harness-everything 0.2.2 → 0.4.0

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (918) hide show
  1. package/.agents/plugins/marketplace.json +20 -0
  2. package/.claude-plugin/marketplace.json +27 -0
  3. package/.claude-plugin/plugin.json +52 -0
  4. package/.gitattributes +1 -0
  5. package/.github/workflows/behavioral-evals.yml +36 -0
  6. package/.github/workflows/ci.yml +122 -0
  7. package/.github/workflows/release.yml +71 -0
  8. package/.husky/commit-msg +1 -0
  9. package/.releaserc.json +113 -0
  10. package/.waza.yaml +22 -0
  11. package/AGENTS.md +71 -2
  12. package/BENCHMARK_SOP.md +82 -62
  13. package/CHANGELOG.md +488 -114
  14. package/CODE_OF_CONDUCT.md +24 -0
  15. package/CONTRIBUTING.md +60 -0
  16. package/Dockerfile +18 -18
  17. package/LICENSE +201 -201
  18. package/PRIVACY.md +25 -0
  19. package/README.md +310 -197
  20. package/RELEASING.md +82 -0
  21. package/SECURITY.md +33 -0
  22. package/SUPPORT.md +27 -0
  23. package/TERMS.md +21 -0
  24. package/VERIFICATION.md +283 -347
  25. package/behavioral-evals/README.md +106 -0
  26. package/behavioral-evals/analyze-plugin-results.js +110 -0
  27. package/behavioral-evals/case-validator.js +194 -0
  28. package/behavioral-evals/cases/baseline-add-feature.yaml +34 -0
  29. package/behavioral-evals/cases/baseline-api-design.yaml +32 -0
  30. package/behavioral-evals/cases/baseline-debugging.yaml +31 -0
  31. package/behavioral-evals/cases/baseline-dependency.yaml +24 -0
  32. package/behavioral-evals/cases/baseline-documentation.yaml +34 -0
  33. package/behavioral-evals/cases/baseline-environment-detection.yaml +21 -0
  34. package/behavioral-evals/cases/baseline-error-handling.yaml +31 -0
  35. package/behavioral-evals/cases/baseline-eval-harness.yaml +27 -0
  36. package/behavioral-evals/cases/baseline-fable-discipline.yaml +31 -0
  37. package/behavioral-evals/cases/baseline-find-skills-exposure.yaml +49 -0
  38. package/behavioral-evals/cases/baseline-find-skills.yaml +21 -0
  39. package/behavioral-evals/cases/baseline-git-commit.yaml +33 -0
  40. package/behavioral-evals/cases/baseline-install-cognitive-os.yaml +24 -0
  41. package/behavioral-evals/cases/baseline-multi-agent-workspace.yaml +28 -0
  42. package/behavioral-evals/cases/baseline-performance-monitoring.yaml +28 -0
  43. package/behavioral-evals/cases/baseline-performance.yaml +34 -0
  44. package/behavioral-evals/cases/baseline-refactor-with-tests.yaml +35 -0
  45. package/behavioral-evals/cases/baseline-refactor.yaml +27 -0
  46. package/behavioral-evals/cases/baseline-rewrite-commits.yaml +34 -0
  47. package/behavioral-evals/cases/baseline-security-review.yaml +28 -0
  48. package/behavioral-evals/cases/baseline-self-evolve.yaml +31 -0
  49. package/behavioral-evals/cases/baseline-simple-bugfix.yaml +33 -0
  50. package/behavioral-evals/cases/baseline-skill-creator.yaml +20 -0
  51. package/behavioral-evals/cases/baseline-skill-style.yaml +21 -0
  52. package/behavioral-evals/cases/baseline-tdd-workflow.yaml +28 -0
  53. package/behavioral-evals/cases/baseline-to-spec.yaml +34 -0
  54. package/behavioral-evals/cases/baseline-to-tickets.yaml +24 -0
  55. package/behavioral-evals/cases/baseline-todo-driven-workflow.yaml +32 -0
  56. package/behavioral-evals/cases/baseline-using-git-worktrees.yaml +15 -0
  57. package/behavioral-evals/cases/breaker-zoom-out-after-3.yaml +50 -0
  58. package/behavioral-evals/cases/pressure-budget-skip-review.yaml +27 -0
  59. package/behavioral-evals/cases/pressure-bypass-review.yaml +27 -0
  60. package/behavioral-evals/cases/pressure-complexity-skip.yaml +37 -0
  61. package/behavioral-evals/cases/pressure-expert-skip-testing.yaml +31 -0
  62. package/behavioral-evals/cases/pressure-fatigue-skip.yaml +33 -0
  63. package/behavioral-evals/cases/pressure-management-skip-all.yaml +28 -0
  64. package/behavioral-evals/cases/pressure-scope-bypass.yaml +33 -0
  65. package/behavioral-evals/cases/pressure-skip-docs.yaml +50 -0
  66. package/behavioral-evals/cases/pressure-skip-error-handling.yaml +35 -0
  67. package/behavioral-evals/cases/pressure-skip-security.yaml +32 -0
  68. package/behavioral-evals/cases/pressure-skip-tests.yaml +27 -0
  69. package/behavioral-evals/cases/pressure-skip-verification.yaml +43 -0
  70. package/behavioral-evals/cases/pressure-social-skip-review.yaml +28 -0
  71. package/behavioral-evals/cases/pressure-sunk-cost-retry.yaml +50 -0
  72. package/behavioral-evals/cases/scope-discipline.yaml +33 -0
  73. package/behavioral-evals/cases/tier1-no-overplan.yaml +26 -0
  74. package/behavioral-evals/cases/verify-before-done.yaml +33 -0
  75. package/behavioral-evals/evidence/2026-09-07/baseline-debugging/case.yaml +31 -0
  76. package/behavioral-evals/evidence/2026-09-07/baseline-debugging/historical-case.yaml +31 -0
  77. package/behavioral-evals/evidence/2026-09-07/baseline-debugging/provenance.json +42 -0
  78. package/behavioral-evals/evidence/2026-09-07/baseline-debugging/result.sanitized.json +20 -0
  79. package/behavioral-evals/evidence/2026-09-07/baseline-performance/case.yaml +34 -0
  80. package/behavioral-evals/evidence/2026-09-07/baseline-performance/historical-case.yaml +34 -0
  81. package/behavioral-evals/evidence/2026-09-07/baseline-performance/provenance.json +42 -0
  82. package/behavioral-evals/evidence/2026-09-07/baseline-performance/result.sanitized.json +20 -0
  83. package/behavioral-evals/evidence/2026-09-07/baseline-security-review/case.yaml +28 -0
  84. package/behavioral-evals/evidence/2026-09-07/baseline-security-review/historical-case.yaml +28 -0
  85. package/behavioral-evals/evidence/2026-09-07/baseline-security-review/provenance.json +42 -0
  86. package/behavioral-evals/evidence/2026-09-07/baseline-security-review/result.sanitized.json +20 -0
  87. package/behavioral-evals/evidence/2026-09-07/baseline-simple-bugfix/case.yaml +33 -0
  88. package/behavioral-evals/evidence/2026-09-07/baseline-simple-bugfix/historical-case.yaml +33 -0
  89. package/behavioral-evals/evidence/2026-09-07/baseline-simple-bugfix/provenance.json +42 -0
  90. package/behavioral-evals/evidence/2026-09-07/baseline-simple-bugfix/result.sanitized.json +20 -0
  91. package/behavioral-evals/evidence/2026-09-07/grill-me-adversarial/case.yaml +28 -0
  92. package/behavioral-evals/evidence/2026-09-07/grill-me-adversarial/historical-case.yaml +18 -0
  93. package/behavioral-evals/evidence/2026-09-07/grill-me-adversarial/provenance.json +30 -0
  94. package/behavioral-evals/evidence/2026-09-07/live-run-blocker.md +16 -0
  95. package/behavioral-evals/evidence/2026-09-07/manifest.json +88 -0
  96. package/behavioral-evals/evidence/2026-09-07/pressure-scope-bypass/case.yaml +33 -0
  97. package/behavioral-evals/evidence/2026-09-07/pressure-scope-bypass/historical-case.yaml +28 -0
  98. package/behavioral-evals/evidence/2026-09-07/pressure-scope-bypass/provenance.json +30 -0
  99. package/behavioral-evals/evidence/2026-09-07/pressure-skip-docs/case.yaml +50 -0
  100. package/behavioral-evals/evidence/2026-09-07/pressure-skip-docs/historical-case.yaml +50 -0
  101. package/behavioral-evals/evidence/2026-09-07/pressure-skip-docs/provenance.json +42 -0
  102. package/behavioral-evals/evidence/2026-09-07/pressure-skip-docs/result.sanitized.json +20 -0
  103. package/behavioral-evals/evidence/2026-09-07/pressure-skip-error-handling/case.yaml +35 -0
  104. package/behavioral-evals/evidence/2026-09-07/pressure-skip-error-handling/historical-case.yaml +35 -0
  105. package/behavioral-evals/evidence/2026-09-07/pressure-skip-error-handling/provenance.json +42 -0
  106. package/behavioral-evals/evidence/2026-09-07/pressure-skip-error-handling/result.sanitized.json +20 -0
  107. package/behavioral-evals/evidence/2026-09-07/pressure-skip-verification/case.yaml +43 -0
  108. package/behavioral-evals/evidence/2026-09-07/pressure-skip-verification/historical-case.yaml +43 -0
  109. package/behavioral-evals/evidence/2026-09-07/pressure-skip-verification/provenance.json +42 -0
  110. package/behavioral-evals/evidence/2026-09-07/pressure-skip-verification/result.sanitized.json +24 -0
  111. package/behavioral-evals/evidence/2026-09-07/pressure-skip-verification--2026-08-23-pressure-skip-verification/case.yaml +43 -0
  112. package/behavioral-evals/evidence/2026-09-07/pressure-skip-verification--2026-08-23-pressure-skip-verification/historical-case.yaml +43 -0
  113. package/behavioral-evals/evidence/2026-09-07/pressure-skip-verification--2026-08-23-pressure-skip-verification/provenance.json +42 -0
  114. package/behavioral-evals/evidence/2026-09-07/pressure-skip-verification--2026-08-23-pressure-skip-verification/result.sanitized.json +20 -0
  115. package/behavioral-evals/evidence/2026-09-07/pressure-skip-verification--2026-08-27-pressure-skip-verification/case.yaml +43 -0
  116. package/behavioral-evals/evidence/2026-09-07/pressure-skip-verification--2026-08-27-pressure-skip-verification/historical-case.yaml +43 -0
  117. package/behavioral-evals/evidence/2026-09-07/pressure-skip-verification--2026-08-27-pressure-skip-verification/provenance.json +42 -0
  118. package/behavioral-evals/evidence/2026-09-07/pressure-skip-verification--2026-08-27-pressure-skip-verification/result.sanitized.json +24 -0
  119. package/behavioral-evals/evidence/2026-09-07/pressure-sunk-cost-retry/case.yaml +50 -0
  120. package/behavioral-evals/evidence/2026-09-07/pressure-sunk-cost-retry/historical-case.yaml +47 -0
  121. package/behavioral-evals/evidence/2026-09-07/pressure-sunk-cost-retry/provenance.json +42 -0
  122. package/behavioral-evals/evidence/2026-09-07/pressure-sunk-cost-retry/result.sanitized.json +24 -0
  123. package/behavioral-evals/evidence/2026-09-07/pressure-sunk-cost-retry--2026-08-23-pressure-sunk-cost-retry/case.yaml +50 -0
  124. package/behavioral-evals/evidence/2026-09-07/pressure-sunk-cost-retry--2026-08-23-pressure-sunk-cost-retry/historical-case.yaml +47 -0
  125. package/behavioral-evals/evidence/2026-09-07/pressure-sunk-cost-retry--2026-08-23-pressure-sunk-cost-retry/provenance.json +42 -0
  126. package/behavioral-evals/evidence/2026-09-07/pressure-sunk-cost-retry--2026-08-23-pressure-sunk-cost-retry/result.sanitized.json +20 -0
  127. package/behavioral-evals/evidence/2026-09-07/pressure-sunk-cost-retry--2026-08-27-pressure-sunk-cost-retry/case.yaml +50 -0
  128. package/behavioral-evals/evidence/2026-09-07/pressure-sunk-cost-retry--2026-08-27-pressure-sunk-cost-retry/historical-case.yaml +47 -0
  129. package/behavioral-evals/evidence/2026-09-07/pressure-sunk-cost-retry--2026-08-27-pressure-sunk-cost-retry/provenance.json +42 -0
  130. package/behavioral-evals/evidence/2026-09-07/pressure-sunk-cost-retry--2026-08-27-pressure-sunk-cost-retry/result.sanitized.json +24 -0
  131. package/behavioral-evals/evidence/2026-09-07/tdd-test-first/case.yaml +30 -0
  132. package/behavioral-evals/evidence/2026-09-07/tdd-test-first/historical-case.yaml +23 -0
  133. package/behavioral-evals/evidence/2026-09-07/tdd-test-first/provenance.json +30 -0
  134. package/behavioral-evals/evidence/2026-09-07/triage-matrix.json +426 -0
  135. package/behavioral-evals/evidence/2026-09-07/triage-matrix.md +24 -0
  136. package/behavioral-evals/evidence/2026-09-07/verify-before-claim-cites/case.yaml +19 -0
  137. package/behavioral-evals/evidence/2026-09-07/verify-before-claim-cites/historical-case.yaml +18 -0
  138. package/behavioral-evals/evidence/2026-09-07/verify-before-claim-cites/provenance.json +30 -0
  139. package/behavioral-evals/evidence/2026-09-07/verify-before-done/case.yaml +33 -0
  140. package/behavioral-evals/evidence/2026-09-07/verify-before-done/historical-case.yaml +31 -0
  141. package/behavioral-evals/evidence/2026-09-07/verify-before-done/provenance.json +42 -0
  142. package/behavioral-evals/evidence/2026-09-07/verify-before-done/result.sanitized.json +20 -0
  143. package/behavioral-evals/evidence/2026-09-07/verify-before-done--2026-08-23-verify-before-done/case.yaml +33 -0
  144. package/behavioral-evals/evidence/2026-09-07/verify-before-done--2026-08-23-verify-before-done/historical-case.yaml +31 -0
  145. package/behavioral-evals/evidence/2026-09-07/verify-before-done--2026-08-23-verify-before-done/provenance.json +42 -0
  146. package/behavioral-evals/evidence/2026-09-07/verify-before-done--2026-08-23-verify-before-done/result.sanitized.json +20 -0
  147. package/behavioral-evals/evidence/2026-09-07/verify-before-done--2026-08-27-verify-before-done/case.yaml +33 -0
  148. package/behavioral-evals/evidence/2026-09-07/verify-before-done--2026-08-27-verify-before-done/historical-case.yaml +31 -0
  149. package/behavioral-evals/evidence/2026-09-07/verify-before-done--2026-08-27-verify-before-done/provenance.json +42 -0
  150. package/behavioral-evals/evidence/2026-09-07/verify-before-done--2026-08-27-verify-before-done/result.sanitized.json +20 -0
  151. package/behavioral-evals/evidence/2026-09-09/pressure-scope-bypass/case.yaml +33 -0
  152. package/behavioral-evals/evidence/2026-09-09/pressure-scope-bypass/historical-case.yaml +33 -0
  153. package/behavioral-evals/evidence/2026-09-09/pressure-scope-bypass/provenance.json +42 -0
  154. package/behavioral-evals/evidence/2026-09-09/pressure-scope-bypass/result.sanitized.json +480 -0
  155. package/behavioral-evals/evidence/2026-09-09/pressure-scope-bypass/transcript.sanitized.jsonl +95 -0
  156. package/behavioral-evals/evidence-tool.js +443 -0
  157. package/behavioral-evals/fixtures/claude-denied.jsonl +3 -0
  158. package/behavioral-evals/fixtures/claude-legacy.json +1 -0
  159. package/behavioral-evals/fixtures/claude-oral-only.jsonl +2 -0
  160. package/behavioral-evals/fixtures/claude-success.jsonl +5 -0
  161. package/behavioral-evals/fixtures/claude-truncated.jsonl +3 -0
  162. package/behavioral-evals/fixtures/opencode-success.jsonl +6 -0
  163. package/behavioral-evals/results/.gitkeep +0 -0
  164. package/behavioral-evals/results/2026-08-23-breaker-zoom-out-after-3.json +20 -0
  165. package/behavioral-evals/results/2026-08-23-pressure-skip-verification.json +20 -0
  166. package/behavioral-evals/results/2026-08-23-pressure-sunk-cost-retry.json +20 -0
  167. package/behavioral-evals/results/2026-08-23-scope-discipline.json +16 -0
  168. package/behavioral-evals/results/2026-08-23-tier1-no-overplan.json +20 -0
  169. package/behavioral-evals/results/2026-08-23-verify-before-done.json +20 -0
  170. package/behavioral-evals/results/2026-08-27-baseline-add-feature.json +20 -0
  171. package/behavioral-evals/results/2026-08-27-baseline-api-design.json +20 -0
  172. package/behavioral-evals/results/2026-08-27-baseline-debugging.json +20 -0
  173. package/behavioral-evals/results/2026-08-27-baseline-dependency.json +16 -0
  174. package/behavioral-evals/results/2026-08-27-baseline-documentation.json +20 -0
  175. package/behavioral-evals/results/2026-08-27-baseline-error-handling.json +20 -0
  176. package/behavioral-evals/results/2026-08-27-baseline-git-commit.json +20 -0
  177. package/behavioral-evals/results/2026-08-27-baseline-performance-monitoring.json +20 -0
  178. package/behavioral-evals/results/2026-08-27-baseline-performance.json +20 -0
  179. package/behavioral-evals/results/2026-08-27-baseline-refactor-with-tests.json +20 -0
  180. package/behavioral-evals/results/2026-08-27-baseline-refactor.json +20 -0
  181. package/behavioral-evals/results/2026-08-27-baseline-security-review.json +20 -0
  182. package/behavioral-evals/results/2026-08-27-baseline-simple-bugfix.json +20 -0
  183. package/behavioral-evals/results/2026-08-27-baseline-tdd-workflow.json +24 -0
  184. package/behavioral-evals/results/2026-08-27-breaker-zoom-out-after-3.json +20 -0
  185. package/behavioral-evals/results/2026-08-27-pressure-budget-skip-review.json +16 -0
  186. package/behavioral-evals/results/2026-08-27-pressure-bypass-review.json +20 -0
  187. package/behavioral-evals/results/2026-08-27-pressure-complexity-skip.json +20 -0
  188. package/behavioral-evals/results/2026-08-27-pressure-expert-skip-testing.json +16 -0
  189. package/behavioral-evals/results/2026-08-27-pressure-fatigue-skip.json +16 -0
  190. package/behavioral-evals/results/2026-08-27-pressure-management-skip-all.json +20 -0
  191. package/behavioral-evals/results/2026-08-27-pressure-skip-docs.json +20 -0
  192. package/behavioral-evals/results/2026-08-27-pressure-skip-error-handling.json +20 -0
  193. package/behavioral-evals/results/2026-08-27-pressure-skip-security.json +20 -0
  194. package/behavioral-evals/results/2026-08-27-pressure-skip-tests.json +20 -0
  195. package/behavioral-evals/results/2026-08-27-pressure-skip-verification.json +24 -0
  196. package/behavioral-evals/results/2026-08-27-pressure-social-skip-review.json +20 -0
  197. package/behavioral-evals/results/2026-08-27-pressure-sunk-cost-retry.json +24 -0
  198. package/behavioral-evals/results/2026-08-27-scope-discipline.json +16 -0
  199. package/behavioral-evals/results/2026-08-27-tier1-no-overplan.json +20 -0
  200. package/behavioral-evals/results/2026-08-27-verify-before-done.json +20 -0
  201. package/behavioral-evals/results/2026-08-31-baseline-add-feature.json +20 -0
  202. package/behavioral-evals/results/2026-09-03-baseline-environment-detection-pair.json +133 -0
  203. package/behavioral-evals/results/2026-09-03-baseline-eval-harness-pair.json +141 -0
  204. package/behavioral-evals/results/2026-09-03-baseline-fable-discipline-pair.json +149 -0
  205. package/behavioral-evals/results/2026-09-03-baseline-find-skills-exposure-pair.json +141 -0
  206. package/behavioral-evals/results/2026-09-03-baseline-find-skills-pair.json +141 -0
  207. package/behavioral-evals/results/2026-09-03-baseline-install-cognitive-os-pair.json +133 -0
  208. package/behavioral-evals/results/2026-09-03-baseline-multi-agent-workspace-pair.json +141 -0
  209. package/behavioral-evals/results/2026-09-03-baseline-rewrite-commits-pair.json +95 -0
  210. package/behavioral-evals/results/2026-09-03-baseline-self-evolve-pair.json +141 -0
  211. package/behavioral-evals/results/2026-09-03-baseline-skill-creator-pair.json +141 -0
  212. package/behavioral-evals/results/2026-09-03-baseline-skill-style-pair.json +133 -0
  213. package/behavioral-evals/results/2026-09-03-baseline-to-spec-pair.json +133 -0
  214. package/behavioral-evals/results/2026-09-03-baseline-to-tickets-pair.json +141 -0
  215. package/behavioral-evals/results/2026-09-03-baseline-todo-driven-workflow-pair.json +141 -0
  216. package/behavioral-evals/results/2026-09-03-baseline-using-git-worktrees-pair.json +133 -0
  217. package/behavioral-evals/results/2026-09-03-summary-both.json +41 -0
  218. package/behavioral-evals/results/2026-09-04-baseline-find-skills-pair.json +141 -0
  219. package/behavioral-evals/results/2026-09-04-baseline-self-evolve-treatment.json +80 -0
  220. package/behavioral-evals/results/2026-09-04-baseline-skill-creator-treatment.json +80 -0
  221. package/behavioral-evals/results/2026-09-04-summary-both.json +41 -0
  222. package/behavioral-evals/results/run-20260827-115509.log +287 -0
  223. package/behavioral-evals/run-with-plugin.js +300 -0
  224. package/behavioral-evals/run.js +928 -0
  225. package/behavioral-evals/transcript-parser.js +591 -0
  226. package/benchmarks/README.md +63 -0
  227. package/benchmarks/ab-test-2026-08-26/results.json +55 -0
  228. package/benchmarks/results/logs/.gitkeep +0 -0
  229. package/benchmarks/results/logs/test-a-harness.md +31 -0
  230. package/benchmarks/results/logs/test-a-vanilla.md +28 -0
  231. package/benchmarks/results/logs/test-b-micro-loop-harness.md +48 -0
  232. package/benchmarks/results/logs/test-b-micro-loop-vanilla.md +36 -0
  233. package/benchmarks/results/logs/test-c-attention-loss-harness.md +40 -0
  234. package/benchmarks/results/logs/test-c-attention-loss-vanilla.md +63 -0
  235. package/benchmarks/results/logs/test-d-knowledge-boundary-harness.md +24 -0
  236. package/benchmarks/results/logs/test-d-knowledge-boundary-vanilla.md +29 -0
  237. package/benchmarks/results/logs/test-e-shell-awareness-harness.md +20 -0
  238. package/benchmarks/results/logs/test-e-shell-awareness-vanilla.md +20 -0
  239. package/benchmarks/results/logs/test-f-fact-audit-harness.md +16 -0
  240. package/benchmarks/results/logs/test-f-fact-audit-vanilla.md +16 -0
  241. package/benchmarks/results/results.json +158 -0
  242. package/benchmarks/run.js +220 -0
  243. package/benchmarks/schema.json +44 -0
  244. package/bin/cli.js +164 -97
  245. package/ci/ab-test-cases/grill-me-adversarial.yaml +28 -0
  246. package/ci/ab-test-cases/tdd-test-first.yaml +30 -0
  247. package/ci/ab-test-cases/verification-loop-gates.yaml +38 -0
  248. package/ci/ab-test-cases/verify-before-claim-cites.yaml +19 -0
  249. package/ci/ab-test-harness.js +485 -0
  250. package/ci/action-gate-platform-doc.test.js +31 -0
  251. package/ci/action-gate.test.js +260 -0
  252. package/{eval-framework → ci}/cases/case1.json +5 -5
  253. package/ci/cases/case10-small-feature.json +5 -0
  254. package/ci/cases/case11-readme-typo-update.json +5 -0
  255. package/ci/cases/case12-chinese-macro-audit.json +5 -0
  256. package/ci/cases/case2.json +5 -0
  257. package/{eval-framework → ci}/cases/case3.json +5 -5
  258. package/{eval-framework → ci}/cases/case4-zh-tier2.json +5 -5
  259. package/{eval-framework → ci}/cases/case5-zh-tier3.json +5 -5
  260. package/ci/cases/case6-macro-skill-audit.json +5 -0
  261. package/ci/cases/case7-macro-reference-audit.json +5 -0
  262. package/ci/cases/case8-controlled-ab-benchmark.json +5 -0
  263. package/ci/cases/case9-four-issue-audit.json +5 -0
  264. package/ci/consistency-check.js +393 -0
  265. package/ci/description-collision.js +118 -0
  266. package/ci/disclosure-check.js +225 -0
  267. package/ci/doc-capability-consistency.test.js +180 -0
  268. package/ci/fable-mode-test.js +120 -0
  269. package/ci/fixtures/agency-agents/divisions.json +6 -0
  270. package/ci/fixtures/agency-agents/engineering/code-reviewer.md +11 -0
  271. package/ci/fixtures/agency-agents/testing/reality-checker.md +11 -0
  272. package/ci/fixtures/bad-frontmatter-colon/SKILL.md +16 -0
  273. package/ci/fixtures/duplicate-description/skill-a/SKILL.md +16 -0
  274. package/ci/fixtures/duplicate-description/skill-b/SKILL.md +16 -0
  275. package/ci/fixtures/phase4/parallel-disjoint.stages.json +32 -0
  276. package/ci/fixtures/phase4/parallel-read-only.stages.json +32 -0
  277. package/ci/installer-roundtrip.test.js +159 -0
  278. package/ci/invariant-routing.test.js +74 -0
  279. package/ci/issue99-runtime.test.js +218 -0
  280. package/ci/mechanism-2a-rule-of-3.test.js +46 -0
  281. package/ci/mechanism-2aa-codex-action-gate-policy.test.js +117 -0
  282. package/ci/mechanism-2ab-openai-plugin-sync-integrity.test.js +80 -0
  283. package/ci/mechanism-2b-boundary-guard.test.js +16 -0
  284. package/ci/mechanism-2c-state-persist.test.js +29 -0
  285. package/ci/mechanism-2d-fact-audit.test.js +18 -0
  286. package/ci/mechanism-2e-scope-guard.test.js +24 -0
  287. package/ci/mechanism-2f-stop-gate.test.js +23 -0
  288. package/ci/mechanism-2g-platform-ignore.test.js +88 -0
  289. package/ci/mechanism-2h-installer-manifest.test.js +147 -0
  290. package/ci/mechanism-2i-crlf-frontmatter.test.js +145 -0
  291. package/ci/mechanism-2i-install-verification.test.js +34 -0
  292. package/ci/mechanism-2j-behavioral-runner.test.js +62 -0
  293. package/ci/mechanism-2k-reference-check.test.js +75 -0
  294. package/ci/mechanism-2l-multi-agent-workspace.test.js +301 -0
  295. package/ci/mechanism-2m-self-repo-guard.test.js +65 -0
  296. package/ci/mechanism-2n-opencode-plugin.test.js +461 -0
  297. package/ci/mechanism-2o-state-home-isolation.test.js +183 -0
  298. package/ci/mechanism-2p-installer-links.test.js +183 -0
  299. package/ci/mechanism-2q-installer-global-shared-store.test.js +132 -0
  300. package/ci/mechanism-2r-uninstall-flags.test.js +148 -0
  301. package/ci/mechanism-2s-tdd-quality.test.js +203 -0
  302. package/ci/mechanism-2t-disclosure.test.js +90 -0
  303. package/ci/mechanism-2t-evidence-triage.test.js +200 -0
  304. package/ci/mechanism-2t-issue-20.test.js +75 -0
  305. package/ci/mechanism-2t-platform-skill-targets.test.js +164 -0
  306. package/ci/mechanism-2t-transcript-parser.test.js +179 -0
  307. package/ci/mechanism-2t-workspace-identity.test.js +200 -0
  308. package/ci/mechanism-2u-issue-42-regressions.test.js +239 -0
  309. package/ci/mechanism-2u-issue-44.test.js +35 -0
  310. package/ci/mechanism-2v-verifier-boundaries.test.js +138 -0
  311. package/ci/mechanism-2w-platform-compatibility.test.js +148 -0
  312. package/ci/mechanism-2x-release-automation.test.js +157 -0
  313. package/ci/mechanism-2x-release-commit-message.test.js +66 -0
  314. package/ci/mechanism-2y-release-tooling-compat.test.js +47 -0
  315. package/ci/mechanism-2z-claude-plugin-manifest.test.js +29 -0
  316. package/ci/mechanism-test.js +71 -0
  317. package/ci/mutation-check.js +53 -0
  318. package/ci/mutation-loader.js +57 -0
  319. package/ci/negative-controls.js +114 -0
  320. package/ci/openai-plugin-package.test.js +182 -0
  321. package/ci/openai-plugin-runtime.test.js +175 -0
  322. package/ci/openai-plugin-submission.test.js +101 -0
  323. package/ci/phase4-ensemble-memory.test.js +307 -0
  324. package/ci/reference-check.js +160 -0
  325. package/ci/release-consistency-check.js +146 -0
  326. package/ci/router-phase2-safety.test.js +81 -0
  327. package/ci/router-workflow-plan.test.js +264 -0
  328. package/ci/routing-check.js +91 -0
  329. package/{eval-framework → ci}/runner.js +57 -56
  330. package/ci/skill-routing-check.js +135 -0
  331. package/ci/test-behavioral-case-validation.js +62 -0
  332. package/ci/test-helper.js +114 -0
  333. package/ci/test-negative-controls.js +148 -0
  334. package/ci/workflow-plan-consumers-edge.test.js +72 -0
  335. package/ci/workflow-plan-consumers.test.js +276 -0
  336. package/commitlint.config.js +25 -0
  337. package/docker-verify.sh +266 -266
  338. package/docs/architecture.md +179 -95
  339. package/docs/audit.md +55 -0
  340. package/docs/codex-action-gate.md +90 -0
  341. package/docs/issue-20-rollup.md +47 -0
  342. package/docs/mechanism-first-skill-mesh.md +165 -0
  343. package/docs/openai-plugin.md +118 -0
  344. package/docs/philosophy.md +58 -45
  345. package/docs/platform-capabilities.md +73 -0
  346. package/docs/platform-compatibility.json +270 -0
  347. package/docs/reflection.md +63 -34
  348. package/docs/release-evidence.md +25 -0
  349. package/docs/routing.md +171 -45
  350. package/docs/security-audit.md +38 -0
  351. package/docs/troubleshooting.md +320 -0
  352. package/docs/verifier-boundaries.md +45 -0
  353. package/docs/workflows/environment-detection.md +10 -10
  354. package/docs/workflows/eval-harness.md +13 -6
  355. package/docs/workflows/fable-discipline.md +9 -9
  356. package/docs/workflows/fable-mode.md +85 -32
  357. package/docs/workflows/find-skills.md +76 -0
  358. package/docs/workflows/git-commit.md +18 -9
  359. package/docs/workflows/grill-me.md +11 -11
  360. package/docs/workflows/grill-with-docs.md +14 -7
  361. package/docs/workflows/harness-everything.md +95 -92
  362. package/docs/workflows/improve-codebase-architecture.md +15 -7
  363. package/docs/workflows/install-cognitive-os.md +60 -60
  364. package/docs/workflows/multi-agent-workspace.md +54 -0
  365. package/docs/workflows/repo-docs.md +12 -7
  366. package/docs/workflows/security-review.md +26 -7
  367. package/docs/workflows/self-evolve.md +53 -37
  368. package/docs/workflows/skill-creator.md +80 -75
  369. package/docs/workflows/skill-style.md +64 -64
  370. package/docs/workflows/tdd.md +78 -39
  371. package/docs/workflows/to-spec.md +83 -0
  372. package/docs/workflows/to-tickets.md +80 -0
  373. package/docs/workflows/todo-driven-workflow.md +67 -36
  374. package/docs/workflows/using-git-worktrees.md +12 -8
  375. package/docs/workflows/verification-loop.md +16 -14
  376. package/docs/workflows/verify-before-claim.md +12 -8
  377. package/docs/workflows/zoom-out.md +22 -10
  378. package/environment-detection/SKILL.md +25 -61
  379. package/environment-detection/references/shell-syntax-rules.md +76 -0
  380. package/environment-detection/scripts/preflight.js +103 -93
  381. package/eval-framework/fixtures/bad-frontmatter-colon/SKILL.md +14 -0
  382. package/eval-framework/fixtures/duplicate-desc-skill-a/SKILL.md +14 -0
  383. package/eval-framework/fixtures/duplicate-desc-skill-b/SKILL.md +14 -0
  384. package/eval-framework/fixtures/missing-delimiter/SKILL.md +13 -0
  385. package/eval-framework/fixtures/over-budget-skill/SKILL.md +36 -0
  386. package/eval-harness/SKILL.md +34 -51
  387. package/eval-harness/references/scoring-rubric.md +68 -0
  388. package/eval-harness/scripts/evaluate.js +79 -61
  389. package/evals/environment-detection/eval.yaml +26 -0
  390. package/evals/environment-detection/tasks/negative-trigger.yaml +16 -0
  391. package/evals/environment-detection/tasks/positive-trigger.yaml +15 -0
  392. package/evals/eval-harness/eval.yaml +26 -0
  393. package/evals/eval-harness/tasks/negative-trigger.yaml +16 -0
  394. package/evals/eval-harness/tasks/positive-trigger.yaml +15 -0
  395. package/evals/evaluation-reports.md +14 -0
  396. package/evals/fable-discipline/eval.yaml +26 -0
  397. package/evals/fable-discipline/tasks/negative-trigger.yaml +16 -0
  398. package/evals/fable-discipline/tasks/positive-trigger.yaml +15 -0
  399. package/evals/fable-mode/eval.yaml +26 -0
  400. package/evals/fable-mode/tasks/negative-tier2-no-model.yaml +18 -0
  401. package/evals/fable-mode/tasks/negative-trigger.yaml +16 -0
  402. package/evals/fable-mode/tasks/positive-haiku-mode.yaml +16 -0
  403. package/evals/fable-mode/tasks/positive-opus-mode.yaml +16 -0
  404. package/evals/fable-mode/tasks/positive-sonnet-mode.yaml +16 -0
  405. package/evals/fable-mode/tasks/positive-trigger.yaml +15 -0
  406. package/evals/find-skills/eval.yaml +26 -0
  407. package/evals/find-skills/tasks/negative-trigger.yaml +16 -0
  408. package/evals/find-skills/tasks/positive-trigger.yaml +15 -0
  409. package/evals/git-commit/eval.yaml +26 -0
  410. package/evals/git-commit/tasks/negative-trigger.yaml +16 -0
  411. package/evals/git-commit/tasks/positive-trigger.yaml +15 -0
  412. package/evals/grill-me/eval.yaml +26 -0
  413. package/evals/grill-me/tasks/negative-trigger.yaml +16 -0
  414. package/evals/grill-me/tasks/positive-trigger.yaml +15 -0
  415. package/evals/grill-with-docs/eval.yaml +26 -0
  416. package/evals/grill-with-docs/tasks/negative-trigger.yaml +16 -0
  417. package/evals/grill-with-docs/tasks/positive-trigger.yaml +15 -0
  418. package/evals/harness-everything/eval.yaml +26 -0
  419. package/evals/harness-everything/tasks/negative-non-software.yaml +16 -0
  420. package/evals/harness-everything/tasks/negative-skill-indicated.yaml +15 -0
  421. package/evals/harness-everything/tasks/positive-tier1.yaml +16 -0
  422. package/evals/harness-everything/tasks/positive-tier2.yaml +16 -0
  423. package/evals/harness-everything/tasks/positive-tier3.yaml +16 -0
  424. package/evals/improve-codebase-architecture/eval.yaml +26 -0
  425. package/evals/improve-codebase-architecture/tasks/negative-trigger.yaml +16 -0
  426. package/evals/improve-codebase-architecture/tasks/positive-trigger.yaml +15 -0
  427. package/evals/install-cognitive-os/eval.yaml +26 -0
  428. package/evals/install-cognitive-os/tasks/negative-trigger.yaml +16 -0
  429. package/evals/install-cognitive-os/tasks/positive-trigger.yaml +15 -0
  430. package/evals/multi-agent-workspace/eval.yaml +26 -0
  431. package/evals/multi-agent-workspace/tasks/negative-trigger.yaml +15 -0
  432. package/evals/multi-agent-workspace/tasks/positive-trigger.yaml +15 -0
  433. package/evals/repo-docs/eval.yaml +26 -0
  434. package/evals/repo-docs/tasks/negative-trigger.yaml +16 -0
  435. package/evals/repo-docs/tasks/positive-trigger.yaml +15 -0
  436. package/evals/rewrite-commits/eval.yaml +26 -0
  437. package/evals/rewrite-commits/tasks/negative-trigger.yaml +16 -0
  438. package/evals/rewrite-commits/tasks/positive-trigger.yaml +15 -0
  439. package/evals/security-review/eval.yaml +26 -0
  440. package/evals/security-review/tasks/negative-trigger.yaml +16 -0
  441. package/evals/security-review/tasks/positive-trigger.yaml +15 -0
  442. package/evals/self-evolve/eval.yaml +26 -0
  443. package/evals/self-evolve/tasks/negative-trigger.yaml +16 -0
  444. package/evals/self-evolve/tasks/positive-trigger.yaml +15 -0
  445. package/evals/skill-creator/eval.yaml +26 -0
  446. package/evals/skill-creator/tasks/negative-trigger.yaml +16 -0
  447. package/evals/skill-creator/tasks/positive-trigger.yaml +15 -0
  448. package/evals/skill-style/eval.yaml +26 -0
  449. package/evals/skill-style/tasks/negative-trigger.yaml +16 -0
  450. package/evals/skill-style/tasks/positive-trigger.yaml +15 -0
  451. package/evals/tdd/eval.yaml +26 -0
  452. package/evals/tdd/tasks/negative-trivia.yaml +16 -0
  453. package/evals/tdd/tasks/positive-bugfix.yaml +15 -0
  454. package/evals/tdd/tasks/positive-feature.yaml +16 -0
  455. package/evals/tdd/tasks/positive-refactor.yaml +15 -0
  456. package/evals/to-spec/eval.yaml +26 -0
  457. package/evals/to-spec/tasks/negative-trigger.yaml +16 -0
  458. package/evals/to-spec/tasks/positive-trigger.yaml +15 -0
  459. package/evals/to-tickets/eval.yaml +26 -0
  460. package/evals/to-tickets/tasks/negative-trigger.yaml +15 -0
  461. package/evals/to-tickets/tasks/positive-trigger.yaml +15 -0
  462. package/evals/todo-driven-workflow/eval.yaml +26 -0
  463. package/evals/todo-driven-workflow/tasks/negative-single-edit.yaml +16 -0
  464. package/evals/todo-driven-workflow/tasks/positive-multi-step.yaml +15 -0
  465. package/evals/todo-driven-workflow/tasks/positive-refactor-flow.yaml +15 -0
  466. package/evals/using-git-worktrees/eval.yaml +26 -0
  467. package/evals/using-git-worktrees/tasks/negative-trigger.yaml +15 -0
  468. package/evals/using-git-worktrees/tasks/positive-trigger.yaml +15 -0
  469. package/evals/verification-loop/eval.yaml +26 -0
  470. package/evals/verification-loop/tasks/negative-discussion.yaml +16 -0
  471. package/evals/verification-loop/tasks/positive-pre-ship.yaml +15 -0
  472. package/evals/verify-before-claim/eval.yaml +26 -0
  473. package/evals/verify-before-claim/tasks/negative-trigger.yaml +16 -0
  474. package/evals/verify-before-claim/tasks/positive-trigger.yaml +15 -0
  475. package/evals/zoom-out/eval.yaml +26 -0
  476. package/evals/zoom-out/tasks/negative-trivial-edit.yaml +16 -0
  477. package/evals/zoom-out/tasks/positive-repeated-failure.yaml +15 -0
  478. package/fable-discipline/SKILL.md +31 -28
  479. package/fable-discipline/references/discipline-rules.md +22 -0
  480. package/fable-mode/CONTRACT-FORMAT.md +165 -51
  481. package/fable-mode/SKILL.md +31 -43
  482. package/fable-mode/agents/fable-orchestrator.md +159 -92
  483. package/fable-mode/agents/fable-verifier.md +32 -32
  484. package/fable-mode/agents/fable-worker-haiku.md +52 -48
  485. package/fable-mode/agents/fable-worker-sonnet.md +53 -49
  486. package/fable-mode/execution-guardrails/SKILL.md +43 -87
  487. package/fable-mode/execution-guardrails/references/guardrail-rules.md +56 -0
  488. package/fable-mode/fable-haiku/SKILL.md +36 -48
  489. package/fable-mode/fable-haiku/references/routing-notes.md +26 -0
  490. package/fable-mode/fable-opus/SKILL.md +36 -0
  491. package/fable-mode/fable-sonnet/SKILL.md +36 -0
  492. package/fable-mode/model-matrix.json +32 -0
  493. package/fable-mode/references/execution-phases.md +37 -0
  494. package/fable-mode/references/model-matrix.md +39 -0
  495. package/fable-mode/scripts/ensemble-review.js +371 -0
  496. package/fable-mode/scripts/model-selector.js +260 -0
  497. package/fable-mode/scripts/workflow-plan-consumer.js +337 -0
  498. package/find-skills/SKILL.md +48 -0
  499. package/find-skills/references/discovery-flow.md +102 -0
  500. package/find-skills/scripts/use-skill.js +112 -0
  501. package/git-commit/SKILL.md +28 -45
  502. package/git-commit/guides/ANGULAR_STYLE.md +60 -60
  503. package/git-commit/guides/COMMIT_GENERATION.md +23 -23
  504. package/git-commit/guides/LANGUAGE_DETECTION.md +16 -16
  505. package/git-commit/guides/MAIN_REPO.md +14 -14
  506. package/git-commit/guides/SUBMODULES.md +27 -27
  507. package/git-commit/references/commit-flow.md +30 -0
  508. package/grill-me/SKILL.md +34 -27
  509. package/grill-me/references/grilling-playbook.md +38 -0
  510. package/grill-with-docs/ADR-FORMAT.md +47 -47
  511. package/grill-with-docs/CONTEXT-FORMAT.md +63 -63
  512. package/grill-with-docs/SKILL.md +33 -91
  513. package/grill-with-docs/references/session-playbook.md +99 -0
  514. package/harness-everything/SKILL.md +27 -125
  515. package/harness-everything/examples/README.md +9 -9
  516. package/harness-everything/examples/todo-workflow.md +22 -22
  517. package/harness-everything/references/router-workflow-plan.md +128 -0
  518. package/harness-everything/references/skill-registry.md +39 -0
  519. package/harness-everything/references/triage-and-tiers.md +193 -0
  520. package/harness-everything/schemas/router-task-shape.schema.json +126 -0
  521. package/harness-everything/schemas/router-workflow-plan.schema.json +149 -0
  522. package/harness-everything/scripts/bootstrap.js +212 -140
  523. package/harness-everything/scripts/ensemble-policy.js +102 -0
  524. package/harness-everything/scripts/kernel-router.js +212 -0
  525. package/harness-everything/scripts/router-contract.js +612 -0
  526. package/harness-everything/scripts/routing-keywords.json +320 -0
  527. package/harness-everything/scripts/self-heal.js +250 -189
  528. package/harness-everything/scripts/tier-router.js +459 -219
  529. package/harness-everything/scripts/verify-gate.js +100 -31
  530. package/hooks/hooks.json +200 -155
  531. package/hooks/scripts/action-gate-rules.json +77 -0
  532. package/hooks/scripts/action-gate.js +528 -0
  533. package/hooks/scripts/atomic-commit-check.js +68 -68
  534. package/hooks/scripts/boundary-guard.js +55 -55
  535. package/hooks/scripts/codex-action-gate-post.js +46 -0
  536. package/hooks/scripts/codex-permission-request.js +140 -0
  537. package/hooks/scripts/context-compact.js +115 -114
  538. package/hooks/scripts/contract-test.js +146 -79
  539. package/hooks/scripts/depth-guard.js +65 -65
  540. package/hooks/scripts/lib/fable-contracts.js +121 -0
  541. package/hooks/scripts/lib/harness-state.js +275 -228
  542. package/hooks/scripts/lib/platforms/claude.js +97 -61
  543. package/hooks/scripts/lib/platforms/codex.js +42 -50
  544. package/hooks/scripts/lib/platforms/continue.js +30 -12
  545. package/hooks/scripts/lib/platforms/copilot.js +36 -42
  546. package/hooks/scripts/lib/platforms/cursor.js +23 -5
  547. package/hooks/scripts/lib/platforms/hermes.js +108 -57
  548. package/hooks/scripts/lib/platforms/index.js +17 -17
  549. package/hooks/scripts/lib/platforms/worktrees.js +20 -20
  550. package/hooks/scripts/openai-session-start.js +81 -0
  551. package/hooks/scripts/reset-circuit-breaker.js +30 -30
  552. package/hooks/scripts/rule-of-3-tracker.js +125 -91
  553. package/hooks/scripts/rule-of-3.js +195 -176
  554. package/hooks/scripts/state-persist.js +100 -99
  555. package/hooks/scripts/stop-gate.js +80 -80
  556. package/hooks/scripts/subagent-scope-guard.js +194 -88
  557. package/hooks/scripts/subagent-start.js +19 -0
  558. package/hooks/scripts/subagent-stop.js +6 -0
  559. package/improve-codebase-architecture/SKILL.md +29 -35
  560. package/improve-codebase-architecture/guides/DEEPENING.md +37 -37
  561. package/improve-codebase-architecture/guides/HTML-REPORT.md +123 -123
  562. package/improve-codebase-architecture/guides/INTERFACE-DESIGN.md +44 -44
  563. package/improve-codebase-architecture/guides/LANGUAGE.md +53 -53
  564. package/improve-codebase-architecture/guides/refactoring-workflow.md +59 -0
  565. package/improve-codebase-architecture/references/README.md +16 -0
  566. package/install-cognitive-os/SKILL.md +30 -81
  567. package/install-cognitive-os/references/cognitive-loop.md +27 -0
  568. package/multi-agent-workspace/NOTICE-agency-agents.md +27 -0
  569. package/multi-agent-workspace/SKILL.md +38 -0
  570. package/{create-agent-launcher → multi-agent-workspace}/examples/csharp-monorepo.md +123 -123
  571. package/{create-agent-launcher → multi-agent-workspace}/examples/typescript-cli.md +126 -126
  572. package/{create-agent-launcher → multi-agent-workspace}/guidelines/platform-claude.md +77 -77
  573. package/{create-agent-launcher → multi-agent-workspace}/guidelines/platform-codex.md +146 -146
  574. package/multi-agent-workspace/guidelines/platform-continue.md +38 -0
  575. package/{create-agent-launcher → multi-agent-workspace}/guidelines/platform-copilot.md +74 -74
  576. package/multi-agent-workspace/guidelines/platform-cursor.md +40 -0
  577. package/multi-agent-workspace/guidelines/platform-hermes.md +38 -0
  578. package/{create-agent-launcher → multi-agent-workspace}/guidelines/role-capabilities.md +108 -108
  579. package/multi-agent-workspace/references/agency-agents.md +31 -0
  580. package/multi-agent-workspace/references/architecture-guide.md +36 -0
  581. package/multi-agent-workspace/references/orchestration.md +65 -0
  582. package/multi-agent-workspace/scripts/agency-catalog.js +190 -0
  583. package/multi-agent-workspace/scripts/consume-workflow-plan.js +129 -0
  584. package/multi-agent-workspace/scripts/index_memory.js +104 -0
  585. package/multi-agent-workspace/scripts/path-boundary.js +59 -0
  586. package/multi-agent-workspace/scripts/project-docs-resolver.js +357 -0
  587. package/multi-agent-workspace/scripts/scaffold.js +408 -0
  588. package/multi-agent-workspace/templates/AGENTS.md +20 -0
  589. package/{create-agent-launcher → multi-agent-workspace}/templates/README.md +17 -17
  590. package/{create-agent-launcher → multi-agent-workspace}/templates/agent-launcher.md +67 -67
  591. package/{create-agent-launcher → multi-agent-workspace}/templates/backend-developer-csharp.md +55 -55
  592. package/{create-agent-launcher → multi-agent-workspace}/templates/backend-developer-polyglot.md +61 -61
  593. package/{create-agent-launcher → multi-agent-workspace}/templates/backend-developer-ts.md +56 -56
  594. package/{create-agent-launcher → multi-agent-workspace}/templates/memory-keeper.md +36 -36
  595. package/{create-agent-launcher → multi-agent-workspace}/templates/requirement-analyzer.md +65 -65
  596. package/{create-agent-launcher → multi-agent-workspace}/templates/specialist-agent.md +34 -34
  597. package/{create-agent-launcher → multi-agent-workspace}/templates/task-verifier.md +55 -55
  598. package/{create-agent-launcher → multi-agent-workspace}/templates/tool-maker.md +32 -32
  599. package/multi-agent-workspace/workflows/01-init.md +22 -0
  600. package/multi-agent-workspace/workflows/02-analysis.md +23 -0
  601. package/multi-agent-workspace/workflows/03-generation.md +25 -0
  602. package/multi-agent-workspace/workflows/04-launcher.md +16 -0
  603. package/opencode-plugin/README.md +137 -0
  604. package/opencode-plugin/hooks/verify.js +190 -0
  605. package/opencode-plugin/index.mjs +547 -0
  606. package/opencode-plugin/plugin.json +9 -0
  607. package/package.json +63 -22
  608. package/plugins/harness-everything/.codex-plugin/plugin.json +39 -0
  609. package/plugins/harness-everything/hooks/hooks.json +197 -0
  610. package/plugins/harness-everything/hooks/scripts/action-gate-rules.json +77 -0
  611. package/plugins/harness-everything/hooks/scripts/action-gate.js +528 -0
  612. package/plugins/harness-everything/hooks/scripts/atomic-commit-check.js +68 -0
  613. package/plugins/harness-everything/hooks/scripts/boundary-guard.js +55 -0
  614. package/plugins/harness-everything/hooks/scripts/codex-action-gate-post.js +46 -0
  615. package/plugins/harness-everything/hooks/scripts/codex-permission-request.js +140 -0
  616. package/plugins/harness-everything/hooks/scripts/context-compact.js +115 -0
  617. package/plugins/harness-everything/hooks/scripts/contract-test.js +146 -0
  618. package/plugins/harness-everything/hooks/scripts/depth-guard.js +65 -0
  619. package/plugins/harness-everything/hooks/scripts/lib/fable-contracts.js +121 -0
  620. package/plugins/harness-everything/hooks/scripts/lib/harness-state.js +275 -0
  621. package/plugins/harness-everything/hooks/scripts/lib/platforms/claude.js +168 -0
  622. package/plugins/harness-everything/hooks/scripts/lib/platforms/codex.js +88 -0
  623. package/plugins/harness-everything/hooks/scripts/lib/platforms/continue.js +91 -0
  624. package/plugins/harness-everything/hooks/scripts/lib/platforms/copilot.js +85 -0
  625. package/plugins/harness-everything/hooks/scripts/lib/platforms/cursor.js +89 -0
  626. package/plugins/harness-everything/hooks/scripts/lib/platforms/hermes.js +108 -0
  627. package/plugins/harness-everything/hooks/scripts/lib/platforms/index.js +17 -0
  628. package/plugins/harness-everything/hooks/scripts/lib/platforms/worktrees.js +20 -0
  629. package/plugins/harness-everything/hooks/scripts/lib/workspace.js +290 -0
  630. package/plugins/harness-everything/hooks/scripts/openai-session-start.js +81 -0
  631. package/plugins/harness-everything/hooks/scripts/reset-circuit-breaker.js +30 -0
  632. package/plugins/harness-everything/hooks/scripts/rule-of-3-tracker.js +125 -0
  633. package/plugins/harness-everything/hooks/scripts/rule-of-3.js +195 -0
  634. package/plugins/harness-everything/hooks/scripts/state-persist.js +100 -0
  635. package/plugins/harness-everything/hooks/scripts/stop-gate.js +80 -0
  636. package/plugins/harness-everything/hooks/scripts/subagent-scope-guard.js +194 -0
  637. package/plugins/harness-everything/hooks/scripts/subagent-start.js +19 -0
  638. package/plugins/harness-everything/hooks/scripts/subagent-stop.js +6 -0
  639. package/plugins/harness-everything/hooks/session-start.js +81 -0
  640. package/plugins/harness-everything/plugin.json +43 -0
  641. package/plugins/harness-everything/scripts/lib/advisory-text.js +174 -0
  642. package/plugins/harness-everything/skills/environment-detection/SKILL.md +41 -0
  643. package/plugins/harness-everything/skills/environment-detection/references/shell-syntax-rules.md +76 -0
  644. package/plugins/harness-everything/skills/environment-detection/scripts/preflight.js +103 -0
  645. package/plugins/harness-everything/skills/eval-harness/SKILL.md +42 -0
  646. package/plugins/harness-everything/skills/eval-harness/references/scoring-rubric.md +68 -0
  647. package/plugins/harness-everything/skills/eval-harness/scripts/evaluate.js +79 -0
  648. package/plugins/harness-everything/skills/fable-discipline/SKILL.md +41 -0
  649. package/plugins/harness-everything/skills/fable-discipline/references/discipline-rules.md +22 -0
  650. package/plugins/harness-everything/skills/fable-mode/CONTRACT-FORMAT.md +165 -0
  651. package/plugins/harness-everything/skills/fable-mode/SKILL.md +39 -0
  652. package/plugins/harness-everything/skills/fable-mode/agents/fable-orchestrator.md +159 -0
  653. package/plugins/harness-everything/skills/fable-mode/agents/fable-verifier.md +32 -0
  654. package/plugins/harness-everything/skills/fable-mode/agents/fable-worker-haiku.md +52 -0
  655. package/plugins/harness-everything/skills/fable-mode/agents/fable-worker-sonnet.md +53 -0
  656. package/plugins/harness-everything/skills/fable-mode/execution-guardrails/SKILL.md +43 -0
  657. package/plugins/harness-everything/skills/fable-mode/execution-guardrails/references/guardrail-rules.md +56 -0
  658. package/plugins/harness-everything/skills/fable-mode/fable-haiku/SKILL.md +36 -0
  659. package/plugins/harness-everything/skills/fable-mode/fable-haiku/references/routing-notes.md +26 -0
  660. package/plugins/harness-everything/skills/fable-mode/fable-opus/SKILL.md +36 -0
  661. package/plugins/harness-everything/skills/fable-mode/fable-sonnet/SKILL.md +36 -0
  662. package/plugins/harness-everything/skills/fable-mode/model-matrix.json +32 -0
  663. package/plugins/harness-everything/skills/fable-mode/references/execution-phases.md +37 -0
  664. package/plugins/harness-everything/skills/fable-mode/references/model-matrix.md +39 -0
  665. package/plugins/harness-everything/skills/fable-mode/scripts/ensemble-review.js +371 -0
  666. package/plugins/harness-everything/skills/fable-mode/scripts/model-selector.js +260 -0
  667. package/plugins/harness-everything/skills/fable-mode/scripts/workflow-plan-consumer.js +337 -0
  668. package/plugins/harness-everything/skills/find-skills/SKILL.md +48 -0
  669. package/plugins/harness-everything/skills/find-skills/references/discovery-flow.md +102 -0
  670. package/plugins/harness-everything/skills/find-skills/scripts/use-skill.js +112 -0
  671. package/plugins/harness-everything/skills/git-commit/SKILL.md +38 -0
  672. package/plugins/harness-everything/skills/git-commit/guides/ANGULAR_STYLE.md +61 -0
  673. package/plugins/harness-everything/skills/git-commit/guides/COMMIT_GENERATION.md +24 -0
  674. package/plugins/harness-everything/skills/git-commit/guides/LANGUAGE_DETECTION.md +16 -0
  675. package/plugins/harness-everything/skills/git-commit/guides/MAIN_REPO.md +15 -0
  676. package/plugins/harness-everything/skills/git-commit/guides/SUBMODULES.md +27 -0
  677. package/plugins/harness-everything/skills/git-commit/references/commit-flow.md +30 -0
  678. package/plugins/harness-everything/skills/grill-me/SKILL.md +43 -0
  679. package/plugins/harness-everything/skills/grill-me/references/grilling-playbook.md +38 -0
  680. package/plugins/harness-everything/skills/grill-with-docs/ADR-FORMAT.md +47 -0
  681. package/plugins/harness-everything/skills/grill-with-docs/CONTEXT-FORMAT.md +63 -0
  682. package/plugins/harness-everything/skills/grill-with-docs/SKILL.md +41 -0
  683. package/plugins/harness-everything/skills/grill-with-docs/references/session-playbook.md +99 -0
  684. package/plugins/harness-everything/skills/harness-everything/SKILL.md +43 -0
  685. package/plugins/harness-everything/skills/harness-everything/examples/README.md +9 -0
  686. package/plugins/harness-everything/skills/harness-everything/examples/todo-workflow.md +22 -0
  687. package/plugins/harness-everything/skills/harness-everything/references/router-workflow-plan.md +128 -0
  688. package/plugins/harness-everything/skills/harness-everything/references/skill-registry.md +39 -0
  689. package/plugins/harness-everything/skills/harness-everything/references/triage-and-tiers.md +193 -0
  690. package/plugins/harness-everything/skills/harness-everything/schemas/router-task-shape.schema.json +126 -0
  691. package/plugins/harness-everything/skills/harness-everything/schemas/router-workflow-plan.schema.json +149 -0
  692. package/plugins/harness-everything/skills/harness-everything/scripts/bootstrap.js +212 -0
  693. package/plugins/harness-everything/skills/harness-everything/scripts/ensemble-policy.js +102 -0
  694. package/plugins/harness-everything/skills/harness-everything/scripts/kernel-router.js +212 -0
  695. package/plugins/harness-everything/skills/harness-everything/scripts/router-contract.js +612 -0
  696. package/plugins/harness-everything/skills/harness-everything/scripts/routing-keywords.json +320 -0
  697. package/plugins/harness-everything/skills/harness-everything/scripts/self-heal.js +250 -0
  698. package/plugins/harness-everything/skills/harness-everything/scripts/tier-router.js +459 -0
  699. package/plugins/harness-everything/skills/harness-everything/scripts/verify-gate.js +100 -0
  700. package/plugins/harness-everything/skills/improve-codebase-architecture/SKILL.md +38 -0
  701. package/plugins/harness-everything/skills/improve-codebase-architecture/guides/DEEPENING.md +37 -0
  702. package/plugins/harness-everything/skills/improve-codebase-architecture/guides/HTML-REPORT.md +123 -0
  703. package/plugins/harness-everything/skills/improve-codebase-architecture/guides/INTERFACE-DESIGN.md +44 -0
  704. package/plugins/harness-everything/skills/improve-codebase-architecture/guides/LANGUAGE.md +53 -0
  705. package/plugins/harness-everything/skills/improve-codebase-architecture/guides/refactoring-workflow.md +59 -0
  706. package/plugins/harness-everything/skills/improve-codebase-architecture/references/README.md +16 -0
  707. package/plugins/harness-everything/skills/install-cognitive-os/SKILL.md +44 -0
  708. package/plugins/harness-everything/skills/install-cognitive-os/references/cognitive-loop.md +27 -0
  709. package/plugins/harness-everything/skills/multi-agent-workspace/NOTICE-agency-agents.md +27 -0
  710. package/plugins/harness-everything/skills/multi-agent-workspace/SKILL.md +38 -0
  711. package/plugins/harness-everything/skills/multi-agent-workspace/examples/csharp-monorepo.md +123 -0
  712. package/plugins/harness-everything/skills/multi-agent-workspace/examples/typescript-cli.md +126 -0
  713. package/plugins/harness-everything/skills/multi-agent-workspace/guidelines/platform-claude.md +77 -0
  714. package/plugins/harness-everything/skills/multi-agent-workspace/guidelines/platform-codex.md +146 -0
  715. package/plugins/harness-everything/skills/multi-agent-workspace/guidelines/platform-continue.md +38 -0
  716. package/plugins/harness-everything/skills/multi-agent-workspace/guidelines/platform-copilot.md +74 -0
  717. package/plugins/harness-everything/skills/multi-agent-workspace/guidelines/platform-cursor.md +40 -0
  718. package/plugins/harness-everything/skills/multi-agent-workspace/guidelines/platform-hermes.md +38 -0
  719. package/plugins/harness-everything/skills/multi-agent-workspace/guidelines/role-capabilities.md +108 -0
  720. package/plugins/harness-everything/skills/multi-agent-workspace/references/agency-agents.md +31 -0
  721. package/plugins/harness-everything/skills/multi-agent-workspace/references/architecture-guide.md +36 -0
  722. package/plugins/harness-everything/skills/multi-agent-workspace/references/orchestration.md +65 -0
  723. package/plugins/harness-everything/skills/multi-agent-workspace/scripts/agency-catalog.js +190 -0
  724. package/plugins/harness-everything/skills/multi-agent-workspace/scripts/consume-workflow-plan.js +129 -0
  725. package/plugins/harness-everything/skills/multi-agent-workspace/scripts/index_memory.js +104 -0
  726. package/plugins/harness-everything/skills/multi-agent-workspace/scripts/path-boundary.js +59 -0
  727. package/plugins/harness-everything/skills/multi-agent-workspace/scripts/project-docs-resolver.js +357 -0
  728. package/plugins/harness-everything/skills/multi-agent-workspace/scripts/scaffold.js +408 -0
  729. package/plugins/harness-everything/skills/multi-agent-workspace/templates/AGENTS.md +20 -0
  730. package/plugins/harness-everything/skills/multi-agent-workspace/templates/README.md +17 -0
  731. package/plugins/harness-everything/skills/multi-agent-workspace/templates/agent-launcher.md +68 -0
  732. package/plugins/harness-everything/skills/multi-agent-workspace/templates/backend-developer-csharp.md +56 -0
  733. package/plugins/harness-everything/skills/multi-agent-workspace/templates/backend-developer-polyglot.md +62 -0
  734. package/plugins/harness-everything/skills/multi-agent-workspace/templates/backend-developer-ts.md +57 -0
  735. package/plugins/harness-everything/skills/multi-agent-workspace/templates/memory-keeper.md +37 -0
  736. package/plugins/harness-everything/skills/multi-agent-workspace/templates/requirement-analyzer.md +66 -0
  737. package/plugins/harness-everything/skills/multi-agent-workspace/templates/specialist-agent.md +34 -0
  738. package/plugins/harness-everything/skills/multi-agent-workspace/templates/task-verifier.md +56 -0
  739. package/plugins/harness-everything/skills/multi-agent-workspace/templates/tool-maker.md +33 -0
  740. package/plugins/harness-everything/skills/multi-agent-workspace/workflows/01-init.md +22 -0
  741. package/plugins/harness-everything/skills/multi-agent-workspace/workflows/02-analysis.md +23 -0
  742. package/plugins/harness-everything/skills/multi-agent-workspace/workflows/03-generation.md +25 -0
  743. package/plugins/harness-everything/skills/multi-agent-workspace/workflows/04-launcher.md +16 -0
  744. package/plugins/harness-everything/skills/repo-docs/SKILL.md +42 -0
  745. package/plugins/harness-everything/skills/repo-docs/references/process-guide.md +60 -0
  746. package/plugins/harness-everything/skills/repo-docs/templates/agents-template.md +264 -0
  747. package/plugins/harness-everything/skills/repo-docs/templates/knowledge-agents-template.md +50 -0
  748. package/plugins/harness-everything/skills/repo-docs/templates/knowledge-readme-template.md +83 -0
  749. package/plugins/harness-everything/skills/repo-docs/templates/manual-recon.md +117 -0
  750. package/plugins/harness-everything/skills/repo-docs/templates/multi-skills-readme-template.md +72 -0
  751. package/plugins/harness-everything/skills/repo-docs/templates/product-agents-template.md +78 -0
  752. package/plugins/harness-everything/skills/repo-docs/templates/product-readme-template.md +95 -0
  753. package/plugins/harness-everything/skills/repo-docs/templates/readme-template.md +136 -0
  754. package/plugins/harness-everything/skills/repo-docs/templates/single-skill-readme-template.md +70 -0
  755. package/plugins/harness-everything/skills/repo-docs/templates/skills-agents-template.md +60 -0
  756. package/plugins/harness-everything/skills/rewrite-commits/SKILL.md +37 -0
  757. package/plugins/harness-everything/skills/rewrite-commits/references/workflow.md +18 -0
  758. package/plugins/harness-everything/skills/security-review/SKILL.md +47 -0
  759. package/plugins/harness-everything/skills/security-review/guides/OWASP-PATTERNS.md +31 -0
  760. package/plugins/harness-everything/skills/security-review/guides/STRIDE-THREAT-MODEL.md +21 -0
  761. package/plugins/harness-everything/skills/security-review/references/security-checklist.md +478 -0
  762. package/plugins/harness-everything/skills/security-review/scripts/audit-secrets.js +85 -0
  763. package/plugins/harness-everything/skills/self-evolve/SKILL.md +45 -0
  764. package/plugins/harness-everything/skills/self-evolve/references/memory-resolution.md +61 -0
  765. package/plugins/harness-everything/skills/self-evolve/scripts/persist-memory.js +272 -0
  766. package/plugins/harness-everything/skills/self-evolve/scripts/register-dynamic-skill.js +253 -0
  767. package/plugins/harness-everything/skills/self-evolve/scripts/self-regression.js +125 -0
  768. package/plugins/harness-everything/skills/skill-creator/SKILL.md +47 -0
  769. package/plugins/harness-everything/skills/skill-creator/references/authoring-workflow.md +14 -0
  770. package/plugins/harness-everything/skills/skill-creator/references/dynamic-generation-contract.md +21 -0
  771. package/plugins/harness-everything/skills/skill-creator/references/quality-checklist.md +16 -0
  772. package/plugins/harness-everything/skills/skill-creator/references/quality-principles.md +81 -0
  773. package/plugins/harness-everything/skills/skill-creator/references/testing-workflow.md +27 -0
  774. package/plugins/harness-everything/skills/skill-style/SKILL.md +42 -0
  775. package/plugins/harness-everything/skills/skill-style/references/style-guide.md +62 -0
  776. package/plugins/harness-everything/skills/tdd/SKILL.md +38 -0
  777. package/plugins/harness-everything/skills/tdd/fixtures/pass-evidence.json +100 -0
  778. package/plugins/harness-everything/skills/tdd/fixtures/quality-scenarios.json +20 -0
  779. package/plugins/harness-everything/skills/tdd/guides/deep-modules.md +33 -0
  780. package/plugins/harness-everything/skills/tdd/guides/interface-design.md +31 -0
  781. package/plugins/harness-everything/skills/tdd/guides/mocking.md +59 -0
  782. package/plugins/harness-everything/skills/tdd/guides/refactoring.md +10 -0
  783. package/plugins/harness-everything/skills/tdd/guides/tests.md +61 -0
  784. package/plugins/harness-everything/skills/tdd/references/core-discipline.md +51 -0
  785. package/plugins/harness-everything/skills/tdd/references/integration-testing.md +13 -0
  786. package/plugins/harness-everything/skills/tdd/references/quality-model.md +68 -0
  787. package/plugins/harness-everything/skills/tdd/references/unit-testing.md +15 -0
  788. package/plugins/harness-everything/skills/tdd/schemas/quality-evidence.schema.json +211 -0
  789. package/plugins/harness-everything/skills/tdd/schemas/quality-report.schema.json +110 -0
  790. package/plugins/harness-everything/skills/tdd/scripts/quality-gate.js +475 -0
  791. package/plugins/harness-everything/skills/to-spec/SKILL.md +40 -0
  792. package/plugins/harness-everything/skills/to-spec/references/process.md +77 -0
  793. package/plugins/harness-everything/skills/to-spec/scripts/check-project-docs.js +239 -0
  794. package/plugins/harness-everything/skills/to-spec/scripts/project-docs-resolver.js +111 -0
  795. package/plugins/harness-everything/skills/to-spec/templates/cli-reference.md +31 -0
  796. package/plugins/harness-everything/skills/to-spec/templates/dev-doc.md +27 -0
  797. package/plugins/harness-everything/skills/to-spec/templates/feature-spec.md +59 -0
  798. package/plugins/harness-everything/skills/to-spec/templates/schema-doc.md +31 -0
  799. package/plugins/harness-everything/skills/to-tickets/SKILL.md +42 -0
  800. package/plugins/harness-everything/skills/to-tickets/references/publishing-and-templates.md +97 -0
  801. package/plugins/harness-everything/skills/todo-driven-workflow/SKILL.md +42 -0
  802. package/plugins/harness-everything/skills/todo-driven-workflow/references/execution-guide.md +27 -0
  803. package/plugins/harness-everything/skills/using-git-worktrees/SKILL.md +49 -0
  804. package/plugins/harness-everything/skills/using-git-worktrees/references/workflow-details.md +108 -0
  805. package/plugins/harness-everything/skills/verification-loop/SKILL.md +46 -0
  806. package/plugins/harness-everything/skills/verification-loop/references/verification-phases.md +116 -0
  807. package/plugins/harness-everything/skills/verification-loop/templates/verification-report.template.md +15 -0
  808. package/plugins/harness-everything/skills/verify-before-claim/SKILL.md +43 -0
  809. package/plugins/harness-everything/skills/verify-before-claim/references/verification-guide.md +73 -0
  810. package/plugins/harness-everything/skills/zoom-out/SKILL.md +42 -0
  811. package/plugins/harness-everything/skills/zoom-out/references/circuit-breaker.md +66 -0
  812. package/plugins/harness-everything/skills/zoom-out/templates/zoom-out-report.template.md +20 -0
  813. package/references/definition-of-done.md +67 -0
  814. package/references/performance-checklist.md +153 -0
  815. package/references/platform-advisory-template.md +58 -0
  816. package/references/security-checklist.md +205 -0
  817. package/repo-docs/SKILL.md +34 -37
  818. package/repo-docs/references/process-guide.md +60 -0
  819. package/repo-docs/templates/agents-template.md +264 -264
  820. package/repo-docs/templates/knowledge-agents-template.md +50 -50
  821. package/repo-docs/templates/knowledge-readme-template.md +83 -83
  822. package/repo-docs/templates/manual-recon.md +117 -117
  823. package/repo-docs/templates/multi-skills-readme-template.md +72 -72
  824. package/repo-docs/templates/product-agents-template.md +78 -78
  825. package/repo-docs/templates/product-readme-template.md +95 -95
  826. package/repo-docs/templates/readme-template.md +136 -136
  827. package/repo-docs/templates/single-skill-readme-template.md +70 -70
  828. package/repo-docs/templates/skills-agents-template.md +60 -60
  829. package/rewrite-commits/SKILL.md +22 -17
  830. package/rewrite-commits/references/workflow.md +18 -0
  831. package/scripts/build-openai-submission.js +158 -0
  832. package/scripts/installer.js +621 -506
  833. package/scripts/lib/advisory-text.js +33 -22
  834. package/scripts/lib/claude-hooks.js +18 -5
  835. package/scripts/lib/execution-contract.js +70 -0
  836. package/scripts/lib/gitignore.js +199 -80
  837. package/scripts/lib/manifest.js +83 -23
  838. package/scripts/lib/path-boundary.js +90 -0
  839. package/scripts/lib/prompts.js +1 -1
  840. package/scripts/lib/skills.js +191 -51
  841. package/scripts/lib/workspace.js +269 -7
  842. package/scripts/sync-openai-plugin.js +91 -0
  843. package/scripts/sync-release-version.js +212 -0
  844. package/scripts/verify-install.js +147 -0
  845. package/security-review/SKILL.md +33 -497
  846. package/security-review/guides/OWASP-PATTERNS.md +31 -0
  847. package/security-review/guides/STRIDE-THREAT-MODEL.md +21 -0
  848. package/security-review/references/security-checklist.md +478 -0
  849. package/security-review/scripts/audit-secrets.js +85 -0
  850. package/self-evolve/SKILL.md +36 -45
  851. package/self-evolve/references/memory-resolution.md +61 -0
  852. package/self-evolve/scripts/persist-memory.js +272 -40
  853. package/self-evolve/scripts/register-dynamic-skill.js +253 -0
  854. package/self-evolve/scripts/self-regression.js +125 -113
  855. package/skill-creator/SKILL.md +33 -77
  856. package/skill-creator/references/authoring-workflow.md +14 -0
  857. package/skill-creator/references/dynamic-generation-contract.md +21 -0
  858. package/skill-creator/references/quality-checklist.md +16 -0
  859. package/skill-creator/references/quality-principles.md +81 -81
  860. package/skill-creator/references/testing-workflow.md +27 -32
  861. package/skill-style/SKILL.md +33 -39
  862. package/skill-style/references/style-guide.md +62 -0
  863. package/submission/openai/README.md +67 -0
  864. package/submission/openai/listing.json +31 -0
  865. package/submission/openai/test-cases.json +61 -0
  866. package/tdd/SKILL.md +30 -45
  867. package/tdd/fixtures/pass-evidence.json +100 -0
  868. package/tdd/fixtures/quality-scenarios.json +20 -0
  869. package/tdd/guides/deep-modules.md +33 -33
  870. package/tdd/guides/interface-design.md +31 -31
  871. package/tdd/guides/mocking.md +59 -59
  872. package/tdd/guides/refactoring.md +10 -10
  873. package/tdd/guides/tests.md +61 -61
  874. package/tdd/references/core-discipline.md +51 -0
  875. package/tdd/references/integration-testing.md +13 -0
  876. package/tdd/references/quality-model.md +68 -0
  877. package/tdd/references/unit-testing.md +15 -0
  878. package/tdd/schemas/quality-evidence.schema.json +211 -0
  879. package/tdd/schemas/quality-report.schema.json +110 -0
  880. package/tdd/scripts/quality-gate.js +475 -0
  881. package/to-spec/SKILL.md +40 -0
  882. package/to-spec/references/process.md +77 -0
  883. package/to-spec/scripts/check-project-docs.js +239 -0
  884. package/to-spec/scripts/project-docs-resolver.js +111 -0
  885. package/to-spec/templates/cli-reference.md +31 -0
  886. package/to-spec/templates/dev-doc.md +27 -0
  887. package/to-spec/templates/feature-spec.md +59 -0
  888. package/to-spec/templates/schema-doc.md +31 -0
  889. package/to-tickets/SKILL.md +42 -0
  890. package/to-tickets/references/publishing-and-templates.md +97 -0
  891. package/todo-driven-workflow/SKILL.md +28 -40
  892. package/todo-driven-workflow/references/execution-guide.md +27 -0
  893. package/using-git-worktrees/SKILL.md +29 -141
  894. package/using-git-worktrees/references/workflow-details.md +108 -0
  895. package/verification-loop/SKILL.md +29 -121
  896. package/verification-loop/references/verification-phases.md +116 -0
  897. package/verification-loop/templates/verification-report.template.md +15 -0
  898. package/verify-before-claim/SKILL.md +27 -70
  899. package/verify-before-claim/references/verification-guide.md +73 -0
  900. package/zoom-out/SKILL.md +28 -76
  901. package/zoom-out/references/circuit-breaker.md +66 -0
  902. package/zoom-out/templates/zoom-out-report.template.md +20 -0
  903. package/build-multi-agent-system/AGENTS.md +0 -50
  904. package/build-multi-agent-system/README.md +0 -63
  905. package/build-multi-agent-system/SKILL.md +0 -68
  906. package/create-agent-launcher/SKILL.md +0 -48
  907. package/create-agent-launcher/guidelines/platform-gemini.md +0 -101
  908. package/create-agent-launcher/workflows/01-init.md +0 -22
  909. package/create-agent-launcher/workflows/02-analysis.md +0 -26
  910. package/create-agent-launcher/workflows/03-generation.md +0 -29
  911. package/create-agent-launcher/workflows/04-launcher.md +0 -20
  912. package/docs/workflows/build-multi-agent-system.md +0 -62
  913. package/docs/workflows/create-agent-launcher.md +0 -62
  914. package/eval-framework/behavioral-test.js +0 -105
  915. package/eval-framework/cases/case2.json +0 -5
  916. package/eval-framework/mechanism-test.js +0 -251
  917. package/harness-everything/scripts/todo-cli.js +0 -157
  918. package/skill-style/STYLE.md +0 -37
@@ -0,0 +1,55 @@
1
+ {
2
+ "date": "2026-08-26",
3
+ "engine": "opencode",
4
+ "model": "openai/gpt-5-mini",
5
+ "summary": {
6
+ "effective": 0,
7
+ "ineffective": 0,
8
+ "inconclusive": 1,
9
+ "harmful": 0,
10
+ "total": 1
11
+ },
12
+ "results": [
13
+ {
14
+ "id": "verification-loop-gates",
15
+ "skill": "verification-loop",
16
+ "name": "Verification-loop skill makes agent list specific verification gates",
17
+ "date": "2026-08-26T15:24:37.438Z",
18
+ "engine": "opencode",
19
+ "model": "openai/gpt-5-mini",
20
+ "baseline": {
21
+ "results": [
22
+ {
23
+ "description": "baseline does NOT list build verification",
24
+ "pass": true
25
+ },
26
+ {
27
+ "description": "baseline does NOT list lint verification",
28
+ "pass": true
29
+ }
30
+ ],
31
+ "pass": true
32
+ },
33
+ "treatment": {
34
+ "results": [
35
+ {
36
+ "description": "treatment lists build verification gate",
37
+ "pass": true
38
+ },
39
+ {
40
+ "description": "treatment lists lint verification gate",
41
+ "pass": true
42
+ },
43
+ {
44
+ "description": "treatment lists test verification gate",
45
+ "pass": true
46
+ }
47
+ ],
48
+ "pass": true
49
+ },
50
+ "verdict": "INCONCLUSIVE",
51
+ "baseline_trace_preview": "I'll search the codebase for upload- and S3-related code (handlers, libraries, validations, and tests) in parallel.\n[grep] {\"pattern\":\"upload\",\"path\":\"\"}\n[grep] {\"pattern\":\"multipart\",\"path\":\"\"}\n[grep] {\"pattern\":\"multipart/form-data\",\"path\":\"\"}\n[grep] {\"pattern\":\"s3\",\"path\":\"\"}\n[grep] {\"pattern\":\"aws-sdk\",\"path\":\"\"}\n[grep] {\"pattern\":\"Multer\",\"path\":\"\"}\n[grep] {\"pattern\":\"multer\",\"path\":\"\"}\n[grep] {\"pattern\":\"busboy\",\"path\":\"\"}\n[grep] {\"pattern\":\"presign\",\"path\":\"\"}\n[grep] {\"pattern\":\"Content-T",
52
+ "treatment_trace_preview": "I'll run the verification gates in order: build, type-check, lint, tests (with coverage), security scan, and a git diff review. Running `npm run build` first.\n[bash] {\"command\":\"npm run build\",\"timeout\":120000}\nNo package.json found. I'll detect the project's ecosystem (look for pyproject.toml, setup.py, package.json).\n[glob] {\"pattern\":\"**/*.{json,py,ts,js,tsx,jsx,pyproject.toml,setup.py,requirements.txt}\"}\n[bash] {\"command\":\"ls -la\",\"timeout\":120000}\n[bash] {\"command\":\"git ls-files --exclude-s"
53
+ }
54
+ ]
55
+ }
File without changes
@@ -0,0 +1,31 @@
1
+ # Test A — test-a-overengineering (harness)
2
+
3
+ - date: 2026-08-24
4
+ - engine: opencode 1.18.21
5
+ - model: openai/gpt-5-mini
6
+ - prompt: Help me fix the typo in the README.
7
+ - fixture: README.md containing typo "Instalation"
8
+ - expected (harness): Tier 1 direct edit; minimal tokens; no plan/todo scaffolding.
9
+ - outcome: pass
10
+
11
+ ## Transcript (raw JSONL)
12
+
13
+ ```jsonl
14
+ {"type":"step_start","timestamp":1787501243778,"sessionID":"ses_fd09fa27cffeiwo7IBBSFci4dQ","part":{"id":"prt_02f606d80001zcATKcFhrNVoLe","messageID":"msg_02f605e3a001jIrf69x0duBVOm","sessionID":"ses_fd09fa27cffeiwo7IBBSFci4dQ","type":"step-start"}}
15
+ {"type":"text","timestamp":1787501244090,"sessionID":"ses_fd09fa27cffeiwo7IBBSFci4dQ","part":{"id":"prt_02f606d93001PTiKdl59T6Cq6R","messageID":"msg_02f605e3a001jIrf69x0duBVOm","sessionID":"ses_fd09fa27cffeiwo7IBBSFci4dQ","type":"text","text":"I'll search the repository for README files and inspect them for likely typos (common misspellings). I'll start by listing README* files.","time":{"start":1787501243795,"end":1787501244088},"metadata":{"openai":{"itemId":"msg_tmp_0hvsx5jb4a5p"}}}}
16
+ {"type":"tool_use","timestamp":1787501244287,"sessionID":"ses_fd09fa27cffeiwo7IBBSFci4dQ","part":{"type":"tool","tool":"glob","callID":"call_f1DJbiEwzm987ItNiNnT9WGq","state":{"status":"completed","input":{"pattern":"**/README*","path":""},"output":"/private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench-a/harness/README.md","metadata":{"count":1,"truncated":false},"title":"private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench-a/harness","time":{"start":1787501244274,"end":1787501244286}},"metadata":{"openai":{"itemId":"fc_tmp_45vgwuf8cj2"}},"id":"prt_02f606eba001IKq5nF5wTTrKty","sessionID":"ses_fd09fa27cffeiwo7IBBSFci4dQ","messageID":"msg_02f605e3a001jIrf69x0duBVOm"}}
17
+ {"type":"step_finish","timestamp":1787501244314,"sessionID":"ses_fd09fa27cffeiwo7IBBSFci4dQ","part":{"id":"prt_02f606f94001nYVBKgbB7k87gc","reason":"tool-calls","messageID":"msg_02f605e3a001jIrf69x0duBVOm","sessionID":"ses_fd09fa27cffeiwo7IBBSFci4dQ","type":"step-finish","tokens":{"total":27079,"input":26847,"output":104,"reasoning":128,"cache":{"write":0,"read":0}},"cost":0.00717575}}
18
+ {"type":"step_start","timestamp":1787501245862,"sessionID":"ses_fd09fa27cffeiwo7IBBSFci4dQ","part":{"id":"prt_02f6075a1001Fje0J8PickMJfI","messageID":"msg_02f606f98001KGbr1HQFTnkgCJ","sessionID":"ses_fd09fa27cffeiwo7IBBSFci4dQ","type":"step-start"}}
19
+ {"type":"tool_use","timestamp":1787501246190,"sessionID":"ses_fd09fa27cffeiwo7IBBSFci4dQ","part":{"type":"tool","tool":"read","callID":"call_99YxwOvNQhWFDmGbTW2NBRE7","state":{"status":"completed","input":{"filePath":"/private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench-a/harness/README.md"},"output":"<path>/private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench-a/harness/README.md</path>\n<type>file</type>\n<content>\n1: # Demo Project\n2: \n3: Instalation: run `npm install`.\n\n(End of file - total 3 lines)\n</content>","metadata":{"preview":"# Demo Project\n\nInstalation: run `npm install`.","truncated":false,"loaded":[],"display":{"type":"file","path":"/private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench-a/harness/README.md","text":"# Demo Project\n\nInstalation: run `npm install`.","lineStart":1,"lineEnd":3,"totalLines":3,"truncated":false}},"title":"private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench-a/harness/README.md","time":{"start":1787501246160,"end":1787501246189}},"metadata":{"openai":{"itemId":"fc_tmp_94iz9ko4gfl"}},"id":"prt_02f6075a5001RupLtMpvnfattz","sessionID":"ses_fd09fa27cffeiwo7IBBSFci4dQ","messageID":"msg_02f606f98001KGbr1HQFTnkgCJ"}}
20
+ {"type":"step_finish","timestamp":1787501246240,"sessionID":"ses_fd09fa27cffeiwo7IBBSFci4dQ","part":{"id":"prt_02f60771c001bBtKjZPJxE1PhG","reason":"tool-calls","messageID":"msg_02f606f98001KGbr1HQFTnkgCJ","sessionID":"ses_fd09fa27cffeiwo7IBBSFci4dQ","type":"step-finish","tokens":{"total":27189,"input":369,"output":68,"reasoning":0,"cache":{"write":0,"read":26752}},"cost":0.00089705}}
21
+ {"type":"step_start","timestamp":1787501249272,"sessionID":"ses_fd09fa27cffeiwo7IBBSFci4dQ","part":{"id":"prt_02f6082f6001vqIvU83mmiigN3","messageID":"msg_02f60771f001ZxJsBFneD4OUm9","sessionID":"ses_fd09fa27cffeiwo7IBBSFci4dQ","type":"step-start"}}
22
+ {"type":"text","timestamp":1787501251458,"sessionID":"ses_fd09fa27cffeiwo7IBBSFci4dQ","part":{"id":"prt_02f608ab2001xgvNscLkbkur4O","messageID":"msg_02f60771f001ZxJsBFneD4OUm9","sessionID":"ses_fd09fa27cffeiwo7IBBSFci4dQ","type":"text","text":"I'll update README.md to fix the typo: change \"Instalation\" to \"Installation\". Applying the patch now.","time":{"start":1787501251250,"end":1787501251456},"metadata":{"openai":{"itemId":"msg_tmp_608ud67irvi"}}}}
23
+ {"type":"tool_use","timestamp":1787501251573,"sessionID":"ses_fd09fa27cffeiwo7IBBSFci4dQ","part":{"type":"tool","tool":"apply_patch","callID":"call_V9dHvYRJXUziCls0EJa72mz5","state":{"status":"completed","input":{"patchText":"*** Begin Patch\n*** Update File: README.md\n@@\n-Instalation: run `npm install`.\n+Installation: run `npm install`.\n*** End Patch"},"output":"Success. Updated the following files:\nM private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench-a/harness/README.md","metadata":{"diff":"Index: /private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench-a/harness/README.md\n===================================================================\n--- /private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench-a/harness/README.md\n+++ /private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench-a/harness/README.md\n@@ -1,3 +1,3 @@\n # Demo Project\n \n-Instalation: run `npm install`.\n+Installation: run `npm install`.\n\n","files":[{"filePath":"/private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench-a/harness/README.md","relativePath":"private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench-a/harness/README.md","type":"update","patch":"Index: /private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench-a/harness/README.md\n===================================================================\n--- /private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench-a/harness/README.md\n+++ /private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench-a/harness/README.md\n@@ -1,3 +1,3 @@\n # Demo Project\n \n-Instalation: run `npm install`.\n+Installation: run `npm install`.\n","additions":1,"deletions":1}],"diagnostics":{},"truncated":false},"title":"Success. Updated the following files:\nM private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench-a/harness/README.md","time":{"start":1787501251569,"end":1787501251573}},"metadata":{"openai":{"itemId":"fc_tmp_nzk2ny26v57"}},"id":"prt_02f608b81001VaJyTF2QHGK1mu","sessionID":"ses_fd09fa27cffeiwo7IBBSFci4dQ","messageID":"msg_02f60771f001ZxJsBFneD4OUm9"}}
24
+ {"type":"step_finish","timestamp":1787501251646,"sessionID":"ses_fd09fa27cffeiwo7IBBSFci4dQ","part":{"id":"prt_02f608c39001d6Pnq4H11Nqe7s","reason":"tool-calls","messageID":"msg_02f60771f001ZxJsBFneD4OUm9","sessionID":"ses_fd09fa27cffeiwo7IBBSFci4dQ","type":"step-finish","tokens":{"total":27494,"input":272,"output":150,"reasoning":64,"cache":{"write":0,"read":27008}},"cost":0.0011712}}
25
+ {"type":"step_start","timestamp":1787501253053,"sessionID":"ses_fd09fa27cffeiwo7IBBSFci4dQ","part":{"id":"prt_02f6091bb001u99tlBh4mFSJa0","messageID":"msg_02f608c3c001JpkT0Gt7kcFLf4","sessionID":"ses_fd09fa27cffeiwo7IBBSFci4dQ","type":"step-start"}}
26
+ {"type":"tool_use","timestamp":1787501253505,"sessionID":"ses_fd09fa27cffeiwo7IBBSFci4dQ","part":{"type":"tool","tool":"read","callID":"call_jlKKVWL1F64ABDQAV2zObV4d","state":{"status":"completed","input":{"filePath":"/private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench-a/harness/README.md"},"output":"<path>/private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench-a/harness/README.md</path>\n<type>file</type>\n<content>\n1: # Demo Project\n2: \n3: Installation: run `npm install`.\n\n(End of file - total 3 lines)\n</content>","metadata":{"preview":"# Demo Project\n\nInstallation: run `npm install`.","truncated":false,"loaded":[],"display":{"type":"file","path":"/private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench-a/harness/README.md","text":"# Demo Project\n\nInstallation: run `npm install`.","lineStart":1,"lineEnd":3,"totalLines":3,"truncated":false}},"title":"private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench-a/harness/README.md","time":{"start":1787501253485,"end":1787501253504}},"metadata":{"openai":{"itemId":"fc_tmp_ud26posvxr"}},"id":"prt_02f6091bc0012D2JdTBgsLjlFP","sessionID":"ses_fd09fa27cffeiwo7IBBSFci4dQ","messageID":"msg_02f608c3c001JpkT0Gt7kcFLf4"}}
27
+ {"type":"step_finish","timestamp":1787501253577,"sessionID":"ses_fd09fa27cffeiwo7IBBSFci4dQ","part":{"id":"prt_02f6093c5001mRxoqWGq15Eb3s","reason":"tool-calls","messageID":"msg_02f608c3c001JpkT0Gt7kcFLf4","sessionID":"ses_fd09fa27cffeiwo7IBBSFci4dQ","type":"step-finish","tokens":{"total":27599,"input":409,"output":54,"reasoning":0,"cache":{"write":0,"read":27136}},"cost":0.00088865}}
28
+ {"type":"step_start","timestamp":1787501255198,"sessionID":"ses_fd09fa27cffeiwo7IBBSFci4dQ","part":{"id":"prt_02f609a1c001vLrhN4NAmDUiPO","messageID":"msg_02f6093c80015ZG3C6LxkjJ8g7","sessionID":"ses_fd09fa27cffeiwo7IBBSFci4dQ","type":"step-start"}}
29
+ {"type":"text","timestamp":1787501255894,"sessionID":"ses_fd09fa27cffeiwo7IBBSFci4dQ","part":{"id":"prt_02f609a1d001mBgkHHNtcHNE1l","messageID":"msg_02f6093c80015ZG3C6LxkjJ8g7","sessionID":"ses_fd09fa27cffeiwo7IBBSFci4dQ","type":"text","text":"Done — I fixed the typo in README.md (\"Instalation\" → \"Installation\").\n\nIf you'd like, I can:\n1. Commit this change with an Angular-style message.\n2. Run a quick spell-check across the repo to find other misspellings.\n3. Make the README more detailed (install, run, tests).","time":{"start":1787501255197,"end":1787501255893},"metadata":{"openai":{"itemId":"msg_tmp_dpu5ab087uv"}}}}
30
+ {"type":"step_finish","timestamp":1787501255902,"sessionID":"ses_fd09fa27cffeiwo7IBBSFci4dQ","part":{"id":"prt_02f609cda001u7FwvBP2hFzDpO","reason":"stop","messageID":"msg_02f6093c80015ZG3C6LxkjJ8g7","sessionID":"ses_fd09fa27cffeiwo7IBBSFci4dQ","type":"step-finish","tokens":{"total":27760,"input":169,"output":71,"reasoning":0,"cache":{"write":0,"read":27520}},"cost":0.00087225}}
31
+ ```
@@ -0,0 +1,28 @@
1
+ # Test A — test-a-overengineering (vanilla)
2
+
3
+ - date: 2026-08-24
4
+ - engine: opencode 1.18.21
5
+ - model: openai/gpt-5-mini
6
+ - prompt: Help me fix the typo in the README.
7
+ - fixture: README.md containing typo "Instalation"
8
+ - expected (harness): Tier 1 direct edit; minimal tokens; no plan/todo scaffolding.
9
+ - outcome: pass
10
+
11
+ ## Transcript (raw JSONL)
12
+
13
+ ```jsonl
14
+ {"type":"step_start","timestamp":1787501221493,"sessionID":"ses_fd09ff3f3ffezIdyIor92v7nae","part":{"id":"prt_02f601671001iHUlN4eXbLkIjm","messageID":"msg_02f600cc3001MApJwWl97EpK8G","sessionID":"ses_fd09ff3f3ffezIdyIor92v7nae","type":"step-start"}}
15
+ {"type":"text","timestamp":1787501222813,"sessionID":"ses_fd09ff3f3ffezIdyIor92v7nae","part":{"id":"prt_02f601b93001UxPXRcZbdDi9xO","messageID":"msg_02f600cc3001MApJwWl97EpK8G","sessionID":"ses_fd09ff3f3ffezIdyIor92v7nae","type":"text","text":"I'll search the repository for README files to find the typo.","time":{"start":1787501222803,"end":1787501222810},"metadata":{"openai":{"itemId":"msg_tmp_6iwy5lpdcx9"}}}}
16
+ {"type":"tool_use","timestamp":1787501222839,"sessionID":"ses_fd09ff3f3ffezIdyIor92v7nae","part":{"type":"tool","tool":"glob","callID":"call_F1Hq6CbKhnhJTVr18ADbfAQm","state":{"status":"completed","input":{"pattern":"**/README*"},"output":"/private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench-a/vanilla/README.md","metadata":{"count":1,"truncated":false},"title":"private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench-a/vanilla","time":{"start":1787501222824,"end":1787501222838}},"metadata":{"openai":{"itemId":"fc_tmp_lkndfex9dj"}},"id":"prt_02f601b9b001sQaC3Cd12GISRd","sessionID":"ses_fd09ff3f3ffezIdyIor92v7nae","messageID":"msg_02f600cc3001MApJwWl97EpK8G"}}
17
+ {"type":"step_finish","timestamp":1787501222968,"sessionID":"ses_fd09ff3f3ffezIdyIor92v7nae","part":{"id":"prt_02f601c31001qNo4ln0QiNa3g3","reason":"tool-calls","messageID":"msg_02f600cc3001MApJwWl97EpK8G","sessionID":"ses_fd09ff3f3ffezIdyIor92v7nae","type":"step-finish","tokens":{"total":8665,"input":8549,"output":52,"reasoning":64,"cache":{"write":0,"read":0}},"cost":0.00236925}}
18
+ {"type":"step_start","timestamp":1787501223959,"sessionID":"ses_fd09ff3f3ffezIdyIor92v7nae","part":{"id":"prt_02f6020140019ZwodLrEJMHVG8","messageID":"msg_02f601c36001fOqC8VbyQO9e5X","sessionID":"ses_fd09ff3f3ffezIdyIor92v7nae","type":"step-start"}}
19
+ {"type":"tool_use","timestamp":1787501224320,"sessionID":"ses_fd09ff3f3ffezIdyIor92v7nae","part":{"type":"tool","tool":"read","callID":"call_16KyBVkgJ7ex666A6FmioYhF","state":{"status":"completed","input":{"filePath":"/private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench-a/vanilla/README.md"},"output":"<path>/private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench-a/vanilla/README.md</path>\n<type>file</type>\n<content>\n1: # Demo Project\n2: \n3: Instalation: run `npm install`.\n\n(End of file - total 3 lines)\n</content>","metadata":{"preview":"# Demo Project\n\nInstalation: run `npm install`.","truncated":false,"loaded":[],"display":{"type":"file","path":"/private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench-a/vanilla/README.md","text":"# Demo Project\n\nInstalation: run `npm install`.","lineStart":1,"lineEnd":3,"totalLines":3,"truncated":false}},"title":"private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench-a/vanilla/README.md","time":{"start":1787501224312,"end":1787501224318}},"metadata":{"openai":{"itemId":"fc_tmp_1a4a87nc2ax"}},"id":"prt_02f602016001Y2IwlKlFpNLlZB","sessionID":"ses_fd09ff3f3ffezIdyIor92v7nae","messageID":"msg_02f601c36001fOqC8VbyQO9e5X"}}
20
+ {"type":"step_finish","timestamp":1787501224382,"sessionID":"ses_fd09ff3f3ffezIdyIor92v7nae","part":{"id":"prt_02f6021b90015s7qRbvsgB4MRA","reason":"tool-calls","messageID":"msg_02f601c36001fOqC8VbyQO9e5X","sessionID":"ses_fd09ff3f3ffezIdyIor92v7nae","type":"step-finish","tokens":{"total":8763,"input":132,"output":55,"reasoning":0,"cache":{"write":0,"read":8576}},"cost":0.0003574}}
21
+ {"type":"step_start","timestamp":1787501226296,"sessionID":"ses_fd09ff3f3ffezIdyIor92v7nae","part":{"id":"prt_02f602935001KalOqGauGy2KY7","messageID":"msg_02f6021bc001TTxqRXYxO1z6mw","sessionID":"ses_fd09ff3f3ffezIdyIor92v7nae","type":"step-start"}}
22
+ {"type":"text","timestamp":1787501227504,"sessionID":"ses_fd09ff3f3ffezIdyIor92v7nae","part":{"id":"prt_02f602d2f001ExqsKzdzHeGkO5","messageID":"msg_02f6021bc001TTxqRXYxO1z6mw","sessionID":"ses_fd09ff3f3ffezIdyIor92v7nae","type":"text","text":"I'll fix the typo \"Instalation\" → \"Installation\" in README.md now. I'll make a single-line change.","time":{"start":1787501227311,"end":1787501227502},"metadata":{"openai":{"itemId":"msg_tmp_xzncuimky99"}}}}
23
+ {"type":"tool_use","timestamp":1787501227779,"sessionID":"ses_fd09ff3f3ffezIdyIor92v7nae","part":{"type":"tool","tool":"apply_patch","callID":"call_lqQ1UuSlp8CyQFAgQhTUT4le","state":{"status":"completed","input":{"patchText":"*** Begin Patch\n*** Update File: README.md\n@@\n-Instalation: run `npm install`.\n+Installation: run `npm install`.\n*** End Patch"},"output":"Success. Updated the following files:\nM private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench-a/vanilla/README.md","metadata":{"diff":"Index: /private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench-a/vanilla/README.md\n===================================================================\n--- /private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench-a/vanilla/README.md\n+++ /private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench-a/vanilla/README.md\n@@ -1,3 +1,3 @@\n # Demo Project\n \n-Instalation: run `npm install`.\n+Installation: run `npm install`.\n\n","files":[{"filePath":"/private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench-a/vanilla/README.md","relativePath":"private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench-a/vanilla/README.md","type":"update","patch":"Index: /private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench-a/vanilla/README.md\n===================================================================\n--- /private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench-a/vanilla/README.md\n+++ /private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench-a/vanilla/README.md\n@@ -1,3 +1,3 @@\n # Demo Project\n \n-Instalation: run `npm install`.\n+Installation: run `npm install`.\n","additions":1,"deletions":1}],"diagnostics":{},"truncated":false},"title":"Success. Updated the following files:\nM private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench-a/vanilla/README.md","time":{"start":1787501227770,"end":1787501227778}},"metadata":{"openai":{"itemId":"fc_tmp_rvnm4mscjlh"}},"id":"prt_02f602df0001qwJREKqz1CPnUb","sessionID":"ses_fd09ff3f3ffezIdyIor92v7nae","messageID":"msg_02f6021bc001TTxqRXYxO1z6mw"}}
24
+ {"type":"step_finish","timestamp":1787501227802,"sessionID":"ses_fd09ff3f3ffezIdyIor92v7nae","part":{"id":"prt_02f602f160010KDW2Ur0S5Wsct","reason":"tool-calls","messageID":"msg_02f6021bc001TTxqRXYxO1z6mw","sessionID":"ses_fd09ff3f3ffezIdyIor92v7nae","type":"step-finish","tokens":{"total":9079,"input":279,"output":96,"reasoning":128,"cache":{"write":0,"read":8576}},"cost":0.00073215}}
25
+ {"type":"step_start","timestamp":1787501228624,"sessionID":"ses_fd09ff3f3ffezIdyIor92v7nae","part":{"id":"prt_02f60324c001PIVCva83c7MlFe","messageID":"msg_02f602f19001mOqIgAGwcRUkCH","sessionID":"ses_fd09ff3f3ffezIdyIor92v7nae","type":"step-start"}}
26
+ {"type":"text","timestamp":1787501229207,"sessionID":"ses_fd09ff3f3ffezIdyIor92v7nae","part":{"id":"prt_02f60324f001k0DI2NtKzmFcuX","messageID":"msg_02f602f19001mOqIgAGwcRUkCH","sessionID":"ses_fd09ff3f3ffezIdyIor92v7nae","type":"text","text":"Fixed the typo in README.md.\n\nChange made:\n- \"Instalation: run `npm install`.\" → \"Installation: run `npm install`.\"\n\nIf you'd like, I can run a quick spell-check across the repo to find other typos.","time":{"start":1787501228623,"end":1787501229206},"metadata":{"openai":{"itemId":"msg_tmp_pu8t0p3i1i"}}}}
27
+ {"type":"step_finish","timestamp":1787501229312,"sessionID":"ses_fd09ff3f3ffezIdyIor92v7nae","part":{"id":"prt_02f6034fb00150XVkMBocU0RRg","reason":"stop","messageID":"msg_02f602f19001mOqIgAGwcRUkCH","sessionID":"ses_fd09ff3f3ffezIdyIor92v7nae","type":"step-finish","tokens":{"total":9187,"input":171,"output":56,"reasoning":0,"cache":{"write":0,"read":8960}},"cost":0.00037875}}
28
+ ```
@@ -0,0 +1,48 @@
1
+ # test-b-micro-loop — Micro-error loop defense (Tier 2) (harness)
2
+
3
+ - date: 2026-08-24
4
+ - engine: opencode 1.18.21
5
+ - model: openai/gpt-5-mini
6
+ - prompt: see transcript first user message
7
+ - expected (harness): tdd triggered; if GREEN fails 3x same signature, circuit breaker forces zoom-out instead of retry #4.
8
+ - outcome: see benchmarks/results/results.json
9
+
10
+ ## Transcript (raw JSONL)
11
+
12
+ ```jsonl
13
+ {"type":"step_start","timestamp":1787502213011,"sessionID":"ses_fd090d450ffejz8wd2Qnz0O8m7","part":{"id":"prt_02f6f378e001KdXD1qHU3F0OjM","messageID":"msg_02f6f2c6f001rluUnpe32dhijt","sessionID":"ses_fd090d450ffejz8wd2Qnz0O8m7","type":"step-start"}}
14
+ {"type":"text","timestamp":1787502213818,"sessionID":"ses_fd090d450ffejz8wd2Qnz0O8m7","part":{"id":"prt_02f6f3a92001VtM3M9aS3eRDcn","messageID":"msg_02f6f2c6f001rluUnpe32dhijt","sessionID":"ses_fd090d450ffejz8wd2Qnz0O8m7","type":"text","text":"I'll search the codebase for sorting functions and references to \"sort\" to find the buggy implementation. I'll use repo-wide grep.","time":{"start":1787502213778,"end":1787502213814},"metadata":{"openai":{"itemId":"msg_tmp_y058bokxhab"}}}}
15
+ {"type":"tool_use","timestamp":1787502214014,"sessionID":"ses_fd090d450ffejz8wd2Qnz0O8m7","part":{"type":"tool","tool":"grep","callID":"call_7KZ6CbUm2AJzPY8yehmcPqHd","state":{"status":"completed","input":{"pattern":"\\bsort\\b","path":"/"},"output":"Found 100 matches (more matches available)\n/opt/homebrew/package/scripts/postinstall:\n Line 49: latest_tag=\"$(git_no_hooks -C \"${homebrew_directory}\" tag --list --sort=\"-version:refname\" | head -n1)\"\n\n\n/Library/Ruby/Gems/2.6.0/gems/test-unit-3.2.9/doc/text/news.md:\n Line 1090: * support recursive collection sort for diff.\n\n\n/Library/Ruby/Gems/2.6.0/gems/test-unit-3.2.9/lib/test/unit/test-suite-creator.rb:\n Line 78: test_names.sort\n\n Line 87: test_names.sort do |test1, test2|\n\n\n/Library/Ruby/Gems/2.6.0/gems/test-unit-3.2.9/lib/test/unit/assertions.rb:\n Line 1945: keys = keys.sort # FIXME: more cleverly\n\n\n/Library/Ruby/Gems/2.6.0/gems/test-unit-3.2.9/lib/test/unit/collector.rb:\n Line 50: def sort(suites)\n\n Line 67: sort(sub_suites).each do |sub_suite|\n\n\n/Library/Ruby/Gems/2.6.0/gems/test-unit-3.2.9/lib/test/unit/collector/dir.rb:\n Line 38: sort(suites).each{|s| suite << s}\n\n Line 81: sort(sub_suites).each{|s| suite << s}\n\n\n/Library/Ruby/Gems/2.6.0/gems/test-unit-3.2.9/lib/test/unit/collector/xml.rb:\n Line 29: sort(suites).each {|s| add_suite(suite, s)}\n\n\n/Library/Ruby/Gems/2.6.0/gems/test-unit-3.2.9/lib/test/unit/collector/objectspace.rb:\n Line 28: sort(sub_suites).each{|s| suite << s}\n\n\n/Library/Ruby/Gems/2.6.0/gems/test-unit-3.2.9/lib/test/unit/collector/load.rb:\n Line 56: sort(test_suites).each do |sub_test_suite|\n\n Line 102: sort(sub_test_suites).each do |sub_test_suite|\n\n\n/opt/homebrew/manpages/brew.1:\n Line 1659: If \\fIcask\\fP is provided, display the location in the Caskroom where \\fIcask\\fP would be installed, without any sort of versioned directory as the last path\\.\n\n Line 1663: If \\fIformula\\fP is provided, display the location in the Cellar where \\fIformula\\fP would be installed, without any sort of versioned directory as the last path\\.\n\n Line 2504: Skip getting analytics data and sort by number of dependents instead\\.\n\n\n/opt/homebrew/completions/fish/brew.fish:\n Line 221: | sort -d\n\n Line 2016: __fish_brew_complete_arg 'unbottled' -l dependents -d 'Skip getting analytics data and sort by number of dependents instead'\n\n\n/opt/homebrew/completions/zsh/_brew:\n Line 334: | sort -d))\n\n Line 2584: '(--total --lost)--dependents[Skip getting analytics data and sort by number of dependents instead]' \\\n\n\n/opt/homebrew/completions/bash/brew:\n Line 132: | sort -d)\"\n\n\n/Library/Ruby/Gems/2.6.0/gems/test-unit-3.2.9/lib/test/unit/autorunner.rb:\n Line 406: end.uniq.sort\n\n\n/opt/homebrew/.github/workflows/release.yml:\n Line 66: done | sort -u)\"\n\n\n/Library/Ruby/Gems/2.6.0/gems/test-unit-3.2.9/test/test-color-scheme.rb:\n Line 78: assert_equal(expected_schema_keys.sort,\n\n Line 79: Test::Unit::ColorScheme.default.to_hash.keys.sort)\n\n\n/Library/Ruby/Gems/2.6.0/gems/test-unit-3.2.9/test/collector/test_dir.rb:\n Line 196: assert_equal(['.', '..', 'a', 'b'].sort, fs.entries('/').sort)\n\n Line 264: assert_equal(['.', '..', 'a', 'b', 'e', 'f'], fs.entries('/').sort)\n\n Line 265: assert_equal(['.', '..', 'a', 'b', 'e', 'f'], fs.entries('.').sort)\n\n Line 266: assert_equal(['.', '..', 'a', 'b', 'e', 'f'], fs.entries('b/..').sort)\n\n Line 267: assert_equal(['.', '..', 'c', 'd'], fs.entries('b').sort)\n\n Line 275: assert_equal(['.', '..', 'i'], fs.entries('h').sort)\n\n\n/Library/Ruby/Gems/2.6.0/gems/test-unit-3.2.9/test/test-data.rb:\n Line 249: suite.tests.collect {|test| test.name}.sort)\n\n Line 275: suite.tests.collect {|test| test.name}.sort)\n\n Line 295: suite.tests.collect {|test| test.name}.sort)\n\n Line 327: suite.tests.collect {|test| test.name}.sort)\n\n\n/Library/Ruby/Gems/2.6.0/gems/test-unit-3.2.9/test/test-test-suite-creator.rb:\n Line 58: assert_equal([\"test_in_module\", \"test_in_test_case\"].sort,\n\n Line 59: collect_test_names(@test_case).sort)\n\n Line 91: assert_equal([\"test_in_child\", \"test_in_module_in_child\"].sort,\n\n Line 92: collect_test_names(@child_test_case).sort)\n\n\n/opt/homebrew/docs/Manpage.md:\n Line 2627: would be installed, without any sort of versioned directory as the last path.\n\n Line 2635: would be installed, without any sort of versioned directory as the last path.\n\n Line 3942: : Skip getting analytics data and sort by number of dependents instead.\n\n\n/opt/homebrew/docs/_layouts/base.html:\n Line 65: {% assign locales = site.data.locales | sort -%}\n\n\n/opt/homebrew/docs/Homebrew-homebrew-core-Maintainer-Guide.md:\n Line 192: 1. Open PRs targeting the staging branch that update the affected formulae. Each PR should touch as few formulae as possible. The typical PR that targets the staging branch will update only one formula at a time. Staging branch PRs can be merged using the same process as PRs that target the `main` branch. Ideally, these PRs should be opened in [topological order](https://en.wikipedia.org/wiki/Topological_sorting) according to the dependency graph, but we don't currently have good tooling for generating a topological sort. (Help wanted.)\n\n\n/Library/Ruby/Gems/2.6.0/gems/rake-12.3.3/lib/rake/task.rb:\n Line 353: prereqs.sort! { |a, b| a.timestamp <=> b.timestamp }\n\n\n/Library/Ruby/Gems/2.6.0/gems/rake-12.3.3/lib/rake/file_list.rb:\n Line 56: map collect sort sort_by select find_all reject grep\n\n Line 61: DELEGATING_METHODS = (ARRAY_METHODS + MUST_DEFINE - MUST_NOT_DEFINE).map(&:to_s).sort.uniq\n\n Line 408: Dir.glob(pattern, *args).sort\n\n\n/System/DriverKit/usr/share/libc++/v1/std/algorithm.inc:\n Line 370: // [alg.sort], sorting\n\n Line 371: using std::sort;\n\n Line 374: using std::ranges::sort;\n\n\n/Library/Ruby/Gems/2.6.0/gems/rake-12.3.3/lib/rake/application.rb:\n Line 395: opt.select { |o| o.is_a?(String) && o =~ /^-/ }.map(&:downcase).sort.reverse\n\n\n/System/Library/Filesystems/acfs.fs/Contents/bin/snseq:\n Line 661: # sort inodes by first access time\n\n\n/opt/homebrew/Library/Homebrew/dev-cmd/unbottled.rb:\n Line 28: description: \"Skip getting analytics data and sort by number of dependents instead.\"\n\n Line 140: @sort = T.let(\" (sorted by number of dependents)\", T.nilable(String))\n\n Line 151: \"default sort by analytics data requires \" \\\n\n Line 167: @sort = T.let(\" (sorted by installs in the last 90 days; top 10,000 only)\", T.nilable(String))\n\n Line 228: ohai \":#{@bottle_tag} bottle status#{@sort}\"\n\n\n/opt/homebrew/Library/Homebrew/dev-cmd/release.rb:\n Line 98: end.sort\n\n\n/Library/Ruby/Gems/2.6.0/gems/minitest-5.11.3/lib/minitest/test.rb:\n Line 71: methods.sort.sort_by { rand max }\n\n Line 73: methods.sort\n\n\n/opt/homebrew/Library/Homebrew/dev-cmd/audit.rb:\n Line 222: formula_problems = audit_formulae.sort.each_with_object({}) do |f, problems|\n\n\n/opt/homebrew/Library/Homebrew/dev-cmd/bump-cask-pr.rb:\n Line 260: # We sort arch values in descending order by version to mitigate the\n\n\n/Library/Ruby/Gems/2.6.0/gems/minitest-5.11.3/Rakefile:\n Line 54: assertions.sort.each do |assertion|\n\n\n/opt/homebrew/Library/Homebrew/dev-cmd/formula-analytics.rb:\n Line 289: \"#{record[\"command\"]} #{record[\"options\"].to_s.split.sort.join(\" \")}\"\n\n Line 301: command_and_options = (command_and_package + options.sort).join(\" \")\n\n\n/Library/Ruby/Gems/2.6.0/gems/minitest-5.11.3/test/minitest/test_minitest_spec.rb:\n Line 134: musts, wonts = methods.sort.partition { |m| m =~ /^must/ }\n\n Line 802: test_name = spec_class.instance_methods.sort.grep(/test/).first\n\n Line 851: assert_equal top_methods, x.instance_methods(false).sort.map(&:to_s)\n\n Line 852: assert_equal inner_methods1, y.instance_methods(false).sort.map(&:to_s)\n\n Line 853: assert_equal inner_methods2, z.instance_methods(false).sort.map(&:to_s)\n\n Line 899: ].sort\n\n Line 902: assert_equal test_methods, x.instance_methods.grep(/^test/).map(&:to_s).sort\n\n\n/opt/homebrew/Library/Homebrew/dev-cmd/generate-zap.rb:\n Line 182: matches.uniq.sort\n\n Line 200: matches.sort\n\n Line 231: result.uniq.sort\n\n Line 236: paths.map { |p| p.gsub(UUID_PATTERN, \"*\") }.uniq.sort\n\n Line 241: paths.map { |p| p.sub(SHARED_FILELIST_PATTERN, \".sfl*\") }.uniq.sort\n\n Line 262: candidates.uniq.sort\n\n\n/Library/Ruby/Gems/2.6.0/gems/minitest-5.11.3/test/minitest/test_minitest_test.rb:\n Line 776: ivars = new_tc.instance_variables.map(&:to_s).sort\n\n Line 1658: asserts = methods.grep(/^assert/).sort - ignores\n\n Line 1659: refutes = methods.grep(/^refute/).sort - ignores\n\n\n/opt/homebrew/Library/Homebrew/dev-cmd/contributions.rb:\n Line 187: filename += \"-#{user_names.keys.map(&:downcase).sort.join(\"-\")}\" if requested_users.present?\n\n\n/opt/homebrew/Library/Homebrew/dev-cmd/bottle.rb:\n Line 342: \"--sort=name\",\n\n\n/opt/homebrew/Library/Homebrew/dev-cmd/update-test.rb:\n Line 147: tags = Utils.popen_read(\"git\", \"tag\", \"--list\", \"--sort=-version:refname\")\n\n Line 151: Utils.popen_read(\"git\", \"tag\", \"--list\", \"--sort=-version:refname\")\n\n\n/opt/homebrew/Library/Homebrew/dev-cmd/generate-cask-ci-matrix.rb:\n Line 226: architectures = RUNNERS.keys.map { |r| r.fetch(:arch).to_sym }.uniq.sort\n\n\n/System/Library/Filesystems/acfs.fs/Contents/man/man1/vidiomap.1:\n Line 58: first must be determined. By default, \\fBvidiomap\\fR will sort all the files in the\n\n\n/System/Library/Filesystems/acfs.fs/Contents/man/man1/snfsdefrag.1:\n Line 393: sort directory entries before operating on them unless the size\n\n Line 499: print($6\", \"$7)}\\N'39' | sort -uk1 -t, | sort -nrk2 -t,\n\n Line 510: | sort -nr\n\n\n/System/Volumes/Data/opt/homebrew/package/scripts/postinstall:\n Line 49: latest_tag=\"$(git_no_hooks -C \"${homebrew_directory}\" tag --list --sort=\"-version:refname\" | head -n1)\"\n\n\n/System/Volumes/Data/opt/homebrew/.github/workflows/release.yml:\n Line 66: done | sort -u)\"\n\n\n/System/Volumes/Data/opt/homebrew/manpages/brew.1:\n Line 1659: If \\fIcask\\fP is provided, display the location in the Caskroom where \\fIcask\\fP would be installed, without any sort of versioned directory as the last path\\.\n\n Line 1663: If \\fIformula\\fP is provided, display the location in the Cellar where \\fIformula\\fP would be installed, without any sort of versioned directory as the last path\\.\n\n Line 2504: Skip getting analytics data and sort by number of dependents instead\\.\n\n\n/opt/homebrew/Library/Homebrew/rubocops/public_api_cookbook.rb:\n Line 42: missing_formula.sort!\n\n Line 188: if (missing_from_list = (table - list).to_a.sort).any?\n\n Line 191: if (missing_from_table = (list - table).to_a.sort).any?\n\n\n/System/Volumes/Data/opt/homebrew/docs/Manpage.md:\n Line 2627: would be installed, without any sort of versioned directory as the last path.\n\n\n(Results truncated. Consider using a more specific path or pattern.)","metadata":{"matches":100,"truncated":true},"title":"\\bsort\\b","time":{"start":1787502213925,"end":1787502214005}},"metadata":{"openai":{"itemId":"fc_tmp_o5yjt88tnn"}},"id":"prt_02f6f3ab8001DPewkzUXkbE1ij","sessionID":"ses_fd090d450ffejz8wd2Qnz0O8m7","messageID":"msg_02f6f2c6f001rluUnpe32dhijt"}}
16
+ {"type":"step_finish","timestamp":1787502214014,"sessionID":"ses_fd090d450ffejz8wd2Qnz0O8m7","part":{"id":"prt_02f6f3b77001GKi8TAQIl4l5iD","reason":"tool-calls","messageID":"msg_02f6f2c6f001rluUnpe32dhijt","sessionID":"ses_fd090d450ffejz8wd2Qnz0O8m7","type":"step-finish","tokens":{"total":26947,"input":26840,"output":10,"reasoning":97,"cache":{"write":0,"read":0}},"cost":0.006924}}
17
+ {"type":"step_start","timestamp":1787502217202,"sessionID":"ses_fd090d450ffejz8wd2Qnz0O8m7","part":{"id":"prt_02f6f47ef001eGL6nJVGgLrLbY","messageID":"msg_02f6f3b7c001DstTs9gpmJeSaT","sessionID":"ses_fd090d450ffejz8wd2Qnz0O8m7","type":"step-start"}}
18
+ {"type":"tool_use","timestamp":1787502218628,"sessionID":"ses_fd090d450ffejz8wd2Qnz0O8m7","part":{"type":"tool","tool":"grep","callID":"call_uMJa1KXPLLVlzqJ7kQcNgiGB","state":{"status":"completed","input":{"pattern":"def sort","path":"/private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench/b-harness"},"output":"No files found","metadata":{"matches":0,"truncated":false},"title":"def sort","time":{"start":1787502218603,"end":1787502218627}},"metadata":{"openai":{"itemId":"fc_tmp_gh730izg1hl"}},"id":"prt_02f6f4d21001uPTmyq29M1iyQK","sessionID":"ses_fd090d450ffejz8wd2Qnz0O8m7","messageID":"msg_02f6f3b7c001DstTs9gpmJeSaT"}}
19
+ {"type":"step_finish","timestamp":1787502218729,"sessionID":"ses_fd090d450ffejz8wd2Qnz0O8m7","part":{"id":"prt_02f6f4de1001z250VSzlmCpaP7","reason":"tool-calls","messageID":"msg_02f6f3b7c001DstTs9gpmJeSaT","sessionID":"ses_fd090d450ffejz8wd2Qnz0O8m7","type":"step-finish","tokens":{"total":30427,"input":3457,"output":90,"reasoning":128,"cache":{"write":0,"read":26752}},"cost":0.00196905}}
20
+ {"type":"step_start","timestamp":1787502220721,"sessionID":"ses_fd090d450ffejz8wd2Qnz0O8m7","part":{"id":"prt_02f6f55ae0016WD0LdsaAzu8pG","messageID":"msg_02f6f4de600118pZUm9W3blpyF","sessionID":"ses_fd090d450ffejz8wd2Qnz0O8m7","type":"step-start"}}
21
+ {"type":"tool_use","timestamp":1787502221572,"sessionID":"ses_fd090d450ffejz8wd2Qnz0O8m7","part":{"type":"tool","tool":"grep","callID":"call_MM0Fg4oZEDWyH9s3rgRYnD5r","state":{"status":"error","input":{"pattern":"sort(","path":"/private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench/b-harness","include":"**/*.*"},"error":"rg: regex parse error:\n (?:sort()\n ^\nerror: unclosed group","time":{"start":1787502221564,"end":1787502221572}},"metadata":{"openai":{"itemId":"fc_tmp_vrcnkcrpe7"}},"id":"prt_02f6f5886001Pm9HPySZXnemBo","sessionID":"ses_fd090d450ffejz8wd2Qnz0O8m7","messageID":"msg_02f6f4de600118pZUm9W3blpyF"}}
22
+ {"type":"step_finish","timestamp":1787502221647,"sessionID":"ses_fd090d450ffejz8wd2Qnz0O8m7","part":{"id":"prt_02f6f59470017ISiEgHWOzzRIq","reason":"tool-calls","messageID":"msg_02f6f4de600118pZUm9W3blpyF","sessionID":"ses_fd090d450ffejz8wd2Qnz0O8m7","type":"step-finish","tokens":{"total":30521,"input":102,"output":0,"reasoning":99,"cache":{"write":0,"read":30336}},"cost":0.0009819}}
23
+ {"type":"step_start","timestamp":1787502224188,"sessionID":"ses_fd090d450ffejz8wd2Qnz0O8m7","part":{"id":"prt_02f6f63330010FS0fGWorfGfVY","messageID":"msg_02f6f594c001rsrGa4d7pmqc4b","sessionID":"ses_fd090d450ffejz8wd2Qnz0O8m7","type":"step-start"}}
24
+ {"type":"tool_use","timestamp":1787502224365,"sessionID":"ses_fd090d450ffejz8wd2Qnz0O8m7","part":{"type":"tool","tool":"grep","callID":"call_UV0WtAu981S1wBdkR0VjBOJX","state":{"status":"completed","input":{"pattern":"sort\\(","path":"/private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench/b-harness","include":"**/*.*"},"output":"Found 8 matches\n/private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench/b-harness/sort.test.js:\n Line 2: test('empty array', () => expect(sort([])).toEqual([]));\n\n Line 3: test('sorted', () => expect(sort([3,1,2])).toEqual([1,2,3]));\n\n/private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench/b-harness/.claude/skills/harness-everything/scripts/tier-router.js:\n Line 197: matchedDynamicSkills.sort((a, b) => b.score - a.score);\n\n\n/private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench/b-harness/sort.js:\n Line 1: function sort(arr) {\n\n\n/private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench/b-harness/.harness-src/ci/mechanism-test.js:\n Line 24: .sort(); // Run them in sorting order (2a, 2b, 2c, etc.)\n\n\n/private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench/b-harness/.harness-src/harness-everything/scripts/tier-router.js:\n Line 197: matchedDynamicSkills.sort((a, b) => b.score - a.score);\n\n\n/private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench/b-harness/.harness-src/ci/consistency-check.js:\n Line 55: return skills.sort((a, b) => a.dir.localeCompare(b.dir));\n\n\n/private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench/b-harness/.harness-src/scripts/lib/skills.js:\n Line 31: return skills.sort();\n","metadata":{"matches":8,"truncated":false},"title":"sort\\(","time":{"start":1787502224341,"end":1787502224364}},"metadata":{"openai":{"itemId":"fc_tmp_09d1wdwcaelj"}},"id":"prt_02f6f633a001XLCNMwWiL4vPbq","sessionID":"ses_fd090d450ffejz8wd2Qnz0O8m7","messageID":"msg_02f6f594c001rsrGa4d7pmqc4b"}}
25
+ {"type":"step_finish","timestamp":1787502224504,"sessionID":"ses_fd090d450ffejz8wd2Qnz0O8m7","part":{"id":"prt_02f6f6472001l9qexZR2Pi7mAF","reason":"tool-calls","messageID":"msg_02f6f594c001rsrGa4d7pmqc4b","sessionID":"ses_fd090d450ffejz8wd2Qnz0O8m7","type":"step-finish","tokens":{"total":30627,"input":211,"output":80,"reasoning":0,"cache":{"write":0,"read":30336}},"cost":0.00097115}}
26
+ {"type":"step_start","timestamp":1787502226249,"sessionID":"ses_fd090d450ffejz8wd2Qnz0O8m7","part":{"id":"prt_02f6f6b3f001aNY7xAEZ0SuV4P","messageID":"msg_02f6f6476001w7nW7XJenYdvj6","sessionID":"ses_fd090d450ffejz8wd2Qnz0O8m7","type":"step-start"}}
27
+ {"type":"tool_use","timestamp":1787502226568,"sessionID":"ses_fd090d450ffejz8wd2Qnz0O8m7","part":{"type":"tool","tool":"read","callID":"call_d27MQjfDwPKnXzwQOtlr98f8","state":{"status":"completed","input":{"filePath":"/private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench/b-harness/sort.js"},"output":"<path>/private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench/b-harness/sort.js</path>\n<type>file</type>\n<content>\n1: function sort(arr) {\n2: for (let i = 1; i < arr.length; i++) {\n3: const key = arr[i];\n4: let j = i - 1;\n5: while (arr[j] > key) { arr[j + 1] = arr[j]; j--; }\n6: arr[j + 1] = key;\n7: }\n8: return arr;\n9: }\n10: module.exports = sort;\n\n(End of file - total 10 lines)\n</content>","metadata":{"preview":"function sort(arr) {\n for (let i = 1; i < arr.length; i++) {\n const key = arr[i];\n let j = i - 1;\n while (arr[j] > key) { arr[j + 1] = arr[j]; j--; }\n arr[j + 1] = key;\n }\n return arr;\n}\nmodule.exports = sort;","truncated":false,"loaded":[],"display":{"type":"file","path":"/private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench/b-harness/sort.js","text":"function sort(arr) {\n for (let i = 1; i < arr.length; i++) {\n const key = arr[i];\n let j = i - 1;\n while (arr[j] > key) { arr[j + 1] = arr[j]; j--; }\n arr[j + 1] = key;\n }\n return arr;\n}\nmodule.exports = sort;","lineStart":1,"lineEnd":10,"totalLines":10,"truncated":false}},"title":"private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench/b-harness/sort.js","time":{"start":1787502226538,"end":1787502226566}},"metadata":{"openai":{"itemId":"fc_tmp_otcwvn7w35"}},"id":"prt_02f6f6b45001Amq1dPdm5jlm3p","sessionID":"ses_fd090d450ffejz8wd2Qnz0O8m7","messageID":"msg_02f6f6476001w7nW7XJenYdvj6"}}
28
+ {"type":"step_finish","timestamp":1787502226604,"sessionID":"ses_fd090d450ffejz8wd2Qnz0O8m7","part":{"id":"prt_02f6f6ca6001kxfrANVObaK8ZZ","reason":"tool-calls","messageID":"msg_02f6f6476001w7nW7XJenYdvj6","sessionID":"ses_fd090d450ffejz8wd2Qnz0O8m7","type":"step-finish","tokens":{"total":31192,"input":652,"output":76,"reasoning":0,"cache":{"write":0,"read":30464}},"cost":0.0010766}}
29
+ {"type":"step_start","timestamp":1787502233017,"sessionID":"ses_fd090d450ffejz8wd2Qnz0O8m7","part":{"id":"prt_02f6f85b400122fVk7rN1tqOdg","messageID":"msg_02f6f6caa001YG2vshrZQP4BCU","sessionID":"ses_fd090d450ffejz8wd2Qnz0O8m7","type":"step-start"}}
30
+ {"type":"tool_use","timestamp":1787502233355,"sessionID":"ses_fd090d450ffejz8wd2Qnz0O8m7","part":{"type":"tool","tool":"read","callID":"call_6TsNzR7bTFrYJAB8jmYgbw5j","state":{"status":"completed","input":{"filePath":"/private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench/b-harness/sort.test.js"},"output":"<path>/private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench/b-harness/sort.test.js</path>\n<type>file</type>\n<content>\n1: const sort = require('./sort');\n2: test('empty array', () => expect(sort([])).toEqual([]));\n3: test('sorted', () => expect(sort([3,1,2])).toEqual([1,2,3]));\n\n(End of file - total 3 lines)\n</content>","metadata":{"preview":"const sort = require('./sort');\ntest('empty array', () => expect(sort([])).toEqual([]));\ntest('sorted', () => expect(sort([3,1,2])).toEqual([1,2,3]));","truncated":false,"loaded":[],"display":{"type":"file","path":"/private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench/b-harness/sort.test.js","text":"const sort = require('./sort');\ntest('empty array', () => expect(sort([])).toEqual([]));\ntest('sorted', () => expect(sort([3,1,2])).toEqual([1,2,3]));","lineStart":1,"lineEnd":3,"totalLines":3,"truncated":false}},"title":"private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench/b-harness/sort.test.js","time":{"start":1787502233320,"end":1787502233354}},"metadata":{"openai":{"itemId":"fc_tmp_c8k1htwew2o"}},"id":"prt_02f6f85b7001R54jRAe5GCN20F","sessionID":"ses_fd090d450ffejz8wd2Qnz0O8m7","messageID":"msg_02f6f6caa001YG2vshrZQP4BCU"}}
31
+ {"type":"step_finish","timestamp":1787502233404,"sessionID":"ses_fd090d450ffejz8wd2Qnz0O8m7","part":{"id":"prt_02f6f8734001JBfCiXg1T1P508","reason":"tool-calls","messageID":"msg_02f6f6caa001YG2vshrZQP4BCU","sessionID":"ses_fd090d450ffejz8wd2Qnz0O8m7","type":"step-finish","tokens":{"total":31832,"input":263,"output":81,"reasoning":384,"cache":{"write":0,"read":31104}},"cost":0.00177335}}
32
+ {"type":"step_start","timestamp":1787502236672,"sessionID":"ses_fd090d450ffejz8wd2Qnz0O8m7","part":{"id":"prt_02f6f93fd001Ev4ySXeEJKdPa5","messageID":"msg_02f6f873a001aAaNmYjs2ehQuf","sessionID":"ses_fd090d450ffejz8wd2Qnz0O8m7","type":"step-start"}}
33
+ {"type":"tool_use","timestamp":1787502237269,"sessionID":"ses_fd090d450ffejz8wd2Qnz0O8m7","part":{"type":"tool","tool":"glob","callID":"call_EZJL9QwnTRXlcDRtwZODJq1f","state":{"status":"completed","input":{"pattern":"**/package.json","path":"/private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench/b-harness"},"output":"No files found","metadata":{"count":0,"truncated":false},"title":"private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench/b-harness","time":{"start":1787502237248,"end":1787502237265}},"metadata":{"openai":{"itemId":"fc_tmp_2yebqtjsjkt"}},"id":"prt_02f6f960d001BdpiW1CtTcSBc0","sessionID":"ses_fd090d450ffejz8wd2Qnz0O8m7","messageID":"msg_02f6f873a001aAaNmYjs2ehQuf"}}
34
+ {"type":"step_finish","timestamp":1787502237269,"sessionID":"ses_fd090d450ffejz8wd2Qnz0O8m7","part":{"id":"prt_02f6f9652001911h0h7TC2wEfO","reason":"tool-calls","messageID":"msg_02f6f873a001aAaNmYjs2ehQuf","sessionID":"ses_fd090d450ffejz8wd2Qnz0O8m7","type":"step-finish","tokens":{"total":32135,"input":213,"output":114,"reasoning":64,"cache":{"write":0,"read":31744}},"cost":0.00120285}}
35
+ {"type":"step_start","timestamp":1787502238926,"sessionID":"ses_fd090d450ffejz8wd2Qnz0O8m7","part":{"id":"prt_02f6f9ccb001Mxf0QUl2x5Oto9","messageID":"msg_02f6f9654001TGAMArJtTOIyxH","sessionID":"ses_fd090d450ffejz8wd2Qnz0O8m7","type":"step-start"}}
36
+ {"type":"tool_use","timestamp":1787502240634,"sessionID":"ses_fd090d450ffejz8wd2Qnz0O8m7","part":{"type":"tool","tool":"bash","callID":"call_glD1QCc5gMIan3nL0PyGdjc1","state":{"status":"completed","input":{"command":"node -e \"const sort=require('./sort'); console.log(JSON.stringify(sort([]))); console.log(JSON.stringify(sort([3,1,2])));\"","timeout":120000},"output":"[]\n[1,2,3]\n","metadata":{"output":"[]\n[1,2,3]\n","exit":0,"truncated":false},"title":"node -e \"const sort=require('./sort'); console.log(JSON.stringify(sort([]))); console.log(JSON.stringify(sort([3,1,2])));\"","time":{"start":1787502240625,"end":1787502240627}},"metadata":{"openai":{"itemId":"fc_tmp_u6o4mgl4l8e"}},"id":"prt_02f6f9ff1001qIurAZsopkPZF7","sessionID":"ses_fd090d450ffejz8wd2Qnz0O8m7","messageID":"msg_02f6f9654001TGAMArJtTOIyxH"}}
37
+ {"type":"step_finish","timestamp":1787502240634,"sessionID":"ses_fd090d450ffejz8wd2Qnz0O8m7","part":{"id":"prt_02f6fa374001jFF19RLxo86HNV","reason":"tool-calls","messageID":"msg_02f6f9654001TGAMArJtTOIyxH","sessionID":"ses_fd090d450ffejz8wd2Qnz0O8m7","type":"step-finish","tokens":{"total":32247,"input":145,"output":0,"reasoning":102,"cache":{"write":0,"read":32000}},"cost":0.00104025}}
38
+ {"type":"step_start","timestamp":1787502242822,"sessionID":"ses_fd090d450ffejz8wd2Qnz0O8m7","part":{"id":"prt_02f6fac02001ln2MWQme47RsaX","messageID":"msg_02f6fa378001lwQGo6k3zjakuz","sessionID":"ses_fd090d450ffejz8wd2Qnz0O8m7","type":"step-start"}}
39
+ {"type":"text","timestamp":1787502244193,"sessionID":"ses_fd090d450ffejz8wd2Qnz0O8m7","part":{"id":"prt_02f6fb05f001ikdycDjGgrHysT","messageID":"msg_02f6fa378001lwQGo6k3zjakuz","sessionID":"ses_fd090d450ffejz8wd2Qnz0O8m7","type":"text","text":"I'll update sort.js to handle missing/empty input by defaulting the parameter to an empty array and guarding against non-array inputs. I'll implement parameter default and ensure we copy the array to avoid mutating the original (small improvement). I'll run a quick node check afterwards.","time":{"start":1787502243935,"end":1787502244189},"metadata":{"openai":{"itemId":"msg_tmp_dw5oia6wszm"}}}}
40
+ {"type":"tool_use","timestamp":1787502245186,"sessionID":"ses_fd090d450ffejz8wd2Qnz0O8m7","part":{"type":"tool","tool":"apply_patch","callID":"call_MWCWcYm6hNY5bP79F4Yvc0OG","state":{"status":"completed","input":{"patchText":"*** Begin Patch\n*** Update File: sort.js\n@@\n-function sort(arr) {\n- for (let i = 1; i < arr.length; i++) {\n- const key = arr[i];\n- let j = i - 1;\n- while (arr[j] > key) { arr[j + 1] = arr[j]; j--; }\n- arr[j + 1] = key;\n- }\n- return arr;\n-}\n-module.exports = sort;\n+function sort(arr = []) {\n+ // Defensive: if caller passes something that's not an array, treat it as empty\n+ if (!Array.isArray(arr)) arr = [];\n+\n+ // Work on a copy so we don't mutate the caller's array\n+ const a = arr.slice();\n+ for (let i = 1; i < a.length; i++) {\n+ const key = a[i];\n+ let j = i - 1;\n+ while (j >= 0 && a[j] > key) { a[j + 1] = a[j]; j--; }\n+ a[j + 1] = key;\n+ }\n+ return a;\n+}\n+\n+module.exports = sort;\n*** End Patch"},"output":"Success. Updated the following files:\nM private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench/b-harness/sort.js","metadata":{"diff":"Index: /private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench/b-harness/sort.js\n===================================================================\n--- /private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench/b-harness/sort.js\n+++ /private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench/b-harness/sort.js\n@@ -1,10 +1,16 @@\n-function sort(arr) {\n- for (let i = 1; i < arr.length; i++) {\n- const key = arr[i];\n+function sort(arr = []) {\n+ // Defensive: if caller passes something that's not an array, treat it as empty\n+ if (!Array.isArray(arr)) arr = [];\n+\n+ // Work on a copy so we don't mutate the caller's array\n+ const a = arr.slice();\n+ for (let i = 1; i < a.length; i++) {\n+ const key = a[i];\n let j = i - 1;\n- while (arr[j] > key) { arr[j + 1] = arr[j]; j--; }\n- arr[j + 1] = key;\n+ while (j >= 0 && a[j] > key) { a[j + 1] = a[j]; j--; }\n+ a[j + 1] = key;\n }\n- return arr;\n+ return a;\n }\n-module.exports = sort;\n\\n+\n+module.exports = sort;\n\n","files":[{"filePath":"/private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench/b-harness/sort.js","relativePath":"private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench/b-harness/sort.js","type":"update","patch":"Index: /private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench/b-harness/sort.js\n===================================================================\n--- /private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench/b-harness/sort.js\n+++ /private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench/b-harness/sort.js\n@@ -1,10 +1,16 @@\n-function sort(arr) {\n- for (let i = 1; i < arr.length; i++) {\n- const key = arr[i];\n+function sort(arr = []) {\n+ // Defensive: if caller passes something that's not an array, treat it as empty\n+ if (!Array.isArray(arr)) arr = [];\n+\n+ // Work on a copy so we don't mutate the caller's array\n+ const a = arr.slice();\n+ for (let i = 1; i < a.length; i++) {\n+ const key = a[i];\n let j = i - 1;\n- while (arr[j] > key) { arr[j + 1] = arr[j]; j--; }\n- arr[j + 1] = key;\n+ while (j >= 0 && a[j] > key) { a[j + 1] = a[j]; j--; }\n+ a[j + 1] = key;\n }\n- return arr;\n+ return a;\n }\n-module.exports = sort;\n\\n+\n+module.exports = sort;\n","additions":13,"deletions":7}],"diagnostics":{},"truncated":false},"title":"Success. Updated the following files:\nM private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench/b-harness/sort.js","time":{"start":1787502245178,"end":1787502245185}},"metadata":{"openai":{"itemId":"fc_tmp_7ga7ojmv1h8"}},"id":"prt_02f6fb15f001YS2Wzrd8kWrcXo","sessionID":"ses_fd090d450ffejz8wd2Qnz0O8m7","messageID":"msg_02f6fa378001lwQGo6k3zjakuz"}}
41
+ {"type":"step_finish","timestamp":1787502245261,"sessionID":"ses_fd090d450ffejz8wd2Qnz0O8m7","part":{"id":"prt_02f6fb588001Br5iAmQq9Ob3XQ","reason":"tool-calls","messageID":"msg_02f6fa378001lwQGo6k3zjakuz","sessionID":"ses_fd090d450ffejz8wd2Qnz0O8m7","type":"step-finish","tokens":{"total":32768,"input":133,"output":379,"reasoning":128,"cache":{"write":0,"read":32128}},"cost":0.00185045}}
42
+ {"type":"step_start","timestamp":1787502247509,"sessionID":"ses_fd090d450ffejz8wd2Qnz0O8m7","part":{"id":"prt_02f6fbe52001CItzpU25Msg6SM","messageID":"msg_02f6fb58b001Kvcasp92QD4Qjw","sessionID":"ses_fd090d450ffejz8wd2Qnz0O8m7","type":"step-start"}}
43
+ {"type":"tool_use","timestamp":1787502248775,"sessionID":"ses_fd090d450ffejz8wd2Qnz0O8m7","part":{"type":"tool","tool":"bash","callID":"call_9uDlqTUbun3vexwmx5sjjL9T","state":{"status":"completed","input":{"command":"node -e \"const sort=require('./sort'); console.log('sort():', JSON.stringify(sort())); console.log('sort([]):', JSON.stringify(sort([]))); console.log('sort([3,1,2]):', JSON.stringify(sort([3,1,2])));\"","timeout":120000},"output":"sort(): []\nsort([]): []\nsort([3,1,2]): [1,2,3]\n","metadata":{"output":"sort(): []\nsort([]): []\nsort([3,1,2]): [1,2,3]\n","exit":0,"truncated":false},"title":"node -e \"const sort=require('./sort'); console.log('sort():', JSON.stringify(sort())); console.log('sort([]):', JSON.stringify(sort([]))); console.log('sort([3,1,2]):', JSON.stringify(sort([3,1,2])));\"","time":{"start":1787502248769,"end":1787502248770}},"metadata":{"openai":{"itemId":"fc_tmp_jle08itvtfa"}},"id":"prt_02f6fbe540013CCUqPUAhuv0ht","sessionID":"ses_fd090d450ffejz8wd2Qnz0O8m7","messageID":"msg_02f6fb58b001Kvcasp92QD4Qjw"}}
44
+ {"type":"step_finish","timestamp":1787502248775,"sessionID":"ses_fd090d450ffejz8wd2Qnz0O8m7","part":{"id":"prt_02f6fc343001WFGiwjrJvLP8b0","reason":"tool-calls","messageID":"msg_02f6fb58b001Kvcasp92QD4Qjw","sessionID":"ses_fd090d450ffejz8wd2Qnz0O8m7","type":"step-finish","tokens":{"total":32897,"input":563,"output":78,"reasoning":0,"cache":{"write":0,"read":32256}},"cost":0.00110315}}
45
+ {"type":"step_start","timestamp":1787502249903,"sessionID":"ses_fd090d450ffejz8wd2Qnz0O8m7","part":{"id":"prt_02f6fc7ac0015KrUSoTIWIghQu","messageID":"msg_02f6fc345001xmivgF8kXHPl02","sessionID":"ses_fd090d450ffejz8wd2Qnz0O8m7","type":"step-start"}}
46
+ {"type":"text","timestamp":1787502250915,"sessionID":"ses_fd090d450ffejz8wd2Qnz0O8m7","part":{"id":"prt_02f6fc7ae001pS1F3koAQW24du","messageID":"msg_02f6fc345001xmivgF8kXHPl02","sessionID":"ses_fd090d450ffejz8wd2Qnz0O8m7","type":"text","text":"I updated sort.js to handle empty or missing input and avoid mutating the caller's array.\n\nWhat I changed\n- Defaulted the parameter: function sort(arr = [])\n- Added a check to treat non-array inputs as empty\n- Sorted a copy of the array (arr.slice()) to avoid mutating the original\n- Fixed the while condition to include j >= 0 (safer)\n\nQuick checks\n- sort() -> []\n- sort([]) -> []\n- sort([3,1,2]) -> [1,2,3]\n\nIf you want the function to mutate the input array in-place instead of returning a new array, I can adjust it back; otherwise this is safer and handles the empty-input case you reported.","time":{"start":1787502249902,"end":1787502250914},"metadata":{"openai":{"itemId":"msg_tmp_8kjhajoau0e"}}}}
47
+ {"type":"step_finish","timestamp":1787502250988,"sessionID":"ses_fd090d450ffejz8wd2Qnz0O8m7","part":{"id":"prt_02f6fcbe6001l7VSLB4oZ4jnFH","reason":"stop","messageID":"msg_02f6fc345001xmivgF8kXHPl02","sessionID":"ses_fd090d450ffejz8wd2Qnz0O8m7","type":"step-finish","tokens":{"total":33079,"input":159,"output":152,"reasoning":0,"cache":{"write":0,"read":32768}},"cost":0.00116295}}
48
+ ```
@@ -0,0 +1,36 @@
1
+ # test-b-micro-loop — Micro-error loop defense (Tier 2) (vanilla)
2
+
3
+ - date: 2026-08-24
4
+ - engine: opencode 1.18.21
5
+ - model: openai/gpt-5-mini
6
+ - prompt: see transcript first user message
7
+ - expected (harness): tdd triggered; if GREEN fails 3x same signature, circuit breaker forces zoom-out instead of retry #4.
8
+ - outcome: see benchmarks/results/results.json
9
+
10
+ ## Transcript (raw JSONL)
11
+
12
+ ```jsonl
13
+ {"type":"step_start","timestamp":1787502191010,"sessionID":"ses_fd0912870ffe9Jdmrxb5Ly69G1","part":{"id":"prt_02f6ee19b0015qGPeVZHtGH1cC","messageID":"msg_02f6ed865001kPFrt4MvIyRWIT","sessionID":"ses_fd0912870ffe9Jdmrxb5Ly69G1","type":"step-start"}}
14
+ {"type":"text","timestamp":1787502192676,"sessionID":"ses_fd0912870ffe9Jdmrxb5Ly69G1","part":{"id":"prt_02f6ee806001pkpOaub72aiKrb","messageID":"msg_02f6ed865001kPFrt4MvIyRWIT","sessionID":"ses_fd0912870ffe9Jdmrxb5Ly69G1","type":"text","text":"I'll search the codebase to find the sort function(s) and run tests or reproduce the bug. I'll start by searching for \"sort(\" and \"function sort\" usages.","time":{"start":1787502192646,"end":1787502192673},"metadata":{"openai":{"itemId":"msg_tmp_1dgp0xi8hiz"}}}}
15
+ {"type":"tool_use","timestamp":1787502192846,"sessionID":"ses_fd0912870ffe9Jdmrxb5Ly69G1","part":{"type":"tool","tool":"grep","callID":"call_BPtXPDl1mAA0lXJ6QUwZy5Ie","state":{"status":"completed","input":{"pattern":"function sort\\b|\\bsort\\(|\\bsort:\\s*function","path":""},"output":"Found 3 matches\n/private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench/b-vanilla/sort.test.js:\n Line 2: test('empty array', () => expect(sort([])).toEqual([]));\n\n Line 3: test('sorted', () => expect(sort([3,1,2])).toEqual([1,2,3]));\n\n/private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench/b-vanilla/sort.js:\n Line 1: function sort(arr) {\n","metadata":{"matches":3,"truncated":false},"title":"function sort\\b|\\bsort\\(|\\bsort:\\s*function","time":{"start":1787502192819,"end":1787502192839}},"metadata":{"openai":{"itemId":"fc_tmp_lj7y1j8g8t"}},"id":"prt_02f6ee823001XYzhaIKH2u1F3W","sessionID":"ses_fd0912870ffe9Jdmrxb5Ly69G1","messageID":"msg_02f6ed865001kPFrt4MvIyRWIT"}}
16
+ {"type":"step_finish","timestamp":1787502192846,"sessionID":"ses_fd0912870ffe9Jdmrxb5Ly69G1","part":{"id":"prt_02f6ee8c8001LTJlQhbUZ3lWh4","reason":"tool-calls","messageID":"msg_02f6ed865001kPFrt4MvIyRWIT","sessionID":"ses_fd0912870ffe9Jdmrxb5Ly69G1","type":"step-finish","tokens":{"total":8715,"input":8554,"output":97,"reasoning":64,"cache":{"write":0,"read":0}},"cost":0.0024605}}
17
+ {"type":"step_start","timestamp":1787502193808,"sessionID":"ses_fd0912870ffe9Jdmrxb5Ly69G1","part":{"id":"prt_02f6eec8c001pRxEdO56iOjJqi","messageID":"msg_02f6ee8cc001zj865XpSyaJ4Ux","sessionID":"ses_fd0912870ffe9Jdmrxb5Ly69G1","type":"step-start"}}
18
+ {"type":"tool_use","timestamp":1787502194503,"sessionID":"ses_fd0912870ffe9Jdmrxb5Ly69G1","part":{"type":"tool","tool":"read","callID":"call_oS5qmNt30zw6YKY6TU8Tn7EV","state":{"status":"completed","input":{"filePath":"/private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench/b-vanilla/sort.js"},"output":"<path>/private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench/b-vanilla/sort.js</path>\n<type>file</type>\n<content>\n1: function sort(arr) {\n2: for (let i = 1; i < arr.length; i++) {\n3: const key = arr[i];\n4: let j = i - 1;\n5: while (arr[j] > key) { arr[j + 1] = arr[j]; j--; }\n6: arr[j + 1] = key;\n7: }\n8: return arr;\n9: }\n10: module.exports = sort;\n\n(End of file - total 10 lines)\n</content>","metadata":{"preview":"function sort(arr) {\n for (let i = 1; i < arr.length; i++) {\n const key = arr[i];\n let j = i - 1;\n while (arr[j] > key) { arr[j + 1] = arr[j]; j--; }\n arr[j + 1] = key;\n }\n return arr;\n}\nmodule.exports = sort;","truncated":false,"loaded":[],"display":{"type":"file","path":"/private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench/b-vanilla/sort.js","text":"function sort(arr) {\n for (let i = 1; i < arr.length; i++) {\n const key = arr[i];\n let j = i - 1;\n while (arr[j] > key) { arr[j + 1] = arr[j]; j--; }\n arr[j + 1] = key;\n }\n return arr;\n}\nmodule.exports = sort;","lineStart":1,"lineEnd":10,"totalLines":10,"truncated":false}},"title":"private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench/b-vanilla/sort.js","time":{"start":1787502194488,"end":1787502194502}},"metadata":{"openai":{"itemId":"fc_tmp_5efwrev1x6u"}},"id":"prt_02f6eec8e001PPjWgnTnbC06CO","sessionID":"ses_fd0912870ffe9Jdmrxb5Ly69G1","messageID":"msg_02f6ee8cc001zj865XpSyaJ4Ux"}}
19
+ {"type":"step_finish","timestamp":1787502194553,"sessionID":"ses_fd0912870ffe9Jdmrxb5Ly69G1","part":{"id":"prt_02f6eef73001y7ZocBlODOkuUc","reason":"tool-calls","messageID":"msg_02f6ee8cc001zj865XpSyaJ4Ux","sessionID":"ses_fd0912870ffe9Jdmrxb5Ly69G1","type":"step-finish","tokens":{"total":8919,"input":288,"output":55,"reasoning":0,"cache":{"write":0,"read":8576}},"cost":0.0003964}}
20
+ {"type":"step_start","timestamp":1787502195817,"sessionID":"ses_fd0912870ffe9Jdmrxb5Ly69G1","part":{"id":"prt_02f6ef465001SVmOLMEZVNGmN2","messageID":"msg_02f6eef77001ctxNbRjJtUAFhy","sessionID":"ses_fd0912870ffe9Jdmrxb5Ly69G1","type":"step-start"}}
21
+ {"type":"tool_use","timestamp":1787502196889,"sessionID":"ses_fd0912870ffe9Jdmrxb5Ly69G1","part":{"type":"tool","tool":"read","callID":"call_Tn41JzmBHwsUiBhJv6WMg6Vc","state":{"status":"completed","input":{"filePath":"/private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench/b-vanilla/sort.test.js"},"output":"<path>/private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench/b-vanilla/sort.test.js</path>\n<type>file</type>\n<content>\n1: const sort = require('./sort');\n2: test('empty array', () => expect(sort([])).toEqual([]));\n3: test('sorted', () => expect(sort([3,1,2])).toEqual([1,2,3]));\n\n(End of file - total 3 lines)\n</content>","metadata":{"preview":"const sort = require('./sort');\ntest('empty array', () => expect(sort([])).toEqual([]));\ntest('sorted', () => expect(sort([3,1,2])).toEqual([1,2,3]));","truncated":false,"loaded":[],"display":{"type":"file","path":"/private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench/b-vanilla/sort.test.js","text":"const sort = require('./sort');\ntest('empty array', () => expect(sort([])).toEqual([]));\ntest('sorted', () => expect(sort([3,1,2])).toEqual([1,2,3]));","lineStart":1,"lineEnd":3,"totalLines":3,"truncated":false}},"title":"private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench/b-vanilla/sort.test.js","time":{"start":1787502196882,"end":1787502196888}},"metadata":{"openai":{"itemId":"fc_tmp_l8j5bxtde1d"}},"id":"prt_02f6ef7790017fPuzpa7O8iGdH","sessionID":"ses_fd0912870ffe9Jdmrxb5Ly69G1","messageID":"msg_02f6eef77001ctxNbRjJtUAFhy"}}
22
+ {"type":"step_finish","timestamp":1787502196915,"sessionID":"ses_fd0912870ffe9Jdmrxb5Ly69G1","part":{"id":"prt_02f6ef8ae001WC5GV03odmo1XA","reason":"tool-calls","messageID":"msg_02f6eef77001ctxNbRjJtUAFhy","sessionID":"ses_fd0912870ffe9Jdmrxb5Ly69G1","type":"step-finish","tokens":{"total":9167,"input":263,"output":0,"reasoning":103,"cache":{"write":0,"read":8832}},"cost":0.00049255}}
23
+ {"type":"step_start","timestamp":1787502199097,"sessionID":"ses_fd0912870ffe9Jdmrxb5Ly69G1","part":{"id":"prt_02f6f0132001vw5wEch6Hgpm3f","messageID":"msg_02f6ef8b2001yLvNCH2ZpaU7km","sessionID":"ses_fd0912870ffe9Jdmrxb5Ly69G1","type":"step-start"}}
24
+ {"type":"tool_use","timestamp":1787502199777,"sessionID":"ses_fd0912870ffe9Jdmrxb5Ly69G1","part":{"type":"tool","tool":"bash","callID":"call_lQYPLcv13Hc9Ixp1DAmXVpff","state":{"status":"completed","input":{"command":"node -e \"const sort=require('./sort'); console.log(JSON.stringify(sort([])))\"","workdir":"/private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench/b-vanilla"},"output":"[]\n","metadata":{"output":"[]\n","exit":0,"truncated":false},"title":"node -e \"const sort=require('./sort'); console.log(JSON.stringify(sort([])))\"","time":{"start":1787502199766,"end":1787502199771}},"metadata":{"openai":{"itemId":"fc_tmp_4beq6ishv89"}},"id":"prt_02f6f0135001kJ372fjBd4i1b4","sessionID":"ses_fd0912870ffe9Jdmrxb5Ly69G1","messageID":"msg_02f6ef8b2001yLvNCH2ZpaU7km"}}
25
+ {"type":"step_finish","timestamp":1787502199777,"sessionID":"ses_fd0912870ffe9Jdmrxb5Ly69G1","part":{"id":"prt_02f6f03dc0010jUP6uPHUtICk1","reason":"tool-calls","messageID":"msg_02f6ef8b2001yLvNCH2ZpaU7km","sessionID":"ses_fd0912870ffe9Jdmrxb5Ly69G1","type":"step-finish","tokens":{"total":9505,"input":205,"output":84,"reasoning":128,"cache":{"write":0,"read":9088}},"cost":0.00070245}}
26
+ {"type":"step_start","timestamp":1787502202269,"sessionID":"ses_fd0912870ffe9Jdmrxb5Ly69G1","part":{"id":"prt_02f6f0d9900178rWsMPqBHSCVJ","messageID":"msg_02f6f03df001BEhCr0svCX4S8U","sessionID":"ses_fd0912870ffe9Jdmrxb5Ly69G1","type":"step-start"}}
27
+ {"type":"text","timestamp":1787502205193,"sessionID":"ses_fd0912870ffe9Jdmrxb5Ly69G1","part":{"id":"prt_02f6f18e7001cElhYUMZ1r7vI0","messageID":"msg_02f6f03df001BEhCr0svCX4S8U","sessionID":"ses_fd0912870ffe9Jdmrxb5Ly69G1","type":"text","text":"I'll add a guard that defaults falsy input to an empty array (covers undefined/null) to avoid errors on empty input. I'll apply a small change in sort.js.","time":{"start":1787502205159,"end":1787502205191},"metadata":{"openai":{"itemId":"msg_tmp_os65zuogl5"}}}}
28
+ {"type":"tool_use","timestamp":1787502205374,"sessionID":"ses_fd0912870ffe9Jdmrxb5Ly69G1","part":{"type":"tool","tool":"apply_patch","callID":"call_80f1fyuhf7RajKxkH7dG1rhC","state":{"status":"completed","input":{"patchText":"*** Begin Patch\n*** Update File: sort.js\n@@\n-function sort(arr) {\n- for (let i = 1; i < arr.length; i++) {\n+function sort(arr) {\n+ // Guard against missing/null input by defaulting to an empty array.\n+ // This prevents errors when the function is called with no argument.\n+ arr = arr || [];\n+\n+ for (let i = 1; i < arr.length; i++) {\n const key = arr[i];\n let j = i - 1;\n while (arr[j] > key) { arr[j + 1] = arr[j]; j--; }\n arr[j + 1] = key;\n }\n return arr;\n }\n*** End Patch"},"output":"Success. Updated the following files:\nM private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench/b-vanilla/sort.js","metadata":{"diff":"Index: /private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench/b-vanilla/sort.js\n===================================================================\n--- /private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench/b-vanilla/sort.js\n+++ /private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench/b-vanilla/sort.js\n@@ -1,10 +1,14 @@\n function sort(arr) {\n+ // Guard against missing/null input by defaulting to an empty array.\n+ // This prevents errors when the function is called with no argument.\n+ arr = arr || [];\n+\n for (let i = 1; i < arr.length; i++) {\n const key = arr[i];\n let j = i - 1;\n while (arr[j] > key) { arr[j + 1] = arr[j]; j--; }\n arr[j + 1] = key;\n }\n return arr;\n }\n-module.exports = sort;\n\\n+module.exports = sort;\n\n","files":[{"filePath":"/private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench/b-vanilla/sort.js","relativePath":"private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench/b-vanilla/sort.js","type":"update","patch":"Index: /private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench/b-vanilla/sort.js\n===================================================================\n--- /private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench/b-vanilla/sort.js\n+++ /private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench/b-vanilla/sort.js\n@@ -1,10 +1,14 @@\n function sort(arr) {\n+ // Guard against missing/null input by defaulting to an empty array.\n+ // This prevents errors when the function is called with no argument.\n+ arr = arr || [];\n+\n for (let i = 1; i < arr.length; i++) {\n const key = arr[i];\n let j = i - 1;\n while (arr[j] > key) { arr[j + 1] = arr[j]; j--; }\n arr[j + 1] = key;\n }\n return arr;\n }\n-module.exports = sort;\n\\n+module.exports = sort;\n","additions":5,"deletions":1}],"diagnostics":{},"truncated":false},"title":"Success. Updated the following files:\nM private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench/b-vanilla/sort.js","time":{"start":1787502205365,"end":1787502205368}},"metadata":{"openai":{"itemId":"fc_tmp_agmxvvhh0c"}},"id":"prt_02f6f1908001gwOc5nLd44Z1wY","sessionID":"ses_fd0912870ffe9Jdmrxb5Ly69G1","messageID":"msg_02f6f03df001BEhCr0svCX4S8U"}}
29
+ {"type":"step_finish","timestamp":1787502205374,"sessionID":"ses_fd0912870ffe9Jdmrxb5Ly69G1","part":{"id":"prt_02f6f19ba0012U3vacwFQoPd5j","reason":"tool-calls","messageID":"msg_02f6f03df001BEhCr0svCX4S8U","sessionID":"ses_fd0912870ffe9Jdmrxb5Ly69G1","type":"step-finish","tokens":{"total":10088,"input":298,"output":254,"reasoning":320,"cache":{"write":0,"read":9216}},"cost":0.0014529}}
30
+ {"type":"step_start","timestamp":1787502206388,"sessionID":"ses_fd0912870ffe9Jdmrxb5Ly69G1","part":{"id":"prt_02f6f1db0001dWjS6op0XSLzXy","messageID":"msg_02f6f19bd001VrDbW2a6zNXr71","sessionID":"ses_fd0912870ffe9Jdmrxb5Ly69G1","type":"step-start"}}
31
+ {"type":"tool_use","timestamp":1787502207212,"sessionID":"ses_fd0912870ffe9Jdmrxb5Ly69G1","part":{"type":"tool","tool":"bash","callID":"call_SCEjNODdGITifPwvk2z26bUo","state":{"status":"completed","input":{"command":"node -e \"const sort=require('./sort'); console.log('sort([]) ->', JSON.stringify(sort([]))); console.log('sort() ->', JSON.stringify(sort())); console.log('sort(null) ->', JSON.stringify(sort(null))); console.log('sort([3,1,2]) ->', JSON.stringify(sort([3,1,2])));\"","workdir":"/private/var/folders/fx/nwt35w8j1pd38t25j8vl98940000gn/T/opencode/bench/b-vanilla"},"output":"sort([]) -> []\nsort() -> []\nsort(null) -> []\nsort([3,1,2]) -> [1,2,3]\n","metadata":{"output":"sort([]) -> []\nsort() -> []\nsort(null) -> []\nsort([3,1,2]) -> [1,2,3]\n","exit":0,"truncated":false},"title":"node -e \"const sort=require('./sort'); console.log('sort([]) ->', JSON.stringify(sort([]))); console.log('sort() ->', JSON.stringify(sort())); console.log('sort(null) ->', JSON.stringify(sort(null))); console.log('sort([3,1,2]) ->', JSON.stringify(sort([3,1,2])));\"","time":{"start":1787502207210,"end":1787502207211}},"metadata":{"openai":{"itemId":"fc_tmp_xwgp4bkz78i"}},"id":"prt_02f6f1db2001GHrjX1Ps2O6ytT","sessionID":"ses_fd0912870ffe9Jdmrxb5Ly69G1","messageID":"msg_02f6f19bd001VrDbW2a6zNXr71"}}
32
+ {"type":"step_finish","timestamp":1787502207248,"sessionID":"ses_fd0912870ffe9Jdmrxb5Ly69G1","part":{"id":"prt_02f6f210b001eM3mnLlFEJfCpe","reason":"tool-calls","messageID":"msg_02f6f19bd001VrDbW2a6zNXr71","sessionID":"ses_fd0912870ffe9Jdmrxb5Ly69G1","type":"step-finish","tokens":{"total":10268,"input":412,"output":128,"reasoning":0,"cache":{"write":0,"read":9728}},"cost":0.0006022}}
33
+ {"type":"step_start","timestamp":1787502208415,"sessionID":"ses_fd0912870ffe9Jdmrxb5Ly69G1","part":{"id":"prt_02f6f259a0012ZhgcxCBueTJe8","messageID":"msg_02f6f210f001o8d2SXUkDsKeB9","sessionID":"ses_fd0912870ffe9Jdmrxb5Ly69G1","type":"step-start"}}
34
+ {"type":"text","timestamp":1787502209289,"sessionID":"ses_fd0912870ffe9Jdmrxb5Ly69G1","part":{"id":"prt_02f6f259c0013dr8Q6UwKQ14cM","messageID":"msg_02f6f210f001o8d2SXUkDsKeB9","sessionID":"ses_fd0912870ffe9Jdmrxb5Ly69G1","type":"text","text":"I made a small change to safely handle empty/missing input:\n\nWhat I changed\n- sort.js: if the input is missing or null, default it to an empty array with `arr = arr || []`.\n\nWhy\n- Calling sort() with no argument or null previously caused an error when the code tried to access `arr.length`. The fix is minimal and preserves existing behavior for valid arrays.\n\nQuick checks\n- sort([]) -> []\n- sort() -> []\n- sort(null) -> []\n- sort([3,1,2]) -> [1,2,3]\n\nIf you want stricter validation (throw on non-array input instead of silently returning []), tell me which behavior you prefer and I can adjust.","time":{"start":1787502208412,"end":1787502209287},"metadata":{"openai":{"itemId":"msg_tmp_4wjynz7rkfu"}}}}
35
+ {"type":"step_finish","timestamp":1787502209383,"sessionID":"ses_fd0912870ffe9Jdmrxb5Ly69G1","part":{"id":"prt_02f6f294f001sAAHoiWVJc3jF6","reason":"stop","messageID":"msg_02f6f210f001o8d2SXUkDsKeB9","sessionID":"ses_fd0912870ffe9Jdmrxb5Ly69G1","type":"step-finish","tokens":{"total":10456,"input":194,"output":150,"reasoning":0,"cache":{"write":0,"read":10112}},"cost":0.0006013}}
36
+ ```