and CI integration in Go.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "3b6175b68f17af9ab68a3e4275a82a74736d9b6d"}
+{"id": "demo_0218", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Tool-calling traces and CI integration in TypeScript.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "f3b404c1695427e8871ad8a5906b0fc4c088966a"}
+{"id": "demo_0219", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Governance, auditability, and safety gates in Go.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "03e3880ec5391fcf3ecef131ad5aa0e0b53ad63e"}
+{"id": "demo_0220", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Tool-calling traces and CI integration in Rust.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "2ad46a119bbbd13e9f59ca46d56fc6ceffd8c0ad"}
+{"id": "demo_0221", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Diff-first patching and SWE-bench style evaluation in Go.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "1ede0fbd7add911639c09c0d992869248542401b"}
+{"id": "demo_0222", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Diff-first patching and SWE-bench style evaluation in Go.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "0a0e75aa64f2d9953fa6bcddadb57aa49fa969c6"}
+{"id": "demo_0223", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for MoE expert routing for code in TypeScript.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "5d8081fa4e2b4d2c429db2955e8f9dd90080d145"}
+{"id": "demo_0224", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Governance, auditability, and safety gates in Go.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "b0ee387d89f5f635c1f9cc32f9773b4068014798"}
+{"id": "demo_0225", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Governance, auditability, and safety gates in Python.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "2cd0a884e47168413d32f61dc500ce673c8e8ce6"}
+{"id": "demo_0226", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for Self-grading and automated critique in Go.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "5effe6fde62e48aa6e9be1d6e57385f23da7e153"}
+{"id": "demo_0227", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for MoE expert routing for code in Go.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "724a661b7365a7ceedaa2c6eb1ec2248b6dc7be0"}
+{"id": "demo_0228", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for MoE expert routing for code in TypeScript.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "04cf27cd6c7cc60a6ad34383c06c00220df1d7d6"}
+{"id": "demo_0229", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Tool-calling traces and CI integration in Python.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "d867ab5b1b03795a650f259437a08f64961f7a87"}
+{"id": "demo_0230", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Tool-calling traces and CI integration in Rust.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "256c808b21b4248d2f7c4c8c03d35d40bff23d5c"}
+{"id": "demo_0231", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Governance, auditability, and safety gates in TypeScript.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "00ebe4210e2b4c255de24fd7a88ee68a0837ec64"}
+{"id": "demo_0232", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for MoE expert routing for code in Go.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "368f0d7a01cf671c54bc9d9be2e070f6ef61d3dc"}
+{"id": "demo_0233", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for Self-grading and automated critique in TypeScript.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "8c70093964bcfcfa32b3ab7b473fcf5233227529"}
+{"id": "demo_0234", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Go.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "ff4427b1aeb8acedadc80320a21a8044fc27397e"}
+{"id": "demo_0235", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in TypeScript.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "3f20b9c5d6121cdd67b0ae805ccb9f1ee72869a7"}
+{"id": "demo_0236", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Governance, auditability, and safety gates in TypeScript.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "4e706c8a193f8c3c233720689ef537626b4b4d05"}
+{"id": "demo_0237", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Diff-first patching and SWE-bench style evaluation in Go.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "1ede0fbd7add911639c09c0d992869248542401b"}
+{"id": "demo_0238", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Go.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "71257b72d4d5b4e500f00ddada2361456c6ab673"}
+{"id": "demo_0239", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Python.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "ce6f758b1bc428101ae12ad43bff7a4e88c0cfc7"}
+{"id": "demo_0240", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Self-grading and automated critique in Python.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "fed8c1215ad6b2320879502cbbc51dc2e7766e7f"}
+{"id": "demo_0241", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Diff-first patching and SWE-bench style evaluation in TypeScript.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "da67f5c566686221ddbda9c967c5c084e482c0de"}
+{"id": "demo_0242", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Go.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "ff4427b1aeb8acedadc80320a21a8044fc27397e"}
+{"id": "demo_0243", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Diff-first patching and SWE-bench style evaluation in TypeScript.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "7105a1d4b1d96645cc6cc8bb9a9c50b8b512bf53"}
+{"id": "demo_0244", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Tool-calling traces and CI integration in TypeScript.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "b70cf5f1b56ffca136b21a28a0f03cb2993dfc03"}
+{"id": "demo_0245", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Tool-calling traces and CI integration in TypeScript.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "7a36c2ca2b689bcc99d6e92fcc5a9d64f4e84ece"}
+{"id": "demo_0246", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for MoE expert routing for code in Go.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "368f0d7a01cf671c54bc9d9be2e070f6ef61d3dc"}
+{"id": "demo_0247", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Governance, auditability, and safety gates in Python.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "320b4391adf67a3dd800dad890aed898949b23da"}
+{"id": "demo_0248", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Governance, auditability, and safety gates in Rust.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "d2ee9ea5723a120aeb73ff995a456cc5d8f41428"}
+{"id": "demo_0249", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for Governance, auditability, and safety gates in Go.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "f58a9b02ebf152bc1c7f48fa9d5cda2b687e07c8"}
+{"id": "demo_0250", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Self-grading and automated critique in TypeScript.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "d02ca6f3f580c87df094e25ecb3de78fc576555b"}
+{"id": "demo_0251", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for MoE expert routing for code in Python.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "d431a0a2bc3c7de101a6f702ba1a56260d522625"}
+{"id": "demo_0252", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for Governance, auditability, and safety gates in TypeScript.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "e55625866d9d72afa0ca219100fd584e58fd1a20"}
+{"id": "demo_0253", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for Self-grading and automated critique in Rust.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "0fcd85479910d25d7597ad10e547762ff04885d2"}
+{"id": "demo_0254", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Python.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "79a6d964f0af5a0c080ec8cbfafd1114d574e578"}
+{"id": "demo_0255", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Diff-first patching and SWE-bench style evaluation in Rust.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "0de742ee682add3fabe10b41b63c699058d167da"}
+{"id": "demo_0256", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Tool-calling traces and CI integration in Python.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "aa6c313c61194f09ceaa17d370a93b648759db00"}
+{"id": "demo_0257", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Governance, auditability, and safety gates in Python.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "f2b051c203782bf1ac25353b2085f0ee12d6e31e"}
+{"id": "demo_0258", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Governance, auditability, and safety gates in Python.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "e026f2a318ea6750793d9529645190cb6103242d"}
+{"id": "demo_0259", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for MoE expert routing for code in Go.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "c77684ed94a4148f8449f9b073cd70d3bc577a1d"}
+{"id": "demo_0260", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for Governance, auditability, and safety gates in TypeScript.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "e55625866d9d72afa0ca219100fd584e58fd1a20"}
+{"id": "demo_0261", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for MoE expert routing for code in Rust.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "2f0099d7d55722348ce0b94bc5863892091131f8"}
+{"id": "demo_0262", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for MoE expert routing for code in Go.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "724a661b7365a7ceedaa2c6eb1ec2248b6dc7be0"}
+{"id": "demo_0263", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Self-grading and automated critique in Go.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "85d5a5bc6f30d51ffc33cee542d2c5f2d1d15657"}
+{"id": "demo_0264", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Tool-calling traces and CI integration in Go.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "b1f3c1daa457ede0e06e5e190d15d6e50635238b"}
+{"id": "demo_0265", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Governance, auditability, and safety gates in Python.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "320b4391adf67a3dd800dad890aed898949b23da"}
+{"id": "demo_0266", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Self-grading and automated critique in Python.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "54a082fe1d81e29f0620aed5d29e8a89b05a7741"}
+{"id": "demo_0267", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Diff-first patching and SWE-bench style evaluation in Go.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "1ede0fbd7add911639c09c0d992869248542401b"}
+{"id": "demo_0268", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Governance, auditability, and safety gates in Rust.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "407f5e4eeddb9a55f35db2f65cf2cc7ffcb2996e"}
+{"id": "demo_0269", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for Diff-first patching and SWE-bench style evaluation in Python.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "399b590cd03414ba736ffb6888734b51cacb8ca9"}
+{"id": "demo_0270", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Self-grading and automated critique in Rust.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "d2815ca3d0e438822e466acc43474ba2874f5aad"}
+{"id": "demo_0271", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Governance, auditability, and safety gates in Go.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "f58a9b02ebf152bc1c7f48fa9d5cda2b687e07c8"}
+{"id": "demo_0272", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Self-grading and automated critique in Python.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "fed8c1215ad6b2320879502cbbc51dc2e7766e7f"}
+{"id": "demo_0273", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Tool-calling traces and CI integration in TypeScript.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "b70cf5f1b56ffca136b21a28a0f03cb2993dfc03"}
+{"id": "demo_0274", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Tool-calling traces and CI integration in Python.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "aa6c313c61194f09ceaa17d370a93b648759db00"}
+{"id": "demo_0275", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for MoE expert routing for code in Python.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "d10ebbd0771e561742de9157381a089d64c68539"}
+{"id": "demo_0276", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Tool-calling traces and CI integration in Rust.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "6771342ab90a99fa5a829bbba8c24379d3ed4b2d"}
+{"id": "demo_0277", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Self-grading and automated critique in Go.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "85d5a5bc6f30d51ffc33cee542d2c5f2d1d15657"}
+{"id": "demo_0278", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for MoE expert routing for code in Python.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "a7ccacbb4f991696e10c4df1b3a721ca21aaccfa"}
+{"id": "demo_0279", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Governance, auditability, and safety gates in Go.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "f58a9b02ebf152bc1c7f48fa9d5cda2b687e07c8"}
+{"id": "demo_0280", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Go.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "532dd88780c2362d1ef34dea626ff108239e845b"}
+{"id": "demo_0281", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Governance, auditability, and safety gates in TypeScript.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "00ebe4210e2b4c255de24fd7a88ee68a0837ec64"}
+{"id": "demo_0282", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Diff-first patching and SWE-bench style evaluation in Python.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "214c8aa1c81b49da6e98d286de14957000ecb489"}
+{"id": "demo_0283", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Python.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "ce6f758b1bc428101ae12ad43bff7a4e88c0cfc7"}
+{"id": "demo_0284", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for MoE expert routing for code in TypeScript.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "84c33edd4bbd46f9b7ed6f724fa0c0c0891f9673"}
+{"id": "demo_0285", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for MoE expert routing for code in Python.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "cdab494c7ea40443854dc9e76f81d9692cba5fad"}
+{"id": "demo_0286", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Tool-calling traces and CI integration in Rust.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "bfa0ece85ebe9acb845d6bef2bf266c580695c9b"}
+{"id": "demo_0287", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for MoE expert routing for code in Python.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "d10ebbd0771e561742de9157381a089d64c68539"}
+{"id": "demo_0288", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in TypeScript.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "97e3218662e2f00746256643faef6885ea18cd7b"}
+{"id": "demo_0289", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Self-grading and automated critique in TypeScript.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "d02ca6f3f580c87df094e25ecb3de78fc576555b"}
+{"id": "demo_0290", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Diff-first patching and SWE-bench style evaluation in Go.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "9ac511806d0fd80846c9ee66e870bc422000ca5d"}
+{"id": "demo_0291", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Tool-calling traces and CI integration in Rust.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "6771342ab90a99fa5a829bbba8c24379d3ed4b2d"}
+{"id": "demo_0292", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Tool-calling traces and CI integration in Go.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "3b6175b68f17af9ab68a3e4275a82a74736d9b6d"}
+{"id": "demo_0293", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Python.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "1f8f524e0710190859cc6940196c24af4135327a"}
+{"id": "demo_0294", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for MoE expert routing for code in Go.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "724a661b7365a7ceedaa2c6eb1ec2248b6dc7be0"}
+{"id": "demo_0295", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Diff-first patching and SWE-bench style evaluation in Rust.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "0de742ee682add3fabe10b41b63c699058d167da"}
+{"id": "demo_0296", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Governance, auditability, and safety gates in Python.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "e0062d33ddb8b86d135c9ab3dec2288f862109b1"}
+{"id": "demo_0297", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for MoE expert routing for code in TypeScript.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "a4322f9593d84f6099aa35b3061f2aa8a7adab8e"}
+{"id": "demo_0298", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Self-grading and automated critique in Python.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "3adfb9c8e53d1b7119af2ca07c587ce30d29dcb5"}
+{"id": "demo_0299", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for MoE expert routing for code in Rust.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "6a015cfda615e97b9a9735cff067ff27b811115e"}
+{"id": "demo_0300", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in TypeScript.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "aa5e688888ba819372064089899cd28221f9fa41"}
+{"id": "demo_0301", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Diff-first patching and SWE-bench style evaluation in Rust.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "8dacf6e08ac4a42cf38a0c147636cfdd66f96f97"}
+{"id": "demo_0302", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Tool-calling traces and CI integration in Rust.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "bfa0ece85ebe9acb845d6bef2bf266c580695c9b"}
+{"id": "demo_0303", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Self-grading and automated critique in Rust.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "204fb7aa6df38d2d8077eb752afceef4537dabc4"}
+{"id": "demo_0304", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Tool-calling traces and CI integration in Rust.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "bfa0ece85ebe9acb845d6bef2bf266c580695c9b"}
+{"id": "demo_0305", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Self-grading and automated critique in TypeScript.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "5ecb56d92db1e83dd4aeacdea1ea7ed21c892010"}
+{"id": "demo_0306", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Self-grading and automated critique in Go.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "85d5a5bc6f30d51ffc33cee542d2c5f2d1d15657"}
+{"id": "demo_0307", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Tool-calling traces and CI integration in TypeScript.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "f3b404c1695427e8871ad8a5906b0fc4c088966a"}
+{"id": "demo_0308", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for MoE expert routing for code in Python.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "cdab494c7ea40443854dc9e76f81d9692cba5fad"}
+{"id": "demo_0309", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Tool-calling traces and CI integration in TypeScript.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "f3b404c1695427e8871ad8a5906b0fc4c088966a"}
+{"id": "demo_0310", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in TypeScript.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "97e3218662e2f00746256643faef6885ea18cd7b"}
+{"id": "demo_0311", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Diff-first patching and SWE-bench style evaluation in Rust.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "0de742ee682add3fabe10b41b63c699058d167da"}
+{"id": "demo_0312", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Governance, auditability, and safety gates in Rust.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "407f5e4eeddb9a55f35db2f65cf2cc7ffcb2996e"}
+{"id": "demo_0313", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for MoE expert routing for code in Go.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "368f0d7a01cf671c54bc9d9be2e070f6ef61d3dc"}
+{"id": "demo_0314", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Self-grading and automated critique in TypeScript.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "d02ca6f3f580c87df094e25ecb3de78fc576555b"}
+{"id": "demo_0315", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Diff-first patching and SWE-bench style evaluation in Python.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "399b590cd03414ba736ffb6888734b51cacb8ca9"}
+{"id": "demo_0316", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Rust.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "1a0fe8867460b6a6a80121de51265743a509981b"}
+{"id": "demo_0317", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Diff-first patching and SWE-bench style evaluation in Rust.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "882626e1c64172100c8201be60a636d13fc4cd1d"}
+{"id": "demo_0318", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Governance, auditability, and safety gates in Go.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "03e3880ec5391fcf3ecef131ad5aa0e0b53ad63e"}
+{"id": "demo_0319", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Tool-calling traces and CI integration in TypeScript.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "b70cf5f1b56ffca136b21a28a0f03cb2993dfc03"}
+{"id": "demo_0320", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for MoE expert routing for code in Rust.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "2f0099d7d55722348ce0b94bc5863892091131f8"}
+{"id": "demo_0321", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Go.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "ff4427b1aeb8acedadc80320a21a8044fc27397e"}
+{"id": "demo_0322", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Self-grading and automated critique in Rust.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "0fcd85479910d25d7597ad10e547762ff04885d2"}
+{"id": "demo_0323", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Governance, auditability, and safety gates in Rust.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "29b4334a491410a737f41c7f169c9fcc2e01bf01"}
+{"id": "demo_0324", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Tool-calling traces and CI integration in Rust.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "6771342ab90a99fa5a829bbba8c24379d3ed4b2d"}
+{"id": "demo_0325", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for MoE expert routing for code in Rust.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "e7dcb2a63f3bf8de88060e0b2a7a5c6f548c4deb"}
+{"id": "demo_0326", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for MoE expert routing for code in Go.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "724a661b7365a7ceedaa2c6eb1ec2248b6dc7be0"}
+{"id": "demo_0327", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Python.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "ce6f758b1bc428101ae12ad43bff7a4e88c0cfc7"}
+{"id": "demo_0328", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Diff-first patching and SWE-bench style evaluation in Python.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "214c8aa1c81b49da6e98d286de14957000ecb489"}
+{"id": "demo_0329", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for MoE expert routing for code in Go.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "368f0d7a01cf671c54bc9d9be2e070f6ef61d3dc"}
+{"id": "demo_0330", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Python.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "1f8f524e0710190859cc6940196c24af4135327a"}
+{"id": "demo_0331", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Rust.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "0380156739222f10ebd520393e1597c16111cff1"}
+{"id": "demo_0332", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Governance, auditability, and safety gates in Go.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "b0ee387d89f5f635c1f9cc32f9773b4068014798"}
+{"id": "demo_0333", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Go.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "532dd88780c2362d1ef34dea626ff108239e845b"}
+{"id": "demo_0334", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Governance, auditability, and safety gates in TypeScript.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "00ebe4210e2b4c255de24fd7a88ee68a0837ec64"}
+{"id": "demo_0335", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Rust.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "370dec51cc86af4647aa0881c7e9b9f3a6bbcfce"}
+{"id": "demo_0336", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Self-grading and automated critique in Python.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "54a082fe1d81e29f0620aed5d29e8a89b05a7741"}
+{"id": "demo_0337", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Diff-first patching and SWE-bench style evaluation in Rust.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "6aef16a5ce69387a6f198bdb7d380162bb772ca2"}
+{"id": "demo_0338", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Governance, auditability, and safety gates in Rust.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "407f5e4eeddb9a55f35db2f65cf2cc7ffcb2996e"}
+{"id": "demo_0339", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Self-grading and automated critique in Rust.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "d2815ca3d0e438822e466acc43474ba2874f5aad"}
+{"id": "demo_0340", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Tool-calling traces and CI integration in Python.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "3901f67d39afc0bfe5397c2c27f8ead5b2b472e7"}
+{"id": "demo_0341", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Diff-first patching and SWE-bench style evaluation in Python.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "214c8aa1c81b49da6e98d286de14957000ecb489"}
+{"id": "demo_0342", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Tool-calling traces and CI integration in Rust.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "2ad46a119bbbd13e9f59ca46d56fc6ceffd8c0ad"}
+{"id": "demo_0343", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for MoE expert routing for code in Go.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "55f016e0f5b3b2fae7261a93255883cdb7e2e829"}
+{"id": "demo_0344", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Tool-calling traces and CI integration in Go.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "3b6175b68f17af9ab68a3e4275a82a74736d9b6d"}
+{"id": "demo_0345", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Go.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "532dd88780c2362d1ef34dea626ff108239e845b"}
+{"id": "demo_0346", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for MoE expert routing for code in Rust.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "a757555ae2d72f4f9a6c937f6ef6303650b71604"}
+{"id": "demo_0347", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Self-grading and automated critique in Rust.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "a43c2fa4a3ea9a8fc5ec43f2a91fd9d185e88885"}
+{"id": "demo_0348", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Self-grading and automated critique in Python.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "dabe0c58e324c9c8c2bdbfe80d61536a1063518d"}
+{"id": "demo_0349", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Diff-first patching and SWE-bench style evaluation in Python.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "ed6687b23f76fa0808d028f46af13a368a5e5155"}
+{"id": "demo_0350", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Rust.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "1a0fe8867460b6a6a80121de51265743a509981b"}
+{"id": "demo_0351", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Governance, auditability, and safety gates in Go.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "6818e4c49db4ba63a4ff27bae5ee7d4d3383255f"}
+{"id": "demo_0352", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in TypeScript.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "aa5e688888ba819372064089899cd28221f9fa41"}
+{"id": "demo_0353", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Python.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "1f8f524e0710190859cc6940196c24af4135327a"}
+{"id": "demo_0354", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Governance, auditability, and safety gates in Python.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "320b4391adf67a3dd800dad890aed898949b23da"}
+{"id": "demo_0355", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Self-grading and automated critique in Python.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "3adfb9c8e53d1b7119af2ca07c587ce30d29dcb5"}
+{"id": "demo_0356", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Governance, auditability, and safety gates in TypeScript.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "e55625866d9d72afa0ca219100fd584e58fd1a20"}
+{"id": "demo_0357", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Governance, auditability, and safety gates in Rust.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "d2ee9ea5723a120aeb73ff995a456cc5d8f41428"}
+{"id": "demo_0358", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Self-grading and automated critique in Rust.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "0fcd85479910d25d7597ad10e547762ff04885d2"}
+{"id": "demo_0359", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Diff-first patching and SWE-bench style evaluation in Rust.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "8dacf6e08ac4a42cf38a0c147636cfdd66f96f97"}
+{"id": "demo_0360", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for MoE expert routing for code in Go.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "55f016e0f5b3b2fae7261a93255883cdb7e2e829"}
+{"id": "demo_0361", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for MoE expert routing for code in Rust.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "e7dcb2a63f3bf8de88060e0b2a7a5c6f548c4deb"}
+{"id": "demo_0362", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for Self-grading and automated critique in TypeScript.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "8c70093964bcfcfa32b3ab7b473fcf5233227529"}
+{"id": "demo_0363", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Tool-calling traces and CI integration in Python.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "9d80a224ffbc2b265069c9355557fa2452ab1b46"}
+{"id": "demo_0364", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Diff-first patching and SWE-bench style evaluation in Python.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "d4e906daed3e71a92b9936a54b063c36f18f0217"}
+{"id": "demo_0365", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Governance, auditability, and safety gates in Go.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "b0ee387d89f5f635c1f9cc32f9773b4068014798"}
+{"id": "demo_0366", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for Governance, auditability, and safety gates in Go.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "f58a9b02ebf152bc1c7f48fa9d5cda2b687e07c8"}
+{"id": "demo_0367", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for MoE expert routing for code in Rust.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "2f0099d7d55722348ce0b94bc5863892091131f8"}
+{"id": "demo_0368", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Go.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "71257b72d4d5b4e500f00ddada2361456c6ab673"}
+{"id": "demo_0369", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for Governance, auditability, and safety gates in TypeScript.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "e55625866d9d72afa0ca219100fd584e58fd1a20"}
+{"id": "demo_0370", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Governance, auditability, and safety gates in Rust.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "57ef0242e37ea02626901d906d595b09bf7d9d85"}
+{"id": "demo_0371", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Go.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "f4a8b6d07ae6c10d7d031fd3e71902189cfbece4"}
+{"id": "demo_0372", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Self-grading and automated critique in Go.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "85d5a5bc6f30d51ffc33cee542d2c5f2d1d15657"}
+{"id": "demo_0373", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Self-grading and automated critique in Python.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "dabe0c58e324c9c8c2bdbfe80d61536a1063518d"}
+{"id": "demo_0374", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for MoE expert routing for code in Python.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "cdab494c7ea40443854dc9e76f81d9692cba5fad"}
+{"id": "demo_0375", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Python.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "cabf427b92e83fe95318ff3a5c2d46bd31d67e6e"}
+{"id": "demo_0376", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Go.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "ff4427b1aeb8acedadc80320a21a8044fc27397e"}
+{"id": "demo_0377", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Governance, auditability, and safety gates in Rust.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "d2ee9ea5723a120aeb73ff995a456cc5d8f41428"}
+{"id": "demo_0378", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for MoE expert routing for code in Rust.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "e7dcb2a63f3bf8de88060e0b2a7a5c6f548c4deb"}
+{"id": "demo_0379", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in TypeScript.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "fab772571976b248bba6e42050c283d2dcacdc46"}
+{"id": "demo_0380", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Tool-calling traces and CI integration in TypeScript.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "b70cf5f1b56ffca136b21a28a0f03cb2993dfc03"}
+{"id": "demo_0381", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Self-grading and automated critique in Go.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "be168e7e68cd992bda3c031199a780770e3d7833"}
+{"id": "demo_0382", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Self-grading and automated critique in Rust.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "a43c2fa4a3ea9a8fc5ec43f2a91fd9d185e88885"}
+{"id": "demo_0383", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for MoE expert routing for code in TypeScript.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "84c33edd4bbd46f9b7ed6f724fa0c0c0891f9673"}
+{"id": "demo_0384", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Self-grading and automated critique in Go.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "5effe6fde62e48aa6e9be1d6e57385f23da7e153"}
+{"id": "demo_0385", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for MoE expert routing for code in TypeScript.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "84c33edd4bbd46f9b7ed6f724fa0c0c0891f9673"}
+{"id": "demo_0386", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Governance, auditability, and safety gates in Rust.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "57ef0242e37ea02626901d906d595b09bf7d9d85"}
+{"id": "demo_0387", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Governance, auditability, and safety gates in TypeScript.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "00ebe4210e2b4c255de24fd7a88ee68a0837ec64"}
+{"id": "demo_0388", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in TypeScript.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "bb6f07fec726ce3ac5b3ef6b3d61c76abe83fd3d"}
+{"id": "demo_0389", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Governance, auditability, and safety gates in Rust.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "407f5e4eeddb9a55f35db2f65cf2cc7ffcb2996e"}
+{"id": "demo_0390", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Governance, auditability, and safety gates in Go.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "b0ee387d89f5f635c1f9cc32f9773b4068014798"}
+{"id": "demo_0391", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Python.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "ce6f758b1bc428101ae12ad43bff7a4e88c0cfc7"}
+{"id": "demo_0392", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for MoE expert routing for code in Rust.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "6a015cfda615e97b9a9735cff067ff27b811115e"}
+{"id": "demo_0393", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Governance, auditability, and safety gates in Rust.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "57ef0242e37ea02626901d906d595b09bf7d9d85"}
+{"id": "demo_0394", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Self-grading and automated critique in TypeScript.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "5ecb56d92db1e83dd4aeacdea1ea7ed21c892010"}
+{"id": "demo_0395", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Governance, auditability, and safety gates in Go.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "b0ee387d89f5f635c1f9cc32f9773b4068014798"}
+{"id": "demo_0396", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Diff-first patching and SWE-bench style evaluation in Python.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "52094c608f86848d71b36f8b28490eeab830b40f"}
+{"id": "demo_0397", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for MoE expert routing for code in Python.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "cdab494c7ea40443854dc9e76f81d9692cba5fad"}
+{"id": "demo_0398", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Tool-calling traces and CI integration in TypeScript.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "b70cf5f1b56ffca136b21a28a0f03cb2993dfc03"}
+{"id": "demo_0399", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Python.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "79a6d964f0af5a0c080ec8cbfafd1114d574e578"}
+{"id": "demo_0400", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for MoE expert routing for code in TypeScript.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "04cf27cd6c7cc60a6ad34383c06c00220df1d7d6"}
+{"id": "demo_0401", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Governance, auditability, and safety gates in TypeScript.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "4e706c8a193f8c3c233720689ef537626b4b4d05"}
+{"id": "demo_0402", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Governance, auditability, and safety gates in Go.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "b0ee387d89f5f635c1f9cc32f9773b4068014798"}
+{"id": "demo_0403", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for MoE expert routing for code in Rust.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "f230672f0116cff639eb4281d9ead7c3e278e6c9"}
+{"id": "demo_0404", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Diff-first patching and SWE-bench style evaluation in Python.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "ed6687b23f76fa0808d028f46af13a368a5e5155"}
+{"id": "demo_0405", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Governance, auditability, and safety gates in Python.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "e026f2a318ea6750793d9529645190cb6103242d"}
+{"id": "demo_0406", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Tool-calling traces and CI integration in Go.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "3b6175b68f17af9ab68a3e4275a82a74736d9b6d"}
+{"id": "demo_0407", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in TypeScript.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "3f20b9c5d6121cdd67b0ae805ccb9f1ee72869a7"}
+{"id": "demo_0408", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for MoE expert routing for code in Go.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "368f0d7a01cf671c54bc9d9be2e070f6ef61d3dc"}
+{"id": "demo_0409", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Tool-calling traces and CI integration in Rust.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "6771342ab90a99fa5a829bbba8c24379d3ed4b2d"}
+{"id": "demo_0410", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for MoE expert routing for code in Rust.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "2f0099d7d55722348ce0b94bc5863892091131f8"}
+{"id": "demo_0411", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Self-grading and automated critique in Go.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "c012733817f6f8ceeff47e5dc6c8bdcfb44433fe"}
+{"id": "demo_0412", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Diff-first patching and SWE-bench style evaluation in Go.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "8e9b6a48836548d0f9202dc0384ed0d98e123f41"}
+{"id": "demo_0413", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Diff-first patching and SWE-bench style evaluation in Rust.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "8dacf6e08ac4a42cf38a0c147636cfdd66f96f97"}
+{"id": "demo_0414", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in TypeScript.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "fab772571976b248bba6e42050c283d2dcacdc46"}
+{"id": "demo_0415", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Self-grading and automated critique in Go.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "6ddfb66b8ab85ff28b1c9d587350b2c6c891a905"}
+{"id": "demo_0416", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Rust.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "57884ed4c3d00bb495a4382eb0ea01a961e42e16"}
+{"id": "demo_0417", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Self-grading and automated critique in Python.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "cf868e52e7b626f6e1445e520642918b3a2a1910"}
+{"id": "demo_0418", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Diff-first patching and SWE-bench style evaluation in Go.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "844412bf05cbb81a55c5e45e3a96939e61f6efd0"}
+{"id": "demo_0419", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for MoE expert routing for code in Rust.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "e7dcb2a63f3bf8de88060e0b2a7a5c6f548c4deb"}
+{"id": "demo_0420", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Tool-calling traces and CI integration in TypeScript.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "b70cf5f1b56ffca136b21a28a0f03cb2993dfc03"}
+{"id": "demo_0421", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Rust.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "57884ed4c3d00bb495a4382eb0ea01a961e42e16"}
+{"id": "demo_0422", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Self-grading and automated critique in TypeScript.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "f31708cf2dc2bf12377b51649cb192e41c358dac"}
+{"id": "demo_0423", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in TypeScript.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "97e3218662e2f00746256643faef6885ea18cd7b"}
+{"id": "demo_0424", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Diff-first patching and SWE-bench style evaluation in Rust.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "b1045ddb05b3c862b41b79b8eba5e07fecfd99b9"}
+{"id": "demo_0425", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Governance, auditability, and safety gates in TypeScript.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "e55625866d9d72afa0ca219100fd584e58fd1a20"}
+{"id": "demo_0426", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for MoE expert routing for code in Go.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "724a661b7365a7ceedaa2c6eb1ec2248b6dc7be0"}
+{"id": "demo_0427", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Self-grading and automated critique in Rust.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "0fcd85479910d25d7597ad10e547762ff04885d2"}
+{"id": "demo_0428", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for MoE expert routing for code in Python.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "d10ebbd0771e561742de9157381a089d64c68539"}
+{"id": "demo_0429", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in TypeScript.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "3f20b9c5d6121cdd67b0ae805ccb9f1ee72869a7"}
+{"id": "demo_0430", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in TypeScript.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "bb6f07fec726ce3ac5b3ef6b3d61c76abe83fd3d"}
+{"id": "demo_0431", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for MoE expert routing for code in Rust.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "f230672f0116cff639eb4281d9ead7c3e278e6c9"}
+{"id": "demo_0432", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Tool-calling traces and CI integration in Rust.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "6771342ab90a99fa5a829bbba8c24379d3ed4b2d"}
+{"id": "demo_0433", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Tool-calling traces and CI integration in Go.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "da7d498e0aa830fb9d9079a323908d7d3e164681"}
+{"id": "demo_0434", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Diff-first patching and SWE-bench style evaluation in Python.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "d4e906daed3e71a92b9936a54b063c36f18f0217"}
+{"id": "demo_0435", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Tool-calling traces and CI integration in Rust.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "6771342ab90a99fa5a829bbba8c24379d3ed4b2d"}
+{"id": "demo_0436", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Self-grading and automated critique in Python.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "dabe0c58e324c9c8c2bdbfe80d61536a1063518d"}
+{"id": "demo_0437", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Tool-calling traces and CI integration in TypeScript.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "47ed1415b5f3bf9882a334b6c73ccdeb3ada9c95"}
+{"id": "demo_0438", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Governance, auditability, and safety gates in Go.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "6818e4c49db4ba63a4ff27bae5ee7d4d3383255f"}
+{"id": "demo_0439", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Tool-calling traces and CI integration in TypeScript.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "f3b404c1695427e8871ad8a5906b0fc4c088966a"}
+{"id": "demo_0440", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Python.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "ce6f758b1bc428101ae12ad43bff7a4e88c0cfc7"}
+{"id": "demo_0441", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Go.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "71257b72d4d5b4e500f00ddada2361456c6ab673"}
+{"id": "demo_0442", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Diff-first patching and SWE-bench style evaluation in Python.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "52094c608f86848d71b36f8b28490eeab830b40f"}
+{"id": "demo_0443", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in TypeScript.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "97e3218662e2f00746256643faef6885ea18cd7b"}
+{"id": "demo_0444", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for Tool-calling traces and CI integration in Rust.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "256c808b21b4248d2f7c4c8c03d35d40bff23d5c"}
+{"id": "demo_0445", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Governance, auditability, and safety gates in Go.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "6818e4c49db4ba63a4ff27bae5ee7d4d3383255f"}
+{"id": "demo_0446", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Diff-first patching and SWE-bench style evaluation in TypeScript.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "7b045c7e00592cb0936703c6efbe07d24fed1640"}
+{"id": "demo_0447", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Go.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "71257b72d4d5b4e500f00ddada2361456c6ab673"}
+{"id": "demo_0448", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Tool-calling traces and CI integration in TypeScript.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "b70cf5f1b56ffca136b21a28a0f03cb2993dfc03"}
+{"id": "demo_0449", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for Self-grading and automated critique in Rust.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "0fcd85479910d25d7597ad10e547762ff04885d2"}
+{"id": "demo_0450", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Self-grading and automated critique in Python.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "54a082fe1d81e29f0620aed5d29e8a89b05a7741"}
+{"id": "demo_0451", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Go.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "d165292165be3a4c543c7636d89fefd10b078e7c"}
+{"id": "demo_0452", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Governance, auditability, and safety gates in Rust.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "d2ee9ea5723a120aeb73ff995a456cc5d8f41428"}
+{"id": "demo_0453", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Self-grading and automated critique in Rust.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "a43c2fa4a3ea9a8fc5ec43f2a91fd9d185e88885"}
+{"id": "demo_0454", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Tool-calling traces and CI integration in Rust.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "bfa0ece85ebe9acb845d6bef2bf266c580695c9b"}
+{"id": "demo_0455", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for MoE expert routing for code in TypeScript.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "a4322f9593d84f6099aa35b3061f2aa8a7adab8e"}
+{"id": "demo_0456", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for MoE expert routing for code in Rust.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "e7dcb2a63f3bf8de88060e0b2a7a5c6f548c4deb"}
+{"id": "demo_0457", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Python.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "cabf427b92e83fe95318ff3a5c2d46bd31d67e6e"}
+{"id": "demo_0458", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Rust.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "1a0fe8867460b6a6a80121de51265743a509981b"}
+{"id": "demo_0459", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Self-grading and automated critique in TypeScript.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "4fc6dade0f8d7671f7cbe49b182676215b21cf70"}
+{"id": "demo_0460", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for MoE expert routing for code in Rust.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "2f0099d7d55722348ce0b94bc5863892091131f8"}
+{"id": "demo_0461", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Governance, auditability, and safety gates in Rust.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "d2ee9ea5723a120aeb73ff995a456cc5d8f41428"}
+{"id": "demo_0462", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Governance, auditability, and safety gates in Go.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "b5dbf5c5ae7587a140ab91a3043227e4c81100fe"}
+{"id": "demo_0463", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Diff-first patching and SWE-bench style evaluation in Go.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "0a0e75aa64f2d9953fa6bcddadb57aa49fa969c6"}
+{"id": "demo_0464", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Python.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "1f8f524e0710190859cc6940196c24af4135327a"}
+{"id": "demo_0465", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for MoE expert routing for code in TypeScript.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "84c33edd4bbd46f9b7ed6f724fa0c0c0891f9673"}
+{"id": "demo_0466", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Self-grading and automated critique in Rust.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "0cb9a5e8d76f6bfd1e0d104cb5af19b52e4f9fae"}
+{"id": "demo_0467", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Python.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "1f8f524e0710190859cc6940196c24af4135327a"}
+{"id": "demo_0468", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Diff-first patching and SWE-bench style evaluation in Python.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "ed6687b23f76fa0808d028f46af13a368a5e5155"}
+{"id": "demo_0469", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Self-grading and automated critique in Rust.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "d2815ca3d0e438822e466acc43474ba2874f5aad"}
+{"id": "demo_0470", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Tool-calling traces and CI integration in Python.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "9d80a224ffbc2b265069c9355557fa2452ab1b46"}
+{"id": "demo_0471", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Diff-first patching and SWE-bench style evaluation in Go.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "8e9b6a48836548d0f9202dc0384ed0d98e123f41"}
+{"id": "demo_0472", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Diff-first patching and SWE-bench style evaluation in Python.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "214c8aa1c81b49da6e98d286de14957000ecb489"}
+{"id": "demo_0473", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Self-grading and automated critique in Python.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "cf868e52e7b626f6e1445e520642918b3a2a1910"}
+{"id": "demo_0474", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Self-grading and automated critique in Go.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "c012733817f6f8ceeff47e5dc6c8bdcfb44433fe"}
+{"id": "demo_0475", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for MoE expert routing for code in Rust.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "6a015cfda615e97b9a9735cff067ff27b811115e"}
+{"id": "demo_0476", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Governance, auditability, and safety gates in TypeScript.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "4e706c8a193f8c3c233720689ef537626b4b4d05"}
+{"id": "demo_0477", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Python.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "79a6d964f0af5a0c080ec8cbfafd1114d574e578"}
+{"id": "demo_0478", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for MoE expert routing for code in Python.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "a7ccacbb4f991696e10c4df1b3a721ca21aaccfa"}
+{"id": "demo_0479", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Tool-calling traces and CI integration in Go.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "b1f3c1daa457ede0e06e5e190d15d6e50635238b"}
+{"id": "demo_0480", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for MoE expert routing for code in Python.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "cdab494c7ea40443854dc9e76f81d9692cba5fad"}
+{"id": "demo_0481", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Go.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "71257b72d4d5b4e500f00ddada2361456c6ab673"}
+{"id": "demo_0482", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Self-grading and automated critique in Rust.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "0fcd85479910d25d7597ad10e547762ff04885d2"}
+{"id": "demo_0483", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for MoE expert routing for code in Go.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "368f0d7a01cf671c54bc9d9be2e070f6ef61d3dc"}
+{"id": "demo_0484", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for MoE expert routing for code in Rust.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "6a015cfda615e97b9a9735cff067ff27b811115e"}
+{"id": "demo_0485", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Go.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "f4a8b6d07ae6c10d7d031fd3e71902189cfbece4"}
+{"id": "demo_0486", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Governance, auditability, and safety gates in Go.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "b0ee387d89f5f635c1f9cc32f9773b4068014798"}
+{"id": "demo_0487", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Tool-calling traces and CI integration in Rust.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "2ad46a119bbbd13e9f59ca46d56fc6ceffd8c0ad"}
+{"id": "demo_0488", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for MoE expert routing for code in TypeScript.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "84c33edd4bbd46f9b7ed6f724fa0c0c0891f9673"}
+{"id": "demo_0489", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Go.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "532dd88780c2362d1ef34dea626ff108239e845b"}
+{"id": "demo_0490", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Governance, auditability, and safety gates in Python.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "e026f2a318ea6750793d9529645190cb6103242d"}
+{"id": "demo_0491", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Self-grading and automated critique in Rust.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "d2815ca3d0e438822e466acc43474ba2874f5aad"}
+{"id": "demo_0492", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Go.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "f4a8b6d07ae6c10d7d031fd3e71902189cfbece4"}
+{"id": "demo_0493", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Diff-first patching and SWE-bench style evaluation in Python.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "399b590cd03414ba736ffb6888734b51cacb8ca9"}
+{"id": "demo_0494", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Governance, auditability, and safety gates in TypeScript.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "4e706c8a193f8c3c233720689ef537626b4b4d05"}
+{"id": "demo_0495", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Governance, auditability, and safety gates in Rust.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "29b4334a491410a737f41c7f169c9fcc2e01bf01"}
+{"id": "demo_0496", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Python.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "ce6f758b1bc428101ae12ad43bff7a4e88c0cfc7"}
+{"id": "demo_0497", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for MoE expert routing for code in Go.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "c77684ed94a4148f8449f9b073cd70d3bc577a1d"}
+{"id": "demo_0498", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Governance, auditability, and safety gates in Go.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "6818e4c49db4ba63a4ff27bae5ee7d4d3383255f"}
+{"id": "demo_0499", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in TypeScript.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "bb6f07fec726ce3ac5b3ef6b3d61c76abe83fd3d"}
+{"id": "demo_0500", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Governance, auditability, and safety gates in Rust.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "d2ee9ea5723a120aeb73ff995a456cc5d8f41428"}
+{"id": "demo_0501", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Rust.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "57884ed4c3d00bb495a4382eb0ea01a961e42e16"}
+{"id": "demo_0502", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Governance, auditability, and safety gates in TypeScript.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "740d9e6139f5f9bbc02e9d9236e6d6023f4948ff"}
+{"id": "demo_0503", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in TypeScript.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "3f20b9c5d6121cdd67b0ae805ccb9f1ee72869a7"}
+{"id": "demo_0504", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Governance, auditability, and safety gates in Python.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "e026f2a318ea6750793d9529645190cb6103242d"}
+{"id": "demo_0505", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Tool-calling traces and CI integration in TypeScript.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "f3b404c1695427e8871ad8a5906b0fc4c088966a"}
+{"id": "demo_0506", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Tool-calling traces and CI integration in Python.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "9d80a224ffbc2b265069c9355557fa2452ab1b46"}
+{"id": "demo_0507", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Go.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "532dd88780c2362d1ef34dea626ff108239e845b"}
+{"id": "demo_0508", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Tool-calling traces and CI integration in Go.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "da7d498e0aa830fb9d9079a323908d7d3e164681"}
+{"id": "demo_0509", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for Governance, auditability, and safety gates in Python.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "f2b051c203782bf1ac25353b2085f0ee12d6e31e"}
+{"id": "demo_0510", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Self-grading and automated critique in Go.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "85d5a5bc6f30d51ffc33cee542d2c5f2d1d15657"}
+{"id": "demo_0511", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Rust.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "370dec51cc86af4647aa0881c7e9b9f3a6bbcfce"}
+{"id": "demo_0512", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in TypeScript.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "aa5e688888ba819372064089899cd28221f9fa41"}
+{"id": "demo_0513", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Self-grading and automated critique in Go.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "c012733817f6f8ceeff47e5dc6c8bdcfb44433fe"}
+{"id": "demo_0514", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Tool-calling traces and CI integration in TypeScript.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "7a36c2ca2b689bcc99d6e92fcc5a9d64f4e84ece"}
+{"id": "demo_0515", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Go.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "d165292165be3a4c543c7636d89fefd10b078e7c"}
+{"id": "demo_0516", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Governance, auditability, and safety gates in Rust.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "407f5e4eeddb9a55f35db2f65cf2cc7ffcb2996e"}
+{"id": "demo_0517", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Rust.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "57884ed4c3d00bb495a4382eb0ea01a961e42e16"}
+{"id": "demo_0518", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for MoE expert routing for code in Go.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "c77684ed94a4148f8449f9b073cd70d3bc577a1d"}
+{"id": "demo_0519", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Rust.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "1a0fe8867460b6a6a80121de51265743a509981b"}
+{"id": "demo_0520", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Python.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "cabf427b92e83fe95318ff3a5c2d46bd31d67e6e"}
+{"id": "demo_0521", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Tool-calling traces and CI integration in TypeScript.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "7a36c2ca2b689bcc99d6e92fcc5a9d64f4e84ece"}
+{"id": "demo_0522", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for MoE expert routing for code in Go.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "724a661b7365a7ceedaa2c6eb1ec2248b6dc7be0"}
+{"id": "demo_0523", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in TypeScript.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "97e3218662e2f00746256643faef6885ea18cd7b"}
+{"id": "demo_0524", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for MoE expert routing for code in Go.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "c77684ed94a4148f8449f9b073cd70d3bc577a1d"}
+{"id": "demo_0525", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Python.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "1f8f524e0710190859cc6940196c24af4135327a"}
+{"id": "demo_0526", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Self-grading and automated critique in TypeScript.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "8c70093964bcfcfa32b3ab7b473fcf5233227529"}
+{"id": "demo_0527", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for Governance, auditability, and safety gates in TypeScript.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "e55625866d9d72afa0ca219100fd584e58fd1a20"}
+{"id": "demo_0528", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Governance, auditability, and safety gates in Rust.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "29b4334a491410a737f41c7f169c9fcc2e01bf01"}
+{"id": "demo_0529", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Tool-calling traces and CI integration in Python.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "aa6c313c61194f09ceaa17d370a93b648759db00"}
+{"id": "demo_0530", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Diff-first patching and SWE-bench style evaluation in TypeScript.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "7b045c7e00592cb0936703c6efbe07d24fed1640"}
+{"id": "demo_0531", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Tool-calling traces and CI integration in TypeScript.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "f3b404c1695427e8871ad8a5906b0fc4c088966a"}
+{"id": "demo_0532", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Self-grading and automated critique in TypeScript.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "4fc6dade0f8d7671f7cbe49b182676215b21cf70"}
+{"id": "demo_0533", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for MoE expert routing for code in Python.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "cdab494c7ea40443854dc9e76f81d9692cba5fad"}
+{"id": "demo_0534", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Tool-calling traces and CI integration in Rust.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "bfa0ece85ebe9acb845d6bef2bf266c580695c9b"}
+{"id": "demo_0535", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for MoE expert routing for code in Rust.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "a757555ae2d72f4f9a6c937f6ef6303650b71604"}
+{"id": "demo_0536", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Diff-first patching and SWE-bench style evaluation in Go.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "9ac511806d0fd80846c9ee66e870bc422000ca5d"}
+{"id": "demo_0537", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Self-grading and automated critique in TypeScript.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "d02ca6f3f580c87df094e25ecb3de78fc576555b"}
+{"id": "demo_0538", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Tool-calling traces and CI integration in Go.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "da7d498e0aa830fb9d9079a323908d7d3e164681"}
+{"id": "demo_0539", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Diff-first patching and SWE-bench style evaluation in TypeScript.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "1971ecc0a5017fee37a2751e068efa0ba9bdb6b7"}
+{"id": "demo_0540", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Diff-first patching and SWE-bench style evaluation in TypeScript.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "1971ecc0a5017fee37a2751e068efa0ba9bdb6b7"}
+{"id": "demo_0541", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Governance, auditability, and safety gates in Go.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "b5dbf5c5ae7587a140ab91a3043227e4c81100fe"}
+{"id": "demo_0542", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for MoE expert routing for code in Rust.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "f230672f0116cff639eb4281d9ead7c3e278e6c9"}
+{"id": "demo_0543", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Python.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "79a6d964f0af5a0c080ec8cbfafd1114d574e578"}
+{"id": "demo_0544", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Self-grading and automated critique in Python.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "3adfb9c8e53d1b7119af2ca07c587ce30d29dcb5"}
+{"id": "demo_0545", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Self-grading and automated critique in Python.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "3adfb9c8e53d1b7119af2ca07c587ce30d29dcb5"}
+{"id": "demo_0546", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for MoE expert routing for code in Go.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "441e2206b28951463f35934572603bd6fa2398d7"}
+{"id": "demo_0547", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for MoE expert routing for code in Python.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "fd55fdc9ecf12df2b0b5a36e7f1f0f93d303dbb3"}
+{"id": "demo_0548", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Tool-calling traces and CI integration in Go.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "da7d498e0aa830fb9d9079a323908d7d3e164681"}
+{"id": "demo_0549", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for MoE expert routing for code in TypeScript.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "84c33edd4bbd46f9b7ed6f724fa0c0c0891f9673"}
+{"id": "demo_0550", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for MoE expert routing for code in Python.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "d10ebbd0771e561742de9157381a089d64c68539"}
+{"id": "demo_0551", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Go.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "71257b72d4d5b4e500f00ddada2361456c6ab673"}
+{"id": "demo_0552", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Tool-calling traces and CI integration in Rust.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "2ad46a119bbbd13e9f59ca46d56fc6ceffd8c0ad"}
+{"id": "demo_0553", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Diff-first patching and SWE-bench style evaluation in Go.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "1ede0fbd7add911639c09c0d992869248542401b"}
+{"id": "demo_0554", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for MoE expert routing for code in Python.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "fd55fdc9ecf12df2b0b5a36e7f1f0f93d303dbb3"}
+{"id": "demo_0555", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Diff-first patching and SWE-bench style evaluation in Python.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "52094c608f86848d71b36f8b28490eeab830b40f"}
+{"id": "demo_0556", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for MoE expert routing for code in Rust.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "6a015cfda615e97b9a9735cff067ff27b811115e"}
+{"id": "demo_0557", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Diff-first patching and SWE-bench style evaluation in Rust.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "6aef16a5ce69387a6f198bdb7d380162bb772ca2"}
+{"id": "demo_0558", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Governance, auditability, and safety gates in Go.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "6818e4c49db4ba63a4ff27bae5ee7d4d3383255f"}
+{"id": "demo_0559", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Governance, auditability, and safety gates in Python.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "e026f2a318ea6750793d9529645190cb6103242d"}
+{"id": "demo_0560", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for MoE expert routing for code in Python.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "a7ccacbb4f991696e10c4df1b3a721ca21aaccfa"}
+{"id": "demo_0561", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Tool-calling traces and CI integration in Rust.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "bfa0ece85ebe9acb845d6bef2bf266c580695c9b"}
+{"id": "demo_0562", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Go.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "532dd88780c2362d1ef34dea626ff108239e845b"}
+{"id": "demo_0563", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Self-grading and automated critique in Go.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "be168e7e68cd992bda3c031199a780770e3d7833"}
+{"id": "demo_0564", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Self-grading and automated critique in Python.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "cf868e52e7b626f6e1445e520642918b3a2a1910"}
+{"id": "demo_0565", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Diff-first patching and SWE-bench style evaluation in Rust.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "6aef16a5ce69387a6f198bdb7d380162bb772ca2"}
+{"id": "demo_0566", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Diff-first patching and SWE-bench style evaluation in Python.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "d4e906daed3e71a92b9936a54b063c36f18f0217"}
+{"id": "demo_0567", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Governance, auditability, and safety gates in Rust.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "29b4334a491410a737f41c7f169c9fcc2e01bf01"}
+{"id": "demo_0568", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Diff-first patching and SWE-bench style evaluation in Go.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "1ede0fbd7add911639c09c0d992869248542401b"}
+{"id": "demo_0569", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Tool-calling traces and CI integration in Rust.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "2ad46a119bbbd13e9f59ca46d56fc6ceffd8c0ad"}
+{"id": "demo_0570", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Tool-calling traces and CI integration in Python.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "aa6c313c61194f09ceaa17d370a93b648759db00"}
+{"id": "demo_0571", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Diff-first patching and SWE-bench style evaluation in Go.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "9ac511806d0fd80846c9ee66e870bc422000ca5d"}
+{"id": "demo_0572", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in TypeScript.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "fab772571976b248bba6e42050c283d2dcacdc46"}
+{"id": "demo_0573", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Diff-first patching and SWE-bench style evaluation in Python.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "ed6687b23f76fa0808d028f46af13a368a5e5155"}
+{"id": "demo_0574", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in TypeScript.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "fab772571976b248bba6e42050c283d2dcacdc46"}
+{"id": "demo_0575", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for MoE expert routing for code in Go.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "441e2206b28951463f35934572603bd6fa2398d7"}
+{"id": "demo_0576", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in TypeScript.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "bb6f07fec726ce3ac5b3ef6b3d61c76abe83fd3d"}
+{"id": "demo_0577", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for Diff-first patching and SWE-bench style evaluation in TypeScript.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "1971ecc0a5017fee37a2751e068efa0ba9bdb6b7"}
+{"id": "demo_0578", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Diff-first patching and SWE-bench style evaluation in Rust.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "8dacf6e08ac4a42cf38a0c147636cfdd66f96f97"}
+{"id": "demo_0579", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Tool-calling traces and CI integration in TypeScript.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "b70cf5f1b56ffca136b21a28a0f03cb2993dfc03"}
+{"id": "demo_0580", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Tool-calling traces and CI integration in Go.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "01ecda677e27b8918134d21a0d1cd85ef04617f9"}
+{"id": "demo_0581", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for MoE expert routing for code in Rust.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "f230672f0116cff639eb4281d9ead7c3e278e6c9"}
+{"id": "demo_0582", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for MoE expert routing for code in TypeScript.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "a4322f9593d84f6099aa35b3061f2aa8a7adab8e"}
+{"id": "demo_0583", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for MoE expert routing for code in TypeScript.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "c6a3c7c940db60ab9c0aad92f2280b426d27d44d"}
+{"id": "demo_0584", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for MoE expert routing for code in TypeScript.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "5d8081fa4e2b4d2c429db2955e8f9dd90080d145"}
+{"id": "demo_0585", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Tool-calling traces and CI integration in Go.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "3b6175b68f17af9ab68a3e4275a82a74736d9b6d"}
+{"id": "demo_0586", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for MoE expert routing for code in Rust.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "e7dcb2a63f3bf8de88060e0b2a7a5c6f548c4deb"}
+{"id": "demo_0587", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Diff-first patching and SWE-bench style evaluation in Python.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "214c8aa1c81b49da6e98d286de14957000ecb489"}
+{"id": "demo_0588", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Governance, auditability, and safety gates in Go.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "6818e4c49db4ba63a4ff27bae5ee7d4d3383255f"}
+{"id": "demo_0589", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for MoE expert routing for code in Go.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "55f016e0f5b3b2fae7261a93255883cdb7e2e829"}
+{"id": "demo_0590", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Governance, auditability, and safety gates in Rust.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "d2ee9ea5723a120aeb73ff995a456cc5d8f41428"}
+{"id": "demo_0591", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Tool-calling traces and CI integration in Python.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "3901f67d39afc0bfe5397c2c27f8ead5b2b472e7"}
+{"id": "demo_0592", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for MoE expert routing for code in Rust.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "a757555ae2d72f4f9a6c937f6ef6303650b71604"}
+{"id": "demo_0593", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Tool-calling traces and CI integration in Rust.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "2ad46a119bbbd13e9f59ca46d56fc6ceffd8c0ad"}
+{"id": "demo_0594", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Diff-first patching and SWE-bench style evaluation in Rust.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "b1045ddb05b3c862b41b79b8eba5e07fecfd99b9"}
+{"id": "demo_0595", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Governance, auditability, and safety gates in TypeScript.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "e55625866d9d72afa0ca219100fd584e58fd1a20"}
+{"id": "demo_0596", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Rust.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "8cd98daa7ed94eceb808aace257ad77a00c093c1"}
+{"id": "demo_0597", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for MoE expert routing for code in Go.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "55f016e0f5b3b2fae7261a93255883cdb7e2e829"}
+{"id": "demo_0598", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Rust.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "57884ed4c3d00bb495a4382eb0ea01a961e42e16"}
+{"id": "demo_0599", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for MoE expert routing for code in Rust.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "f230672f0116cff639eb4281d9ead7c3e278e6c9"}
+{"id": "demo_0600", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for Self-grading and automated critique in Go.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "5effe6fde62e48aa6e9be1d6e57385f23da7e153"}
+{"id": "demo_0601", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Diff-first patching and SWE-bench style evaluation in Go.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "0a0e75aa64f2d9953fa6bcddadb57aa49fa969c6"}
+{"id": "demo_0602", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Self-grading and automated critique in Python.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "cf868e52e7b626f6e1445e520642918b3a2a1910"}
+{"id": "demo_0603", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for MoE expert routing for code in Go.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "368f0d7a01cf671c54bc9d9be2e070f6ef61d3dc"}
+{"id": "demo_0604", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Python.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "1f8f524e0710190859cc6940196c24af4135327a"}
+{"id": "demo_0605", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Tool-calling traces and CI integration in Python.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "9d80a224ffbc2b265069c9355557fa2452ab1b46"}
+{"id": "demo_0606", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Tool-calling traces and CI integration in Rust.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "6b80883d024afe8123d755ffa41e5b3cc0dcfb74"}
+{"id": "demo_0607", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Rust.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "1a0fe8867460b6a6a80121de51265743a509981b"}
+{"id": "demo_0608", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for Diff-first patching and SWE-bench style evaluation in Python.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "399b590cd03414ba736ffb6888734b51cacb8ca9"}
+{"id": "demo_0609", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for MoE expert routing for code in Go.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "c77684ed94a4148f8449f9b073cd70d3bc577a1d"}
+{"id": "demo_0610", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Self-grading and automated critique in TypeScript.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "f31708cf2dc2bf12377b51649cb192e41c358dac"}
+{"id": "demo_0611", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Rust.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "8cd98daa7ed94eceb808aace257ad77a00c093c1"}
+{"id": "demo_0612", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Self-grading and automated critique in Go.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "be168e7e68cd992bda3c031199a780770e3d7833"}
+{"id": "demo_0613", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Tool-calling traces and CI integration in TypeScript.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "7a36c2ca2b689bcc99d6e92fcc5a9d64f4e84ece"}
+{"id": "demo_0614", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for MoE expert routing for code in Python.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "fd55fdc9ecf12df2b0b5a36e7f1f0f93d303dbb3"}
+{"id": "demo_0615", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in TypeScript.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "97e3218662e2f00746256643faef6885ea18cd7b"}
+{"id": "demo_0616", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Python.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "b7325d57cdc539740421eea3ad320dc75e3a5809"}
+{"id": "demo_0617", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Governance, auditability, and safety gates in TypeScript.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "00ebe4210e2b4c255de24fd7a88ee68a0837ec64"}
+{"id": "demo_0618", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Governance, auditability, and safety gates in Rust.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "407f5e4eeddb9a55f35db2f65cf2cc7ffcb2996e"}
+{"id": "demo_0619", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Python.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "1f8f524e0710190859cc6940196c24af4135327a"}
+{"id": "demo_0620", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for MoE expert routing for code in Go.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "441e2206b28951463f35934572603bd6fa2398d7"}
+{"id": "demo_0621", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Governance, auditability, and safety gates in Go.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "b0ee387d89f5f635c1f9cc32f9773b4068014798"}
+{"id": "demo_0622", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Tool-calling traces and CI integration in Go.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "01ecda677e27b8918134d21a0d1cd85ef04617f9"}
+{"id": "demo_0623", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Diff-first patching and SWE-bench style evaluation in Go.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "8e9b6a48836548d0f9202dc0384ed0d98e123f41"}
+{"id": "demo_0624", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Diff-first patching and SWE-bench style evaluation in Python.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "52094c608f86848d71b36f8b28490eeab830b40f"}
+{"id": "demo_0625", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Diff-first patching and SWE-bench style evaluation in Rust.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "882626e1c64172100c8201be60a636d13fc4cd1d"}
+{"id": "demo_0626", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Governance, auditability, and safety gates in Python.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "f2b051c203782bf1ac25353b2085f0ee12d6e31e"}
+{"id": "demo_0627", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for Self-grading and automated critique in Go.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "5effe6fde62e48aa6e9be1d6e57385f23da7e153"}
+{"id": "demo_0628", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in TypeScript.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "aa5e688888ba819372064089899cd28221f9fa41"}
+{"id": "demo_0629", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Governance, auditability, and safety gates in Python.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "f2b051c203782bf1ac25353b2085f0ee12d6e31e"}
+{"id": "demo_0630", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Diff-first patching and SWE-bench style evaluation in Rust.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "6aef16a5ce69387a6f198bdb7d380162bb772ca2"}
+{"id": "demo_0631", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for MoE expert routing for code in Go.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "441e2206b28951463f35934572603bd6fa2398d7"}
+{"id": "demo_0632", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for Governance, auditability, and safety gates in Go.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "f58a9b02ebf152bc1c7f48fa9d5cda2b687e07c8"}
+{"id": "demo_0633", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for MoE expert routing for code in TypeScript.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "84c33edd4bbd46f9b7ed6f724fa0c0c0891f9673"}
+{"id": "demo_0634", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Governance, auditability, and safety gates in Go.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "b0ee387d89f5f635c1f9cc32f9773b4068014798"}
+{"id": "demo_0635", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Diff-first patching and SWE-bench style evaluation in Python.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "214c8aa1c81b49da6e98d286de14957000ecb489"}
+{"id": "demo_0636", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Self-grading and automated critique in Rust.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "0fcd85479910d25d7597ad10e547762ff04885d2"}
+{"id": "demo_0637", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Python.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "cabf427b92e83fe95318ff3a5c2d46bd31d67e6e"}
+{"id": "demo_0638", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Governance, auditability, and safety gates in Rust.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "29b4334a491410a737f41c7f169c9fcc2e01bf01"}
+{"id": "demo_0639", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Governance, auditability, and safety gates in Rust.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "407f5e4eeddb9a55f35db2f65cf2cc7ffcb2996e"}
+{"id": "demo_0640", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Self-grading and automated critique in Python.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "54a082fe1d81e29f0620aed5d29e8a89b05a7741"}
+{"id": "demo_0641", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Go.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "532dd88780c2362d1ef34dea626ff108239e845b"}
+{"id": "demo_0642", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Self-grading and automated critique in Python.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "3adfb9c8e53d1b7119af2ca07c587ce30d29dcb5"}
+{"id": "demo_0643", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Tool-calling traces and CI integration in TypeScript.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "2fd96b170ecf4c762b8fa01b03e48f8dc581d55a"}
+{"id": "demo_0644", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Diff-first patching and SWE-bench style evaluation in Rust.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "6aef16a5ce69387a6f198bdb7d380162bb772ca2"}
+{"id": "demo_0645", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Governance, auditability, and safety gates in Go.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "b5dbf5c5ae7587a140ab91a3043227e4c81100fe"}
+{"id": "demo_0646", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Governance, auditability, and safety gates in TypeScript.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "00ebe4210e2b4c255de24fd7a88ee68a0837ec64"}
+{"id": "demo_0647", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for Governance, auditability, and safety gates in TypeScript.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "e55625866d9d72afa0ca219100fd584e58fd1a20"}
+{"id": "demo_0648", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in TypeScript.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "3f20b9c5d6121cdd67b0ae805ccb9f1ee72869a7"}
+{"id": "demo_0649", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Governance, auditability, and safety gates in Rust.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "d2ee9ea5723a120aeb73ff995a456cc5d8f41428"}
+{"id": "demo_0650", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for MoE expert routing for code in Rust.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "a757555ae2d72f4f9a6c937f6ef6303650b71604"}
+{"id": "demo_0651", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Self-grading and automated critique in TypeScript.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "f31708cf2dc2bf12377b51649cb192e41c358dac"}
+{"id": "demo_0652", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in TypeScript.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "bb6f07fec726ce3ac5b3ef6b3d61c76abe83fd3d"}
+{"id": "demo_0653", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Python.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "cabf427b92e83fe95318ff3a5c2d46bd31d67e6e"}
+{"id": "demo_0654", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Governance, auditability, and safety gates in Python.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "e0062d33ddb8b86d135c9ab3dec2288f862109b1"}
+{"id": "demo_0655", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Rust.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "1a0fe8867460b6a6a80121de51265743a509981b"}
+{"id": "demo_0656", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Self-grading and automated critique in Rust.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "a43c2fa4a3ea9a8fc5ec43f2a91fd9d185e88885"}
+{"id": "demo_0657", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Governance, auditability, and safety gates in Python.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "320b4391adf67a3dd800dad890aed898949b23da"}
+{"id": "demo_0658", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Governance, auditability, and safety gates in Python.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "320b4391adf67a3dd800dad890aed898949b23da"}
+{"id": "demo_0659", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Tool-calling traces and CI integration in Go.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "01ecda677e27b8918134d21a0d1cd85ef04617f9"}
+{"id": "demo_0660", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for MoE expert routing for code in Rust.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "2f0099d7d55722348ce0b94bc5863892091131f8"}
+{"id": "demo_0661", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Governance, auditability, and safety gates in Go.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "6818e4c49db4ba63a4ff27bae5ee7d4d3383255f"}
+{"id": "demo_0662", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Diff-first patching and SWE-bench style evaluation in Go.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "0a0e75aa64f2d9953fa6bcddadb57aa49fa969c6"}
+{"id": "demo_0663", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Go.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "ff4427b1aeb8acedadc80320a21a8044fc27397e"}
+{"id": "demo_0664", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Self-grading and automated critique in TypeScript.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "8c70093964bcfcfa32b3ab7b473fcf5233227529"}
+{"id": "demo_0665", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for MoE expert routing for code in Python.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "fd55fdc9ecf12df2b0b5a36e7f1f0f93d303dbb3"}
+{"id": "demo_0666", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Diff-first patching and SWE-bench style evaluation in Go.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "8e9b6a48836548d0f9202dc0384ed0d98e123f41"}
+{"id": "demo_0667", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for MoE expert routing for code in Rust.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "6a015cfda615e97b9a9735cff067ff27b811115e"}
+{"id": "demo_0668", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Self-grading and automated critique in Go.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "85d5a5bc6f30d51ffc33cee542d2c5f2d1d15657"}
+{"id": "demo_0669", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Tool-calling traces and CI integration in TypeScript.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "47ed1415b5f3bf9882a334b6c73ccdeb3ada9c95"}
+{"id": "demo_0670", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Tool-calling traces and CI integration in Rust.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "6b80883d024afe8123d755ffa41e5b3cc0dcfb74"}
+{"id": "demo_0671", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Tool-calling traces and CI integration in Rust.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "256c808b21b4248d2f7c4c8c03d35d40bff23d5c"}
+{"id": "demo_0672", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Self-grading and automated critique in TypeScript.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "8c70093964bcfcfa32b3ab7b473fcf5233227529"}
+{"id": "demo_0673", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Diff-first patching and SWE-bench style evaluation in Python.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "399b590cd03414ba736ffb6888734b51cacb8ca9"}
+{"id": "demo_0674", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for Governance, auditability, and safety gates in Go.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "f58a9b02ebf152bc1c7f48fa9d5cda2b687e07c8"}
+{"id": "demo_0675", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Diff-first patching and SWE-bench style evaluation in Python.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "d4e906daed3e71a92b9936a54b063c36f18f0217"}
+{"id": "demo_0676", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Python.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "cabf427b92e83fe95318ff3a5c2d46bd31d67e6e"}
+{"id": "demo_0677", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Self-grading and automated critique in Rust.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "0cb9a5e8d76f6bfd1e0d104cb5af19b52e4f9fae"}
+{"id": "demo_0678", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Diff-first patching and SWE-bench style evaluation in Go.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "8e9b6a48836548d0f9202dc0384ed0d98e123f41"}
+{"id": "demo_0679", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Self-grading and automated critique in Go.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "5effe6fde62e48aa6e9be1d6e57385f23da7e153"}
+{"id": "demo_0680", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Diff-first patching and SWE-bench style evaluation in Go.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "844412bf05cbb81a55c5e45e3a96939e61f6efd0"}
+{"id": "demo_0681", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Tool-calling traces and CI integration in Rust.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "bfa0ece85ebe9acb845d6bef2bf266c580695c9b"}
+{"id": "demo_0682", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Diff-first patching and SWE-bench style evaluation in Python.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "d4e906daed3e71a92b9936a54b063c36f18f0217"}
+{"id": "demo_0683", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Rust.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "8cd98daa7ed94eceb808aace257ad77a00c093c1"}
+{"id": "demo_0684", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Diff-first patching and SWE-bench style evaluation in Python.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "ed6687b23f76fa0808d028f46af13a368a5e5155"}
+{"id": "demo_0685", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for MoE expert routing for code in TypeScript.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "c6a3c7c940db60ab9c0aad92f2280b426d27d44d"}
+{"id": "demo_0686", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for MoE expert routing for code in TypeScript.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "04cf27cd6c7cc60a6ad34383c06c00220df1d7d6"}
+{"id": "demo_0687", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for MoE expert routing for code in Go.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "441e2206b28951463f35934572603bd6fa2398d7"}
+{"id": "demo_0688", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Tool-calling traces and CI integration in Go.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "ed2736226ef8b9ff056b4302095ccecddcf966bd"}
+{"id": "demo_0689", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Self-grading and automated critique in Python.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "cf868e52e7b626f6e1445e520642918b3a2a1910"}
+{"id": "demo_0690", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Diff-first patching and SWE-bench style evaluation in Go.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "844412bf05cbb81a55c5e45e3a96939e61f6efd0"}
+{"id": "demo_0691", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Diff-first patching and SWE-bench style evaluation in TypeScript.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "7b045c7e00592cb0936703c6efbe07d24fed1640"}
+{"id": "demo_0692", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Diff-first patching and SWE-bench style evaluation in Python.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "52094c608f86848d71b36f8b28490eeab830b40f"}
+{"id": "demo_0693", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Tool-calling traces and CI integration in TypeScript.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "7a36c2ca2b689bcc99d6e92fcc5a9d64f4e84ece"}
+{"id": "demo_0694", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Diff-first patching and SWE-bench style evaluation in Rust.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "6aef16a5ce69387a6f198bdb7d380162bb772ca2"}
+{"id": "demo_0695", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Governance, auditability, and safety gates in Rust.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "57ef0242e37ea02626901d906d595b09bf7d9d85"}
+{"id": "demo_0696", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Diff-first patching and SWE-bench style evaluation in Rust.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "882626e1c64172100c8201be60a636d13fc4cd1d"}
+{"id": "demo_0697", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Tool-calling traces and CI integration in Go.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "3b6175b68f17af9ab68a3e4275a82a74736d9b6d"}
+{"id": "demo_0698", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Tool-calling traces and CI integration in Go.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "da7d498e0aa830fb9d9079a323908d7d3e164681"}
+{"id": "demo_0699", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for Tool-calling traces and CI integration in Go.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "b1f3c1daa457ede0e06e5e190d15d6e50635238b"}
+{"id": "demo_0700", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Self-grading and automated critique in Rust.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "0cb9a5e8d76f6bfd1e0d104cb5af19b52e4f9fae"}
+{"id": "demo_0701", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for MoE expert routing for code in Rust.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "e7dcb2a63f3bf8de88060e0b2a7a5c6f548c4deb"}
+{"id": "demo_0702", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Governance, auditability, and safety gates in Python.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "2cd0a884e47168413d32f61dc500ce673c8e8ce6"}
+{"id": "demo_0703", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Go.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "532dd88780c2362d1ef34dea626ff108239e845b"}
+{"id": "demo_0704", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Governance, auditability, and safety gates in Python.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "e0062d33ddb8b86d135c9ab3dec2288f862109b1"}
+{"id": "demo_0705", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Governance, auditability, and safety gates in Rust.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "66b4ca1544895342a9abd1a6c3586898a0140df4"}
+{"id": "demo_0706", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Diff-first patching and SWE-bench style evaluation in Go.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "1ede0fbd7add911639c09c0d992869248542401b"}
+{"id": "demo_0707", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Governance, auditability, and safety gates in Python.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "2cd0a884e47168413d32f61dc500ce673c8e8ce6"}
+{"id": "demo_0708", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Go.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "532dd88780c2362d1ef34dea626ff108239e845b"}
+{"id": "demo_0709", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Self-grading and automated critique in Go.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "6ddfb66b8ab85ff28b1c9d587350b2c6c891a905"}
+{"id": "demo_0710", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Governance, auditability, and safety gates in TypeScript.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "00ebe4210e2b4c255de24fd7a88ee68a0837ec64"}
+{"id": "demo_0711", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Rust.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "57884ed4c3d00bb495a4382eb0ea01a961e42e16"}
+{"id": "demo_0712", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Rust.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "8cd98daa7ed94eceb808aace257ad77a00c093c1"}
+{"id": "demo_0713", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Tool-calling traces and CI integration in TypeScript.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "f3b404c1695427e8871ad8a5906b0fc4c088966a"}
+{"id": "demo_0714", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Diff-first patching and SWE-bench style evaluation in Go.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "844412bf05cbb81a55c5e45e3a96939e61f6efd0"}
+{"id": "demo_0715", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Diff-first patching and SWE-bench style evaluation in Go.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "844412bf05cbb81a55c5e45e3a96939e61f6efd0"}
+{"id": "demo_0716", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Self-grading and automated critique in Go.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "be168e7e68cd992bda3c031199a780770e3d7833"}
+{"id": "demo_0717", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Go.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "f4a8b6d07ae6c10d7d031fd3e71902189cfbece4"}
+{"id": "demo_0718", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Rust.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "8cd98daa7ed94eceb808aace257ad77a00c093c1"}
+{"id": "demo_0719", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Governance, auditability, and safety gates in TypeScript.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "4e706c8a193f8c3c233720689ef537626b4b4d05"}
+{"id": "demo_0720", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Governance, auditability, and safety gates in Rust.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "407f5e4eeddb9a55f35db2f65cf2cc7ffcb2996e"}
+{"id": "demo_0721", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Diff-first patching and SWE-bench style evaluation in Go.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "1ede0fbd7add911639c09c0d992869248542401b"}
+{"id": "demo_0722", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for MoE expert routing for code in Python.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "cdab494c7ea40443854dc9e76f81d9692cba5fad"}
+{"id": "demo_0723", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for MoE expert routing for code in Rust.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "2f0099d7d55722348ce0b94bc5863892091131f8"}
+{"id": "demo_0724", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Tool-calling traces and CI integration in Python.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "d867ab5b1b03795a650f259437a08f64961f7a87"}
+{"id": "demo_0725", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Self-grading and automated critique in Python.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "dabe0c58e324c9c8c2bdbfe80d61536a1063518d"}
+{"id": "demo_0726", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Tool-calling traces and CI integration in TypeScript.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "7a36c2ca2b689bcc99d6e92fcc5a9d64f4e84ece"}
+{"id": "demo_0727", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for MoE expert routing for code in TypeScript.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "5d8081fa4e2b4d2c429db2955e8f9dd90080d145"}
+{"id": "demo_0728", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Diff-first patching and SWE-bench style evaluation in Go.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "8e9b6a48836548d0f9202dc0384ed0d98e123f41"}
+{"id": "demo_0729", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Tool-calling traces and CI integration in Rust.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "bfa0ece85ebe9acb845d6bef2bf266c580695c9b"}
+{"id": "demo_0730", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Governance, auditability, and safety gates in Python.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "f2b051c203782bf1ac25353b2085f0ee12d6e31e"}
+{"id": "demo_0731", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Tool-calling traces and CI integration in Go.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "01ecda677e27b8918134d21a0d1cd85ef04617f9"}
+{"id": "demo_0732", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Governance, auditability, and safety gates in Go.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "b5dbf5c5ae7587a140ab91a3043227e4c81100fe"}
+{"id": "demo_0733", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Tool-calling traces and CI integration in TypeScript.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "b70cf5f1b56ffca136b21a28a0f03cb2993dfc03"}
+{"id": "demo_0734", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Self-grading and automated critique in Python.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "cf868e52e7b626f6e1445e520642918b3a2a1910"}
+{"id": "demo_0735", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Tool-calling traces and CI integration in Rust.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "6771342ab90a99fa5a829bbba8c24379d3ed4b2d"}
+{"id": "demo_0736", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Diff-first patching and SWE-bench style evaluation in Go.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "1ede0fbd7add911639c09c0d992869248542401b"}
+{"id": "demo_0737", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Python.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "b7325d57cdc539740421eea3ad320dc75e3a5809"}
+{"id": "demo_0738", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for MoE expert routing for code in TypeScript.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "c6a3c7c940db60ab9c0aad92f2280b426d27d44d"}
+{"id": "demo_0739", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Go.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "d165292165be3a4c543c7636d89fefd10b078e7c"}
+{"id": "demo_0740", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Rust.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "57884ed4c3d00bb495a4382eb0ea01a961e42e16"}
+{"id": "demo_0741", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Diff-first patching and SWE-bench style evaluation in Python.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "52094c608f86848d71b36f8b28490eeab830b40f"}
+{"id": "demo_0742", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for Tool-calling traces and CI integration in Go.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "b1f3c1daa457ede0e06e5e190d15d6e50635238b"}
+{"id": "demo_0743", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Python.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "1f8f524e0710190859cc6940196c24af4135327a"}
+{"id": "demo_0744", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Governance, auditability, and safety gates in Rust.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "d2ee9ea5723a120aeb73ff995a456cc5d8f41428"}
+{"id": "demo_0745", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Self-grading and automated critique in Go.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "be168e7e68cd992bda3c031199a780770e3d7833"}
+{"id": "demo_0746", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Governance, auditability, and safety gates in Python.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "e0062d33ddb8b86d135c9ab3dec2288f862109b1"}
+{"id": "demo_0747", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Diff-first patching and SWE-bench style evaluation in Rust.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "b1045ddb05b3c862b41b79b8eba5e07fecfd99b9"}
+{"id": "demo_0748", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Governance, auditability, and safety gates in Python.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "e0062d33ddb8b86d135c9ab3dec2288f862109b1"}
+{"id": "demo_0749", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Python.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "1f8f524e0710190859cc6940196c24af4135327a"}
+{"id": "demo_0750", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for Tool-calling traces and CI integration in Rust.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "256c808b21b4248d2f7c4c8c03d35d40bff23d5c"}
+{"id": "demo_0751", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Rust.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "0380156739222f10ebd520393e1597c16111cff1"}
+{"id": "demo_0752", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Tool-calling traces and CI integration in Python.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "3901f67d39afc0bfe5397c2c27f8ead5b2b472e7"}
+{"id": "demo_0753", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for MoE expert routing for code in Rust.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "2f0099d7d55722348ce0b94bc5863892091131f8"}
+{"id": "demo_0754", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Tool-calling traces and CI integration in TypeScript.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "7a36c2ca2b689bcc99d6e92fcc5a9d64f4e84ece"}
+{"id": "demo_0755", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Self-grading and automated critique in TypeScript.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "5ecb56d92db1e83dd4aeacdea1ea7ed21c892010"}
+{"id": "demo_0756", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Diff-first patching and SWE-bench style evaluation in Go.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "0a0e75aa64f2d9953fa6bcddadb57aa49fa969c6"}
+{"id": "demo_0757", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Governance, auditability, and safety gates in Go.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "b5dbf5c5ae7587a140ab91a3043227e4c81100fe"}
+{"id": "demo_0758", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Governance, auditability, and safety gates in Go.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "b5dbf5c5ae7587a140ab91a3043227e4c81100fe"}
+{"id": "demo_0759", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Tool-calling traces and CI integration in Go.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "da7d498e0aa830fb9d9079a323908d7d3e164681"}
+{"id": "demo_0760", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Diff-first patching and SWE-bench style evaluation in TypeScript.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "da67f5c566686221ddbda9c967c5c084e482c0de"}
+{"id": "demo_0761", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for MoE expert routing for code in Go.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "724a661b7365a7ceedaa2c6eb1ec2248b6dc7be0"}
+{"id": "demo_0762", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Governance, auditability, and safety gates in Python.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "2cd0a884e47168413d32f61dc500ce673c8e8ce6"}
+{"id": "demo_0763", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Diff-first patching and SWE-bench style evaluation in TypeScript.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "7105a1d4b1d96645cc6cc8bb9a9c50b8b512bf53"}
+{"id": "demo_0764", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Self-grading and automated critique in Python.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "54a082fe1d81e29f0620aed5d29e8a89b05a7741"}
+{"id": "demo_0765", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Tool-calling traces and CI integration in TypeScript.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "47ed1415b5f3bf9882a334b6c73ccdeb3ada9c95"}
+{"id": "demo_0766", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Tool-calling traces and CI integration in Python.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "d867ab5b1b03795a650f259437a08f64961f7a87"}
+{"id": "demo_0767", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for MoE expert routing for code in Python.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "cdab494c7ea40443854dc9e76f81d9692cba5fad"}
+{"id": "demo_0768", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Python.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "b7325d57cdc539740421eea3ad320dc75e3a5809"}
+{"id": "demo_0769", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Governance, auditability, and safety gates in Rust.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "57ef0242e37ea02626901d906d595b09bf7d9d85"}
+{"id": "demo_0770", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Rust.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "370dec51cc86af4647aa0881c7e9b9f3a6bbcfce"}
+{"id": "demo_0771", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Self-grading and automated critique in TypeScript.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "d02ca6f3f580c87df094e25ecb3de78fc576555b"}
+{"id": "demo_0772", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Self-grading and automated critique in TypeScript.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "d02ca6f3f580c87df094e25ecb3de78fc576555b"}
+{"id": "demo_0773", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Tool-calling traces and CI integration in Go.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "b1f3c1daa457ede0e06e5e190d15d6e50635238b"}
+{"id": "demo_0774", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for MoE expert routing for code in Rust.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "f230672f0116cff639eb4281d9ead7c3e278e6c9"}
+{"id": "demo_0775", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for MoE expert routing for code in TypeScript.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "5d8081fa4e2b4d2c429db2955e8f9dd90080d145"}
+{"id": "demo_0776", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Rust.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "0380156739222f10ebd520393e1597c16111cff1"}
+{"id": "demo_0777", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for MoE expert routing for code in Python.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "a7ccacbb4f991696e10c4df1b3a721ca21aaccfa"}
+{"id": "demo_0778", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Python.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "b7325d57cdc539740421eea3ad320dc75e3a5809"}
+{"id": "demo_0779", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for MoE expert routing for code in Rust.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "f230672f0116cff639eb4281d9ead7c3e278e6c9"}
+{"id": "demo_0780", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Go.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "532dd88780c2362d1ef34dea626ff108239e845b"}
+{"id": "demo_0781", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Go.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "532dd88780c2362d1ef34dea626ff108239e845b"}
+{"id": "demo_0782", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Rust.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "57884ed4c3d00bb495a4382eb0ea01a961e42e16"}
+{"id": "demo_0783", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Self-grading and automated critique in Python.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "dabe0c58e324c9c8c2bdbfe80d61536a1063518d"}
+{"id": "demo_0784", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for MoE expert routing for code in Go.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "441e2206b28951463f35934572603bd6fa2398d7"}
+{"id": "demo_0785", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Tool-calling traces and CI integration in TypeScript.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "b70cf5f1b56ffca136b21a28a0f03cb2993dfc03"}
+{"id": "demo_0786", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Diff-first patching and SWE-bench style evaluation in Go.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "1ede0fbd7add911639c09c0d992869248542401b"}
+{"id": "demo_0787", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in TypeScript.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "fab772571976b248bba6e42050c283d2dcacdc46"}
+{"id": "demo_0788", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Tool-calling traces and CI integration in Python.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "aa6c313c61194f09ceaa17d370a93b648759db00"}
+{"id": "demo_0789", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Self-grading and automated critique in Python.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "dabe0c58e324c9c8c2bdbfe80d61536a1063518d"}
+{"id": "demo_0790", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Governance, auditability, and safety gates in Rust.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "407f5e4eeddb9a55f35db2f65cf2cc7ffcb2996e"}
+{"id": "demo_0791", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Tool-calling traces and CI integration in Rust.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "6b80883d024afe8123d755ffa41e5b3cc0dcfb74"}
+{"id": "demo_0792", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Tool-calling traces and CI integration in Rust.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "6b80883d024afe8123d755ffa41e5b3cc0dcfb74"}
+{"id": "demo_0793", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Self-grading and automated critique in Python.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "3adfb9c8e53d1b7119af2ca07c587ce30d29dcb5"}
+{"id": "demo_0794", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Self-grading and automated critique in Python.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "3adfb9c8e53d1b7119af2ca07c587ce30d29dcb5"}
+{"id": "demo_0795", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Tool-calling traces and CI integration in Python.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "aa6c313c61194f09ceaa17d370a93b648759db00"}
+{"id": "demo_0796", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for MoE expert routing for code in Python.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "a7ccacbb4f991696e10c4df1b3a721ca21aaccfa"}
+{"id": "demo_0797", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Tool-calling traces and CI integration in Go.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "b1f3c1daa457ede0e06e5e190d15d6e50635238b"}
+{"id": "demo_0798", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Governance, auditability, and safety gates in Python.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "320b4391adf67a3dd800dad890aed898949b23da"}
+{"id": "demo_0799", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for MoE expert routing for code in Python.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "d431a0a2bc3c7de101a6f702ba1a56260d522625"}
+{"id": "demo_0800", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for MoE expert routing for code in Rust.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "a757555ae2d72f4f9a6c937f6ef6303650b71604"}
+{"id": "demo_0801", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Tool-calling traces and CI integration in TypeScript.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "7a36c2ca2b689bcc99d6e92fcc5a9d64f4e84ece"}
+{"id": "demo_0802", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Python.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "b7325d57cdc539740421eea3ad320dc75e3a5809"}
+{"id": "demo_0803", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for Tool-calling traces and CI integration in Python.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "3901f67d39afc0bfe5397c2c27f8ead5b2b472e7"}
+{"id": "demo_0804", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Rust.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "1a0fe8867460b6a6a80121de51265743a509981b"}
+{"id": "demo_0805", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Self-grading and automated critique in Python.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "cf868e52e7b626f6e1445e520642918b3a2a1910"}
+{"id": "demo_0806", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Governance, auditability, and safety gates in TypeScript.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "00ebe4210e2b4c255de24fd7a88ee68a0837ec64"}
+{"id": "demo_0807", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for MoE expert routing for code in Python.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "d431a0a2bc3c7de101a6f702ba1a56260d522625"}
+{"id": "demo_0808", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for MoE expert routing for code in TypeScript.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "84c33edd4bbd46f9b7ed6f724fa0c0c0891f9673"}
+{"id": "demo_0809", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Governance, auditability, and safety gates in Python.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "e0062d33ddb8b86d135c9ab3dec2288f862109b1"}
+{"id": "demo_0810", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Governance, auditability, and safety gates in Python.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "2cd0a884e47168413d32f61dc500ce673c8e8ce6"}
+{"id": "demo_0811", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Governance, auditability, and safety gates in Rust.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "d2ee9ea5723a120aeb73ff995a456cc5d8f41428"}
+{"id": "demo_0812", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Diff-first patching and SWE-bench style evaluation in Rust.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "6aef16a5ce69387a6f198bdb7d380162bb772ca2"}
+{"id": "demo_0813", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for MoE expert routing for code in Go.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "368f0d7a01cf671c54bc9d9be2e070f6ef61d3dc"}
+{"id": "demo_0814", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Self-grading and automated critique in Rust.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "0fcd85479910d25d7597ad10e547762ff04885d2"}
+{"id": "demo_0815", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Governance, auditability, and safety gates in Rust.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "29b4334a491410a737f41c7f169c9fcc2e01bf01"}
+{"id": "demo_0816", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Rust.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "57884ed4c3d00bb495a4382eb0ea01a961e42e16"}
+{"id": "demo_0817", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Self-grading and automated critique in Rust.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "d2815ca3d0e438822e466acc43474ba2874f5aad"}
+{"id": "demo_0818", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Diff-first patching and SWE-bench style evaluation in Rust.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "6aef16a5ce69387a6f198bdb7d380162bb772ca2"}
+{"id": "demo_0819", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Diff-first patching and SWE-bench style evaluation in Python.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "52094c608f86848d71b36f8b28490eeab830b40f"}
+{"id": "demo_0820", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for MoE expert routing for code in Python.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "a7ccacbb4f991696e10c4df1b3a721ca21aaccfa"}
+{"id": "demo_0821", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for MoE expert routing for code in TypeScript.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "5d8081fa4e2b4d2c429db2955e8f9dd90080d145"}
+{"id": "demo_0822", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Governance, auditability, and safety gates in Go.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "f58a9b02ebf152bc1c7f48fa9d5cda2b687e07c8"}
+{"id": "demo_0823", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Tool-calling traces and CI integration in Go.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "01ecda677e27b8918134d21a0d1cd85ef04617f9"}
+{"id": "demo_0824", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Python.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "b7325d57cdc539740421eea3ad320dc75e3a5809"}
+{"id": "demo_0825", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Tool-calling traces and CI integration in Python.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "b12d3c546b09fdd7d4abbaa5a72e2e41deabef1d"}
+{"id": "demo_0826", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Governance, auditability, and safety gates in Go.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "b5dbf5c5ae7587a140ab91a3043227e4c81100fe"}
+{"id": "demo_0827", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Self-grading and automated critique in TypeScript.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "d02ca6f3f580c87df094e25ecb3de78fc576555b"}
+{"id": "demo_0828", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for MoE expert routing for code in Go.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "368f0d7a01cf671c54bc9d9be2e070f6ef61d3dc"}
+{"id": "demo_0829", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for MoE expert routing for code in TypeScript.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "04cf27cd6c7cc60a6ad34383c06c00220df1d7d6"}
+{"id": "demo_0830", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Tool-calling traces and CI integration in TypeScript.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "7a36c2ca2b689bcc99d6e92fcc5a9d64f4e84ece"}
+{"id": "demo_0831", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Tool-calling traces and CI integration in Rust.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "2ad46a119bbbd13e9f59ca46d56fc6ceffd8c0ad"}
+{"id": "demo_0832", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in TypeScript.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "97e3218662e2f00746256643faef6885ea18cd7b"}
+{"id": "demo_0833", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for MoE expert routing for code in Rust.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "e7dcb2a63f3bf8de88060e0b2a7a5c6f548c4deb"}
+{"id": "demo_0834", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Tool-calling traces and CI integration in Go.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "01ecda677e27b8918134d21a0d1cd85ef04617f9"}
+{"id": "demo_0835", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Governance, auditability, and safety gates in Rust.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "d2ee9ea5723a120aeb73ff995a456cc5d8f41428"}
+{"id": "demo_0836", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Diff-first patching and SWE-bench style evaluation in Go.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "0a0e75aa64f2d9953fa6bcddadb57aa49fa969c6"}
+{"id": "demo_0837", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Self-grading and automated critique in Go.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "6ddfb66b8ab85ff28b1c9d587350b2c6c891a905"}
+{"id": "demo_0838", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Tool-calling traces and CI integration in Rust.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "6b80883d024afe8123d755ffa41e5b3cc0dcfb74"}
+{"id": "demo_0839", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for Diff-first patching and SWE-bench style evaluation in Rust.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "b1045ddb05b3c862b41b79b8eba5e07fecfd99b9"}
+{"id": "demo_0840", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Tool-calling traces and CI integration in Python.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "b12d3c546b09fdd7d4abbaa5a72e2e41deabef1d"}
+{"id": "demo_0841", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Rust.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "57884ed4c3d00bb495a4382eb0ea01a961e42e16"}
+{"id": "demo_0842", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Diff-first patching and SWE-bench style evaluation in Go.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "0a0e75aa64f2d9953fa6bcddadb57aa49fa969c6"}
+{"id": "demo_0843", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Diff-first patching and SWE-bench style evaluation in TypeScript.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "1971ecc0a5017fee37a2751e068efa0ba9bdb6b7"}
+{"id": "demo_0844", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Self-grading and automated critique in Python.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "dabe0c58e324c9c8c2bdbfe80d61536a1063518d"}
+{"id": "demo_0845", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Rust.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "1a0fe8867460b6a6a80121de51265743a509981b"}
+{"id": "demo_0846", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Diff-first patching and SWE-bench style evaluation in Python.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "399b590cd03414ba736ffb6888734b51cacb8ca9"}
+{"id": "demo_0847", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Go.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "ff4427b1aeb8acedadc80320a21a8044fc27397e"}
+{"id": "demo_0848", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for MoE expert routing for code in Go.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "55f016e0f5b3b2fae7261a93255883cdb7e2e829"}
+{"id": "demo_0849", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for MoE expert routing for code in Go.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "55f016e0f5b3b2fae7261a93255883cdb7e2e829"}
+{"id": "demo_0850", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for MoE expert routing for code in TypeScript.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "c6a3c7c940db60ab9c0aad92f2280b426d27d44d"}
+{"id": "demo_0851", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for MoE expert routing for code in Python.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "d10ebbd0771e561742de9157381a089d64c68539"}
+{"id": "demo_0852", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Diff-first patching and SWE-bench style evaluation in Python.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "ed6687b23f76fa0808d028f46af13a368a5e5155"}
+{"id": "demo_0853", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Diff-first patching and SWE-bench style evaluation in Python.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "d4e906daed3e71a92b9936a54b063c36f18f0217"}
+{"id": "demo_0854", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Diff-first patching and SWE-bench style evaluation in Rust.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "6aef16a5ce69387a6f198bdb7d380162bb772ca2"}
+{"id": "demo_0855", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Self-grading and automated critique in Go.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "6ddfb66b8ab85ff28b1c9d587350b2c6c891a905"}
+{"id": "demo_0856", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Rust.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "57884ed4c3d00bb495a4382eb0ea01a961e42e16"}
+{"id": "demo_0857", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Self-grading and automated critique in TypeScript.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "5ecb56d92db1e83dd4aeacdea1ea7ed21c892010"}
+{"id": "demo_0858", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for Governance, auditability, and safety gates in Go.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "f58a9b02ebf152bc1c7f48fa9d5cda2b687e07c8"}
+{"id": "demo_0859", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for MoE expert routing for code in TypeScript.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "c6a3c7c940db60ab9c0aad92f2280b426d27d44d"}
+{"id": "demo_0860", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Diff-first patching and SWE-bench style evaluation in Go.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "0a0e75aa64f2d9953fa6bcddadb57aa49fa969c6"}
+{"id": "demo_0861", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Governance, auditability, and safety gates in Python.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "2cd0a884e47168413d32f61dc500ce673c8e8ce6"}
+{"id": "demo_0862", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for Diff-first patching and SWE-bench style evaluation in Go.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "844412bf05cbb81a55c5e45e3a96939e61f6efd0"}
+{"id": "demo_0863", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for Governance, auditability, and safety gates in Rust.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "57ef0242e37ea02626901d906d595b09bf7d9d85"}
+{"id": "demo_0864", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Tool-calling traces and CI integration in Python.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "d867ab5b1b03795a650f259437a08f64961f7a87"}
+{"id": "demo_0865", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Governance, auditability, and safety gates in Go.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "f58a9b02ebf152bc1c7f48fa9d5cda2b687e07c8"}
+{"id": "demo_0866", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Self-grading and automated critique in Python.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "3adfb9c8e53d1b7119af2ca07c587ce30d29dcb5"}
+{"id": "demo_0867", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for Diff-first patching and SWE-bench style evaluation in Go.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "844412bf05cbb81a55c5e45e3a96939e61f6efd0"}
+{"id": "demo_0868", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Tool-calling traces and CI integration in TypeScript.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "47ed1415b5f3bf9882a334b6c73ccdeb3ada9c95"}
+{"id": "demo_0869", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Tool-calling traces and CI integration in TypeScript.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "2fd96b170ecf4c762b8fa01b03e48f8dc581d55a"}
+{"id": "demo_0870", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for MoE expert routing for code in Go.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "368f0d7a01cf671c54bc9d9be2e070f6ef61d3dc"}
+{"id": "demo_0871", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Tool-calling traces and CI integration in Python.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "3901f67d39afc0bfe5397c2c27f8ead5b2b472e7"}
+{"id": "demo_0872", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Diff-first patching and SWE-bench style evaluation in TypeScript.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "da67f5c566686221ddbda9c967c5c084e482c0de"}
+{"id": "demo_0873", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for MoE expert routing for code in Rust.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "2f0099d7d55722348ce0b94bc5863892091131f8"}
+{"id": "demo_0874", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Self-grading and automated critique in Rust.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "a43c2fa4a3ea9a8fc5ec43f2a91fd9d185e88885"}
+{"id": "demo_0875", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Diff-first patching and SWE-bench style evaluation in Rust.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "6aef16a5ce69387a6f198bdb7d380162bb772ca2"}
+{"id": "demo_0876", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for MoE expert routing for code in Go.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "724a661b7365a7ceedaa2c6eb1ec2248b6dc7be0"}
+{"id": "demo_0877", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for MoE expert routing for code in Go.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "724a661b7365a7ceedaa2c6eb1ec2248b6dc7be0"}
+{"id": "demo_0878", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for MoE expert routing for code in Rust.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "f230672f0116cff639eb4281d9ead7c3e278e6c9"}
+{"id": "demo_0879", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Self-grading and automated critique in TypeScript.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "f31708cf2dc2bf12377b51649cb192e41c358dac"}
+{"id": "demo_0880", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Self-grading and automated critique in TypeScript.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "f31708cf2dc2bf12377b51649cb192e41c358dac"}
+{"id": "demo_0881", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for Self-grading and automated critique in Python.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "fed8c1215ad6b2320879502cbbc51dc2e7766e7f"}
+{"id": "demo_0882", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Governance, auditability, and safety gates in Go.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "03e3880ec5391fcf3ecef131ad5aa0e0b53ad63e"}
+{"id": "demo_0883", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Diff-first patching and SWE-bench style evaluation in Python.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "ed6687b23f76fa0808d028f46af13a368a5e5155"}
+{"id": "demo_0884", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Governance, auditability, and safety gates in Rust.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "66b4ca1544895342a9abd1a6c3586898a0140df4"}
+{"id": "demo_0885", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for MoE expert routing for code in Go.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "724a661b7365a7ceedaa2c6eb1ec2248b6dc7be0"}
+{"id": "demo_0886", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Self-grading and automated critique in TypeScript.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "4fc6dade0f8d7671f7cbe49b182676215b21cf70"}
+{"id": "demo_0887", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Diff-first patching and SWE-bench style evaluation in TypeScript.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "da67f5c566686221ddbda9c967c5c084e482c0de"}
+{"id": "demo_0888", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Diff-first patching and SWE-bench style evaluation in Rust.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "6aef16a5ce69387a6f198bdb7d380162bb772ca2"}
+{"id": "demo_0889", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Go.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "532dd88780c2362d1ef34dea626ff108239e845b"}
+{"id": "demo_0890", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Tool-calling traces and CI integration in Rust.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "6771342ab90a99fa5a829bbba8c24379d3ed4b2d"}
+{"id": "demo_0891", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for MoE expert routing for code in TypeScript.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "84c33edd4bbd46f9b7ed6f724fa0c0c0891f9673"}
+{"id": "demo_0892", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for MoE expert routing for code in Rust.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "6a015cfda615e97b9a9735cff067ff27b811115e"}
+{"id": "demo_0893", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for MoE expert routing for code in Go.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "368f0d7a01cf671c54bc9d9be2e070f6ef61d3dc"}
+{"id": "demo_0894", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Tool-calling traces and CI integration in Rust.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "6771342ab90a99fa5a829bbba8c24379d3ed4b2d"}
+{"id": "demo_0895", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Tool-calling traces and CI integration in TypeScript.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "b70cf5f1b56ffca136b21a28a0f03cb2993dfc03"}
+{"id": "demo_0896", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Governance, auditability, and safety gates in Go.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "03e3880ec5391fcf3ecef131ad5aa0e0b53ad63e"}
+{"id": "demo_0897", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Self-grading and automated critique in Rust.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "a43c2fa4a3ea9a8fc5ec43f2a91fd9d185e88885"}
+{"id": "demo_0898", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in TypeScript.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "97e3218662e2f00746256643faef6885ea18cd7b"}
+{"id": "demo_0899", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Self-grading and automated critique in Python.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "3adfb9c8e53d1b7119af2ca07c587ce30d29dcb5"}
+{"id": "demo_0900", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Python.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "cabf427b92e83fe95318ff3a5c2d46bd31d67e6e"}
+{"id": "demo_0901", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Rust.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "370dec51cc86af4647aa0881c7e9b9f3a6bbcfce"}
+{"id": "demo_0902", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for MoE expert routing for code in Go.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "c77684ed94a4148f8449f9b073cd70d3bc577a1d"}
+{"id": "demo_0903", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for MoE expert routing for code in Python.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "a7ccacbb4f991696e10c4df1b3a721ca21aaccfa"}
+{"id": "demo_0904", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Governance, auditability, and safety gates in Rust.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "57ef0242e37ea02626901d906d595b09bf7d9d85"}
+{"id": "demo_0905", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Tool-calling traces and CI integration in Go.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "da7d498e0aa830fb9d9079a323908d7d3e164681"}
+{"id": "demo_0906", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Tool-calling traces and CI integration in TypeScript.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "f3b404c1695427e8871ad8a5906b0fc4c088966a"}
+{"id": "demo_0907", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for MoE expert routing for code in Rust.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "e7dcb2a63f3bf8de88060e0b2a7a5c6f548c4deb"}
+{"id": "demo_0908", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Tool-calling traces and CI integration in Rust.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "2ad46a119bbbd13e9f59ca46d56fc6ceffd8c0ad"}
+{"id": "demo_0909", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Self-grading and automated critique in TypeScript.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "d02ca6f3f580c87df094e25ecb3de78fc576555b"}
+{"id": "demo_0910", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for Governance, auditability, and safety gates in Go.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "f58a9b02ebf152bc1c7f48fa9d5cda2b687e07c8"}
+{"id": "demo_0911", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for MoE expert routing for code in Go.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "724a661b7365a7ceedaa2c6eb1ec2248b6dc7be0"}
+{"id": "demo_0912", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Self-grading and automated critique in Python.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "dabe0c58e324c9c8c2bdbfe80d61536a1063518d"}
+{"id": "demo_0913", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Diff-first patching and SWE-bench style evaluation in Python.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "399b590cd03414ba736ffb6888734b51cacb8ca9"}
+{"id": "demo_0914", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Diff-first patching and SWE-bench style evaluation in Rust.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "882626e1c64172100c8201be60a636d13fc4cd1d"}
+{"id": "demo_0915", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Tool-calling traces and CI integration in Go.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "3b6175b68f17af9ab68a3e4275a82a74736d9b6d"}
+{"id": "demo_0916", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Self-grading and automated critique in TypeScript.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "5ecb56d92db1e83dd4aeacdea1ea7ed21c892010"}
+{"id": "demo_0917", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Self-grading and automated critique in TypeScript.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "4fc6dade0f8d7671f7cbe49b182676215b21cf70"}
+{"id": "demo_0918", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Diff-first patching and SWE-bench style evaluation in Rust.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "b1045ddb05b3c862b41b79b8eba5e07fecfd99b9"}
+{"id": "demo_0919", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Diff-first patching and SWE-bench style evaluation in Go.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "0a0e75aa64f2d9953fa6bcddadb57aa49fa969c6"}
+{"id": "demo_0920", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for Tool-calling traces and CI integration in TypeScript.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "2fd96b170ecf4c762b8fa01b03e48f8dc581d55a"}
+{"id": "demo_0921", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Governance, auditability, and safety gates in Rust.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "d2ee9ea5723a120aeb73ff995a456cc5d8f41428"}
+{"id": "demo_0922", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in TypeScript.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "97e3218662e2f00746256643faef6885ea18cd7b"}
+{"id": "demo_0923", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Governance, auditability, and safety gates in Rust.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "d2ee9ea5723a120aeb73ff995a456cc5d8f41428"}
+{"id": "demo_0924", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Self-grading and automated critique in Go.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "be168e7e68cd992bda3c031199a780770e3d7833"}
+{"id": "demo_0925", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Governance, auditability, and safety gates in Rust.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "407f5e4eeddb9a55f35db2f65cf2cc7ffcb2996e"}
+{"id": "demo_0926", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Governance, auditability, and safety gates in Go.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "03e3880ec5391fcf3ecef131ad5aa0e0b53ad63e"}
+{"id": "demo_0927", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Self-grading and automated critique in Go.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "c012733817f6f8ceeff47e5dc6c8bdcfb44433fe"}
+{"id": "demo_0928", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Diff-first patching and SWE-bench style evaluation in TypeScript.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "7105a1d4b1d96645cc6cc8bb9a9c50b8b512bf53"}
+{"id": "demo_0929", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Governance, auditability, and safety gates in TypeScript.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "e55625866d9d72afa0ca219100fd584e58fd1a20"}
+{"id": "demo_0930", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for Diff-first patching and SWE-bench style evaluation in Python.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "399b590cd03414ba736ffb6888734b51cacb8ca9"}
+{"id": "demo_0931", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for Governance, auditability, and safety gates in Go.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "f58a9b02ebf152bc1c7f48fa9d5cda2b687e07c8"}
+{"id": "demo_0932", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for MoE expert routing for code in Rust.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "6a015cfda615e97b9a9735cff067ff27b811115e"}
+{"id": "demo_0933", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in TypeScript.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "bb6f07fec726ce3ac5b3ef6b3d61c76abe83fd3d"}
+{"id": "demo_0934", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Rust.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "0380156739222f10ebd520393e1597c16111cff1"}
+{"id": "demo_0935", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Tool-calling traces and CI integration in TypeScript.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "f3b404c1695427e8871ad8a5906b0fc4c088966a"}
+{"id": "demo_0936", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Tool-calling traces and CI integration in Go.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "01ecda677e27b8918134d21a0d1cd85ef04617f9"}
+{"id": "demo_0937", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Diff-first patching and SWE-bench style evaluation in Go.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "0a0e75aa64f2d9953fa6bcddadb57aa49fa969c6"}
+{"id": "demo_0938", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Tool-calling traces and CI integration in Go.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "da7d498e0aa830fb9d9079a323908d7d3e164681"}
+{"id": "demo_0939", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Go.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "ff4427b1aeb8acedadc80320a21a8044fc27397e"}
+{"id": "demo_0940", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Tool-calling traces and CI integration in Go.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "da7d498e0aa830fb9d9079a323908d7d3e164681"}
+{"id": "demo_0941", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for MoE expert routing for code in Rust.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "2f0099d7d55722348ce0b94bc5863892091131f8"}
+{"id": "demo_0942", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Tool-calling traces and CI integration in Go.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "01ecda677e27b8918134d21a0d1cd85ef04617f9"}
+{"id": "demo_0943", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Rust.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "57884ed4c3d00bb495a4382eb0ea01a961e42e16"}
+{"id": "demo_0944", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for MoE expert routing for code in Go.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "724a661b7365a7ceedaa2c6eb1ec2248b6dc7be0"}
+{"id": "demo_0945", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Rust.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "1a0fe8867460b6a6a80121de51265743a509981b"}
+{"id": "demo_0946", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Tool-calling traces and CI integration in Python.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "aa6c313c61194f09ceaa17d370a93b648759db00"}
+{"id": "demo_0947", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for MoE expert routing for code in Rust.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "a757555ae2d72f4f9a6c937f6ef6303650b71604"}
+{"id": "demo_0948", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for MoE expert routing for code in TypeScript.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "a4322f9593d84f6099aa35b3061f2aa8a7adab8e"}
+{"id": "demo_0949", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Diff-first patching and SWE-bench style evaluation in Python.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "ed6687b23f76fa0808d028f46af13a368a5e5155"}
+{"id": "demo_0950", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for Diff-first patching and SWE-bench style evaluation in Python.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "399b590cd03414ba736ffb6888734b51cacb8ca9"}
+{"id": "demo_0951", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Rust.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "8cd98daa7ed94eceb808aace257ad77a00c093c1"}
+{"id": "demo_0952", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Diff-first patching and SWE-bench style evaluation in Go.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "9ac511806d0fd80846c9ee66e870bc422000ca5d"}
+{"id": "demo_0953", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Self-grading and automated critique in TypeScript.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "d02ca6f3f580c87df094e25ecb3de78fc576555b"}
+{"id": "demo_0954", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for MoE expert routing for code in Rust.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "6a015cfda615e97b9a9735cff067ff27b811115e"}
+{"id": "demo_0955", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for MoE expert routing for code in Rust.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "e7dcb2a63f3bf8de88060e0b2a7a5c6f548c4deb"}
+{"id": "demo_0956", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Governance, auditability, and safety gates in Rust.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "d2ee9ea5723a120aeb73ff995a456cc5d8f41428"}
+{"id": "demo_0957", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for MoE expert routing for code in Go.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "55f016e0f5b3b2fae7261a93255883cdb7e2e829"}
+{"id": "demo_0958", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Go.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "f4a8b6d07ae6c10d7d031fd3e71902189cfbece4"}
+{"id": "demo_0959", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for MoE expert routing for code in TypeScript.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "5d8081fa4e2b4d2c429db2955e8f9dd90080d145"}
+{"id": "demo_0960", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Tool-calling traces and CI integration in TypeScript.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "47ed1415b5f3bf9882a334b6c73ccdeb3ada9c95"}
+{"id": "demo_0961", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Governance, auditability, and safety gates in Rust.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "407f5e4eeddb9a55f35db2f65cf2cc7ffcb2996e"}
+{"id": "demo_0962", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in TypeScript.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "fab772571976b248bba6e42050c283d2dcacdc46"}
+{"id": "demo_0963", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for MoE expert routing for code in Rust.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "f230672f0116cff639eb4281d9ead7c3e278e6c9"}
+{"id": "demo_0964", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for MoE expert routing for code in Python.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "fd55fdc9ecf12df2b0b5a36e7f1f0f93d303dbb3"}
+{"id": "demo_0965", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Tool-calling traces and CI integration in Go.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "3b6175b68f17af9ab68a3e4275a82a74736d9b6d"}
+{"id": "demo_0966", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Diff-first patching and SWE-bench style evaluation in Rust.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "8dacf6e08ac4a42cf38a0c147636cfdd66f96f97"}
+{"id": "demo_0967", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Tool-calling traces and CI integration in Go.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "01ecda677e27b8918134d21a0d1cd85ef04617f9"}
+{"id": "demo_0968", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Diff-first patching and SWE-bench style evaluation in Rust.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "8dacf6e08ac4a42cf38a0c147636cfdd66f96f97"}
+{"id": "demo_0969", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Go.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "f4a8b6d07ae6c10d7d031fd3e71902189cfbece4"}
+{"id": "demo_0970", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Go.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "f4a8b6d07ae6c10d7d031fd3e71902189cfbece4"}
+{"id": "demo_0971", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Self-grading and automated critique in Rust.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "d2815ca3d0e438822e466acc43474ba2874f5aad"}
+{"id": "demo_0972", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for Tool-calling traces and CI integration in Rust.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "6b80883d024afe8123d755ffa41e5b3cc0dcfb74"}
+{"id": "demo_0973", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "eval", "difficulty": "advanced", "instruction": "Demonstrate eval for Tool-calling traces and CI integration in TypeScript.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "7a36c2ca2b689bcc99d6e92fcc5a9d64f4e84ece"}
+{"id": "demo_0974", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Diff-first patching and SWE-bench style evaluation in Rust.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "6aef16a5ce69387a6f198bdb7d380162bb772ca2"}
+{"id": "demo_0975", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Diff-first patching and SWE-bench style evaluation in TypeScript.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "7b045c7e00592cb0936703c6efbe07d24fed1640"}
+{"id": "demo_0976", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Self-grading and automated critique in Go.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "85d5a5bc6f30d51ffc33cee542d2c5f2d1d15657"}
+{"id": "demo_0977", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Demonstrate patch_diff for Self-grading and automated critique in Python.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "dabe0c58e324c9c8c2bdbfe80d61536a1063518d"}
+{"id": "demo_0978", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for MoE expert routing for code in Go.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "55f016e0f5b3b2fae7261a93255883cdb7e2e829"}
+{"id": "demo_0979", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Self-grading and automated critique in Python.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "cf868e52e7b626f6e1445e520642918b3a2a1910"}
+{"id": "demo_0980", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Demonstrate agent_loop for Diff-first patching and SWE-bench style evaluation in Go.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "1ede0fbd7add911639c09c0d992869248542401b"}
+{"id": "demo_0981", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Demonstrate tool_trace for MoE expert routing for code in Rust.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "e7dcb2a63f3bf8de88060e0b2a7a5c6f548c4deb"}
+{"id": "demo_0982", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Self-grading and automated critique in Python.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "54a082fe1d81e29f0620aed5d29e8a89b05a7741"}
+{"id": "demo_0983", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Go.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "532dd88780c2362d1ef34dea626ff108239e845b"}
+{"id": "demo_0984", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Self-grading and automated critique in Python.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "54a082fe1d81e29f0620aed5d29e8a89b05a7741"}
+{"id": "demo_0985", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Self-grading and automated critique in Python.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "54a082fe1d81e29f0620aed5d29e8a89b05a7741"}
+{"id": "demo_0986", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Diff-first patching and SWE-bench style evaluation in Python.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "ed6687b23f76fa0808d028f46af13a368a5e5155"}
+{"id": "demo_0987", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Diff-first patching and SWE-bench style evaluation in Go.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "9ac511806d0fd80846c9ee66e870bc422000ca5d"}
+{"id": "demo_0988", "created": "2026-01-01", "topic": "Self-grading and automated critique", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Self-grading and automated critique in Python.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "cf868e52e7b626f6e1445e520642918b3a2a1910"}
+{"id": "demo_0989", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Rust.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "0380156739222f10ebd520393e1597c16111cff1"}
+{"id": "demo_0990", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Demonstrate patch_diff for MoE expert routing for code in Go.", "input": "", "output": "```diff\n- return a/b\n+ if b==0: raise ValueError('division by zero')\n+ return a/b\n```", "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "368f0d7a01cf671c54bc9d9be2e070f6ef61d3dc"}
+{"id": "demo_0991", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for Diff-first patching and SWE-bench style evaluation in Rust.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "b1045ddb05b3c862b41b79b8eba5e07fecfd99b9"}
+{"id": "demo_0992", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Tool-calling traces and CI integration in Go.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Go", "frontier_demo": true}, "hash": "01ecda677e27b8918134d21a0d1cd85ef04617f9"}
+{"id": "demo_0993", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Python.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "cabf427b92e83fe95318ff3a5c2d46bd31d67e6e"}
+{"id": "demo_0994", "created": "2026-01-01", "topic": "MoE expert routing for code", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for MoE expert routing for code in Python.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "d10ebbd0771e561742de9157381a089d64c68539"}
+{"id": "demo_0995", "created": "2026-01-01", "topic": "Diff-first patching and SWE-bench style evaluation", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Demonstrate tool_trace for Diff-first patching and SWE-bench style evaluation in TypeScript.", "input": "", "output": [{"tool": "checkout", "args": {"commit": "abc123"}}, {"tool": "run_tests", "args": {"suite": "unit"}}, {"tool": "apply_patch", "args": {"files": 1}}, {"tool": "run_tests", "args": {"suite": "full"}}], "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "1971ecc0a5017fee37a2751e068efa0ba9bdb6b7"}
+{"id": "demo_0996", "created": "2026-01-01", "topic": "Tool-calling traces and CI integration", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Demonstrate agent_loop for Tool-calling traces and CI integration in Rust.", "input": "", "output": "Plan \u2192 Edit \u2192 Test \u2192 Reflect \u2192 Human Gate", "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "bfa0ece85ebe9acb845d6bef2bf266c580695c9b"}
+{"id": "demo_0997", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "expert", "instruction": "Demonstrate eval for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Python.", "input": "", "output": "Metrics: pass@k, time-to-green, cost-per-fix, regressions", "metadata": {"brand": "Within Us AI", "language": "Python", "frontier_demo": true}, "hash": "ce6f758b1bc428101ae12ad43bff7a4e88c0cfc7"}
+{"id": "demo_0998", "created": "2026-01-01", "topic": "Governance, auditability, and safety gates", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Demonstrate self_grade for Governance, auditability, and safety gates in TypeScript.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "TypeScript", "frontier_demo": true}, "hash": "fe883ef5866a8ab03c83ebb26b3251f8679aca88"}
+{"id": "demo_0999", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "self_grade", "difficulty": "expert", "instruction": "Demonstrate self_grade for Agentic coding systems (plan\u2192edit\u2192test\u2192reflect) in Rust.", "input": "", "output": {"score": 9, "confidence": 0.9, "notes": "Tests passed; minimal diff; safe."}, "metadata": {"brand": "Within Us AI", "language": "Rust", "frontier_demo": true}, "hash": "0380156739222f10ebd520393e1597c16111cff1"}