agent-hitch 0.2.6 → 0.2.8

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 (562) hide show
  1. package/README.md +30 -3
  2. package/README.zh-CN.md +14 -2
  3. package/dist/bin/hitch.js +25 -3
  4. package/dist/bin/hitch.js.map +1 -1
  5. package/dist/scripts/canary-benchmark.js +54 -0
  6. package/dist/scripts/canary-benchmark.js.map +1 -0
  7. package/dist/scripts/canary-buildkit-secrets.js +166 -0
  8. package/dist/scripts/canary-buildkit-secrets.js.map +1 -0
  9. package/dist/scripts/canary-collision-domains.js +77 -0
  10. package/dist/scripts/canary-collision-domains.js.map +1 -0
  11. package/dist/scripts/canary-docker-image-observation.js +181 -0
  12. package/dist/scripts/canary-docker-image-observation.js.map +1 -0
  13. package/dist/scripts/canary-eval-scheduler-throughput.js +47 -0
  14. package/dist/scripts/canary-eval-scheduler-throughput.js.map +1 -0
  15. package/dist/scripts/canary-gpu-hardware.js +95 -0
  16. package/dist/scripts/canary-gpu-hardware.js.map +1 -0
  17. package/dist/scripts/canary-harbor-load.js +365 -0
  18. package/dist/scripts/canary-harbor-load.js.map +1 -0
  19. package/dist/scripts/canary-packaged-harness.js +75 -0
  20. package/dist/scripts/canary-packaged-harness.js.map +1 -0
  21. package/dist/scripts/canary-resource-load.js +188 -0
  22. package/dist/scripts/canary-resource-load.js.map +1 -0
  23. package/dist/scripts/check-architecture.js +10 -5
  24. package/dist/scripts/check-architecture.js.map +1 -1
  25. package/dist/scripts/docker-reaper-crash-child.js +21 -0
  26. package/dist/scripts/docker-reaper-crash-child.js.map +1 -0
  27. package/dist/src/adapters/catalog.js +3 -1
  28. package/dist/src/adapters/catalog.js.map +1 -1
  29. package/dist/src/adapters/discovery.js +1 -1
  30. package/dist/src/adapters/discovery.js.map +1 -1
  31. package/dist/src/adapters/providers/claude.js +6 -0
  32. package/dist/src/adapters/providers/claude.js.map +1 -1
  33. package/dist/src/adapters/providers/codex-auth.js +29 -0
  34. package/dist/src/adapters/providers/codex-auth.js.map +1 -0
  35. package/dist/src/adapters/providers/codex.js +29 -6
  36. package/dist/src/adapters/providers/codex.js.map +1 -1
  37. package/dist/src/adapters/providers/deepseek.js +6 -0
  38. package/dist/src/adapters/providers/deepseek.js.map +1 -1
  39. package/dist/src/adapters/providers/model-call.js +26 -0
  40. package/dist/src/adapters/providers/model-call.js.map +1 -0
  41. package/dist/src/adapters/providers/opencode.js +6 -0
  42. package/dist/src/adapters/providers/opencode.js.map +1 -1
  43. package/dist/src/adapters/providers/pi.js +6 -0
  44. package/dist/src/adapters/providers/pi.js.map +1 -1
  45. package/dist/src/artifacts/index.js +1 -0
  46. package/dist/src/artifacts/index.js.map +1 -1
  47. package/dist/src/artifacts/preparer.js +7 -2
  48. package/dist/src/artifacts/preparer.js.map +1 -1
  49. package/dist/src/backends/harbor/agent-budget.js +31 -0
  50. package/dist/src/backends/harbor/agent-budget.js.map +1 -0
  51. package/dist/src/backends/harbor/backend.js +59 -107
  52. package/dist/src/backends/harbor/backend.js.map +1 -1
  53. package/dist/src/backends/harbor/environment-config.js +92 -0
  54. package/dist/src/backends/harbor/environment-config.js.map +1 -0
  55. package/dist/src/backends/harbor/index.js +6 -1
  56. package/dist/src/backends/harbor/index.js.map +1 -1
  57. package/dist/src/backends/harbor/model-proxy-config.js +23 -0
  58. package/dist/src/backends/harbor/model-proxy-config.js.map +1 -0
  59. package/dist/src/backends/harbor/process.js +177 -0
  60. package/dist/src/backends/harbor/process.js.map +1 -0
  61. package/dist/src/backends/harbor/regrade.js +77 -0
  62. package/dist/src/backends/harbor/regrade.js.map +1 -0
  63. package/dist/src/backends/harbor/runtime-toolchain.js +30 -0
  64. package/dist/src/backends/harbor/runtime-toolchain.js.map +1 -0
  65. package/dist/src/backends/harbor/supervisor.js +74 -0
  66. package/dist/src/backends/harbor/supervisor.js.map +1 -0
  67. package/dist/src/backends/harbor/task-resources.js +206 -0
  68. package/dist/src/backends/harbor/task-resources.js.map +1 -0
  69. package/dist/src/backends/harbor/tools.js +6 -29
  70. package/dist/src/backends/harbor/tools.js.map +1 -1
  71. package/dist/src/backends/index.js +5 -1
  72. package/dist/src/backends/index.js.map +1 -1
  73. package/dist/src/benchmarks/index.js +3 -0
  74. package/dist/src/benchmarks/index.js.map +1 -0
  75. package/dist/src/benchmarks/loader.js +271 -0
  76. package/dist/src/benchmarks/loader.js.map +1 -0
  77. package/dist/src/benchmarks/metrics.js +16 -0
  78. package/dist/src/benchmarks/metrics.js.map +1 -0
  79. package/dist/src/benchmarks/toml.js +13 -0
  80. package/dist/src/benchmarks/toml.js.map +1 -0
  81. package/dist/src/benchmarks/validation.js +242 -0
  82. package/dist/src/benchmarks/validation.js.map +1 -0
  83. package/dist/src/cli/arguments.js +6 -4
  84. package/dist/src/cli/arguments.js.map +1 -1
  85. package/dist/src/cli/commands/benchmark.js +22 -0
  86. package/dist/src/cli/commands/benchmark.js.map +1 -0
  87. package/dist/src/cli/commands/capabilities.js +18 -0
  88. package/dist/src/cli/commands/capabilities.js.map +1 -0
  89. package/dist/src/cli/commands/daemon.js +135 -8
  90. package/dist/src/cli/commands/daemon.js.map +1 -1
  91. package/dist/src/cli/commands/eval.js +228 -6
  92. package/dist/src/cli/commands/eval.js.map +1 -1
  93. package/dist/src/cli/commands/images.js +40 -0
  94. package/dist/src/cli/commands/images.js.map +1 -0
  95. package/dist/src/cli/commands/run.js +42 -17
  96. package/dist/src/cli/commands/run.js.map +1 -1
  97. package/dist/src/cli/commands/runs.js +25 -3
  98. package/dist/src/cli/commands/runs.js.map +1 -1
  99. package/dist/src/cli/commands/trajectory.js +96 -5
  100. package/dist/src/cli/commands/trajectory.js.map +1 -1
  101. package/dist/src/cli/commands/verifier.js +36 -0
  102. package/dist/src/cli/commands/verifier.js.map +1 -0
  103. package/dist/src/cli/commands/worker.js +79 -0
  104. package/dist/src/cli/commands/worker.js.map +1 -0
  105. package/dist/src/cli/main.js +10 -0
  106. package/dist/src/cli/main.js.map +1 -1
  107. package/dist/src/cli/output.js +84 -4
  108. package/dist/src/cli/output.js.map +1 -1
  109. package/dist/src/control-plane/build-admission.js +69 -0
  110. package/dist/src/control-plane/build-admission.js.map +1 -0
  111. package/dist/src/control-plane/build-records.js +6 -0
  112. package/dist/src/control-plane/build-records.js.map +1 -0
  113. package/dist/src/control-plane/collisions.js +40 -0
  114. package/dist/src/control-plane/collisions.js.map +1 -0
  115. package/dist/src/control-plane/eval-control-state.js +22 -0
  116. package/dist/src/control-plane/eval-control-state.js.map +1 -0
  117. package/dist/src/control-plane/eval-control-work.js +44 -0
  118. package/dist/src/control-plane/eval-control-work.js.map +1 -0
  119. package/dist/src/control-plane/eval-image-resolution.js +52 -0
  120. package/dist/src/control-plane/eval-image-resolution.js.map +1 -0
  121. package/dist/src/control-plane/eval-image-services.js +30 -0
  122. package/dist/src/control-plane/eval-image-services.js.map +1 -0
  123. package/dist/src/control-plane/eval-records.js +252 -0
  124. package/dist/src/control-plane/eval-records.js.map +1 -0
  125. package/dist/src/control-plane/eval-recovery.js +181 -0
  126. package/dist/src/control-plane/eval-recovery.js.map +1 -0
  127. package/dist/src/control-plane/eval-scheduler.js +450 -0
  128. package/dist/src/control-plane/eval-scheduler.js.map +1 -0
  129. package/dist/src/control-plane/index.js +20 -0
  130. package/dist/src/control-plane/index.js.map +1 -0
  131. package/dist/src/control-plane/local-worker.js +49 -0
  132. package/dist/src/control-plane/local-worker.js.map +1 -0
  133. package/dist/src/control-plane/model-capture-planning.js +12 -0
  134. package/dist/src/control-plane/model-capture-planning.js.map +1 -0
  135. package/dist/src/control-plane/remote-result-transport.js +175 -0
  136. package/dist/src/control-plane/remote-result-transport.js.map +1 -0
  137. package/dist/src/control-plane/remote-work-coordinator.js +333 -0
  138. package/dist/src/control-plane/remote-work-coordinator.js.map +1 -0
  139. package/dist/src/control-plane/remote-work-inputs.js +266 -0
  140. package/dist/src/control-plane/remote-work-inputs.js.map +1 -0
  141. package/dist/src/control-plane/remote-work-item.js +62 -0
  142. package/dist/src/control-plane/remote-work-item.js.map +1 -0
  143. package/dist/src/control-plane/remote-work-recovery.js +220 -0
  144. package/dist/src/control-plane/remote-work-recovery.js.map +1 -0
  145. package/dist/src/control-plane/remote-worker-artifacts.js +135 -0
  146. package/dist/src/control-plane/remote-worker-artifacts.js.map +1 -0
  147. package/dist/src/control-plane/remote-worker-client.js +202 -0
  148. package/dist/src/control-plane/remote-worker-client.js.map +1 -0
  149. package/dist/src/control-plane/remote-worker-configuration.js +34 -0
  150. package/dist/src/control-plane/remote-worker-configuration.js.map +1 -0
  151. package/dist/src/control-plane/remote-worker-credentials.js +58 -0
  152. package/dist/src/control-plane/remote-worker-credentials.js.map +1 -0
  153. package/dist/src/control-plane/remote-worker-protocol.js +489 -0
  154. package/dist/src/control-plane/remote-worker-protocol.js.map +1 -0
  155. package/dist/src/control-plane/remote-worker-runner.js +214 -0
  156. package/dist/src/control-plane/remote-worker-runner.js.map +1 -0
  157. package/dist/src/control-plane/remote-workers.js +377 -0
  158. package/dist/src/control-plane/remote-workers.js.map +1 -0
  159. package/dist/src/control-plane/rerun-scheduler.js +448 -0
  160. package/dist/src/control-plane/rerun-scheduler.js.map +1 -0
  161. package/dist/src/control-plane/rerun-submission.js +62 -0
  162. package/dist/src/control-plane/rerun-submission.js.map +1 -0
  163. package/dist/src/control-plane/resources.js +182 -0
  164. package/dist/src/control-plane/resources.js.map +1 -0
  165. package/dist/src/control-plane/scheduler-eval-entry.js +32 -0
  166. package/dist/src/control-plane/scheduler-eval-entry.js.map +1 -0
  167. package/dist/src/control-plane/synthetic-result.js +21 -0
  168. package/dist/src/control-plane/synthetic-result.js.map +1 -0
  169. package/dist/src/control-plane/work-admission.js +23 -0
  170. package/dist/src/control-plane/work-admission.js.map +1 -0
  171. package/dist/src/control-plane/work-dispatcher.js +237 -0
  172. package/dist/src/control-plane/work-dispatcher.js.map +1 -0
  173. package/dist/src/controller-runtime/hash.js +19 -3
  174. package/dist/src/controller-runtime/hash.js.map +1 -1
  175. package/dist/src/controller-runtime/index.js +1 -1
  176. package/dist/src/controller-runtime/index.js.map +1 -1
  177. package/dist/src/controller-runtime/store.js +15 -6
  178. package/dist/src/controller-runtime/store.js.map +1 -1
  179. package/dist/src/daemon/auth.js +5 -0
  180. package/dist/src/daemon/auth.js.map +1 -1
  181. package/dist/src/daemon/client.js.map +1 -1
  182. package/dist/src/daemon/event-stream.js +46 -0
  183. package/dist/src/daemon/event-stream.js.map +1 -0
  184. package/dist/src/daemon/health.js +66 -0
  185. package/dist/src/daemon/health.js.map +1 -0
  186. package/dist/src/daemon/launcher.js +28 -1
  187. package/dist/src/daemon/launcher.js.map +1 -1
  188. package/dist/src/daemon/logging.js +23 -0
  189. package/dist/src/daemon/logging.js.map +1 -0
  190. package/dist/src/daemon/scheduler.js +26 -5
  191. package/dist/src/daemon/scheduler.js.map +1 -1
  192. package/dist/src/daemon/server.js +255 -98
  193. package/dist/src/daemon/server.js.map +1 -1
  194. package/dist/src/daemon/telemetry.js +163 -0
  195. package/dist/src/daemon/telemetry.js.map +1 -0
  196. package/dist/src/daemon/worker-routes.js +219 -0
  197. package/dist/src/daemon/worker-routes.js.map +1 -0
  198. package/dist/src/domain/benchmarks.js +2 -0
  199. package/dist/src/domain/benchmarks.js.map +1 -0
  200. package/dist/src/domain/bundles.js +2 -0
  201. package/dist/src/domain/bundles.js.map +1 -0
  202. package/dist/src/domain/control-plane.js +2 -0
  203. package/dist/src/domain/control-plane.js.map +1 -0
  204. package/dist/src/domain/execution-evidence.js +2 -0
  205. package/dist/src/domain/execution-evidence.js.map +1 -0
  206. package/dist/src/domain/execution-plan.js +2 -0
  207. package/dist/src/domain/execution-plan.js.map +1 -0
  208. package/dist/src/domain/images.js +2 -0
  209. package/dist/src/domain/images.js.map +1 -0
  210. package/dist/src/domain/index.js +13 -0
  211. package/dist/src/domain/index.js.map +1 -1
  212. package/dist/src/domain/interactions.js +2 -0
  213. package/dist/src/domain/interactions.js.map +1 -0
  214. package/dist/src/domain/providers.js +2 -0
  215. package/dist/src/domain/providers.js.map +1 -0
  216. package/dist/src/domain/resources.js +2 -0
  217. package/dist/src/domain/resources.js.map +1 -0
  218. package/dist/src/domain/runs.js.map +1 -1
  219. package/dist/src/domain/validation.js +11 -4
  220. package/dist/src/domain/validation.js.map +1 -1
  221. package/dist/src/domain/verifier-evidence-validation.js +214 -0
  222. package/dist/src/domain/verifier-evidence-validation.js.map +1 -0
  223. package/dist/src/domain/verifier-evidence.js +3 -0
  224. package/dist/src/domain/verifier-evidence.js.map +1 -0
  225. package/dist/src/domain/verifier-score-contract.js +232 -0
  226. package/dist/src/domain/verifier-score-contract.js.map +1 -0
  227. package/dist/src/domain/workers.js +2 -0
  228. package/dist/src/domain/workers.js.map +1 -0
  229. package/dist/src/evals/benchmark-adapter-manifest.js +219 -0
  230. package/dist/src/evals/benchmark-adapter-manifest.js.map +1 -0
  231. package/dist/src/evals/benchmark-candidate.js +36 -0
  232. package/dist/src/evals/benchmark-candidate.js.map +1 -0
  233. package/dist/src/evals/benchmark-run.js +166 -0
  234. package/dist/src/evals/benchmark-run.js.map +1 -0
  235. package/dist/src/evals/collect-only-rerun.js +183 -0
  236. package/dist/src/evals/collect-only-rerun.js.map +1 -0
  237. package/dist/src/evals/directory.js +30 -0
  238. package/dist/src/evals/directory.js.map +1 -0
  239. package/dist/src/evals/docker-engine-stats.js +89 -0
  240. package/dist/src/evals/docker-engine-stats.js.map +1 -0
  241. package/dist/src/evals/docker-ownership.js +51 -0
  242. package/dist/src/evals/docker-ownership.js.map +1 -0
  243. package/dist/src/evals/docker-reaper.js +144 -0
  244. package/dist/src/evals/docker-reaper.js.map +1 -0
  245. package/dist/src/evals/docker-resource-observer.js +207 -0
  246. package/dist/src/evals/docker-resource-observer.js.map +1 -0
  247. package/dist/src/evals/duration-estimator.js +126 -0
  248. package/dist/src/evals/duration-estimator.js.map +1 -0
  249. package/dist/src/evals/environment-image-planning.js +164 -0
  250. package/dist/src/evals/environment-image-planning.js.map +1 -0
  251. package/dist/src/evals/eval-artifact-planning.js +56 -0
  252. package/dist/src/evals/eval-artifact-planning.js.map +1 -0
  253. package/dist/src/evals/eval-finalization.js +10 -0
  254. package/dist/src/evals/eval-finalization.js.map +1 -0
  255. package/dist/src/evals/eval-lifecycle-events.js +29 -0
  256. package/dist/src/evals/eval-lifecycle-events.js.map +1 -0
  257. package/dist/src/evals/eval-logical-plan.js +51 -0
  258. package/dist/src/evals/eval-logical-plan.js.map +1 -0
  259. package/dist/src/evals/eval-result-builder.js +81 -0
  260. package/dist/src/evals/eval-result-builder.js.map +1 -0
  261. package/dist/src/evals/events.js +53 -5
  262. package/dist/src/evals/events.js.map +1 -1
  263. package/dist/src/evals/evolution-baseline.js +191 -0
  264. package/dist/src/evals/evolution-baseline.js.map +1 -0
  265. package/dist/src/evals/execution-evidence.js +116 -0
  266. package/dist/src/evals/execution-evidence.js.map +1 -0
  267. package/dist/src/evals/execution-leases.js +342 -0
  268. package/dist/src/evals/execution-leases.js.map +1 -0
  269. package/dist/src/evals/execution-plan-artifacts.js +72 -0
  270. package/dist/src/evals/execution-plan-artifacts.js.map +1 -0
  271. package/dist/src/evals/execution-plan-images.js +69 -0
  272. package/dist/src/evals/execution-plan-images.js.map +1 -0
  273. package/dist/src/evals/execution-plan-resources.js +174 -0
  274. package/dist/src/evals/execution-plan-resources.js.map +1 -0
  275. package/dist/src/evals/execution-plan.js +451 -0
  276. package/dist/src/evals/execution-plan.js.map +1 -0
  277. package/dist/src/evals/failure-classifier.js +61 -0
  278. package/dist/src/evals/failure-classifier.js.map +1 -0
  279. package/dist/src/evals/harbor-artifact-builder-image.js +99 -0
  280. package/dist/src/evals/harbor-artifact-builder-image.js.map +1 -0
  281. package/dist/src/evals/harbor-artifact-builder.js +358 -0
  282. package/dist/src/evals/harbor-artifact-builder.js.map +1 -0
  283. package/dist/src/evals/harbor-bridge-error.js +58 -3
  284. package/dist/src/evals/harbor-bridge-error.js.map +1 -1
  285. package/dist/src/evals/harbor-node-runtime.js +146 -0
  286. package/dist/src/evals/harbor-node-runtime.js.map +1 -0
  287. package/dist/src/evals/harbor-phase-timings.js +60 -0
  288. package/dist/src/evals/harbor-phase-timings.js.map +1 -0
  289. package/dist/src/evals/index.js +40 -2
  290. package/dist/src/evals/index.js.map +1 -1
  291. package/dist/src/evals/infrastructure-retry.js +208 -84
  292. package/dist/src/evals/infrastructure-retry.js.map +1 -1
  293. package/dist/src/evals/interaction-capture-import.js +49 -0
  294. package/dist/src/evals/interaction-capture-import.js.map +1 -0
  295. package/dist/src/evals/local-docker-provider.js +372 -0
  296. package/dist/src/evals/local-docker-provider.js.map +1 -0
  297. package/dist/src/evals/local-eval-planning.js +80 -0
  298. package/dist/src/evals/local-eval-planning.js.map +1 -0
  299. package/dist/src/evals/model-capture-plan.js +22 -0
  300. package/dist/src/evals/model-capture-plan.js.map +1 -0
  301. package/dist/src/evals/model-capture-runtime.js +114 -0
  302. package/dist/src/evals/model-capture-runtime.js.map +1 -0
  303. package/dist/src/evals/model-proxy-runtime-state.js +53 -0
  304. package/dist/src/evals/model-proxy-runtime-state.js.map +1 -0
  305. package/dist/src/evals/native-phase-evidence.js +440 -0
  306. package/dist/src/evals/native-phase-evidence.js.map +1 -0
  307. package/dist/src/evals/physical-retry-work.js +31 -0
  308. package/dist/src/evals/physical-retry-work.js.map +1 -0
  309. package/dist/src/evals/planned-execution-support.js +66 -0
  310. package/dist/src/evals/planned-execution-support.js.map +1 -0
  311. package/dist/src/evals/planned-execution.js +469 -0
  312. package/dist/src/evals/planned-execution.js.map +1 -0
  313. package/dist/src/evals/planned-progress-publisher.js +62 -0
  314. package/dist/src/evals/planned-progress-publisher.js.map +1 -0
  315. package/dist/src/evals/planned-remote-execution.js +40 -0
  316. package/dist/src/evals/planned-remote-execution.js.map +1 -0
  317. package/dist/src/evals/planned-retry-execution.js +174 -0
  318. package/dist/src/evals/planned-retry-execution.js.map +1 -0
  319. package/dist/src/evals/planned-retry-lifecycle.js +188 -0
  320. package/dist/src/evals/planned-retry-lifecycle.js.map +1 -0
  321. package/dist/src/evals/preparation-rerun.js +225 -0
  322. package/dist/src/evals/preparation-rerun.js.map +1 -0
  323. package/dist/src/evals/prepared-harness.js +76 -0
  324. package/dist/src/evals/prepared-harness.js.map +1 -0
  325. package/dist/src/evals/progress.js +52 -12
  326. package/dist/src/evals/progress.js.map +1 -1
  327. package/dist/src/evals/records.js +20 -3
  328. package/dist/src/evals/records.js.map +1 -1
  329. package/dist/src/evals/recovery.js +242 -0
  330. package/dist/src/evals/recovery.js.map +1 -0
  331. package/dist/src/evals/regrade-evidence.js +77 -0
  332. package/dist/src/evals/regrade-evidence.js.map +1 -0
  333. package/dist/src/evals/remote-infrastructure-retry.js +165 -0
  334. package/dist/src/evals/remote-infrastructure-retry.js.map +1 -0
  335. package/dist/src/evals/request.js +21 -3
  336. package/dist/src/evals/request.js.map +1 -1
  337. package/dist/src/evals/rerun-inputs.js +108 -0
  338. package/dist/src/evals/rerun-inputs.js.map +1 -0
  339. package/dist/src/evals/rerun-slots.js +4 -3
  340. package/dist/src/evals/rerun-slots.js.map +1 -1
  341. package/dist/src/evals/rerun-types.js +45 -0
  342. package/dist/src/evals/rerun-types.js.map +1 -0
  343. package/dist/src/evals/rerun.js +133 -88
  344. package/dist/src/evals/rerun.js.map +1 -1
  345. package/dist/src/evals/result-helpers.js +78 -0
  346. package/dist/src/evals/result-helpers.js.map +1 -0
  347. package/dist/src/evals/resume-state.js +36 -0
  348. package/dist/src/evals/resume-state.js.map +1 -0
  349. package/dist/src/evals/retry-backoff.js +20 -0
  350. package/dist/src/evals/retry-backoff.js.map +1 -0
  351. package/dist/src/evals/retry-state.js +195 -0
  352. package/dist/src/evals/retry-state.js.map +1 -0
  353. package/dist/src/evals/scheduler-metrics.js +95 -0
  354. package/dist/src/evals/scheduler-metrics.js.map +1 -0
  355. package/dist/src/evals/scheduler-trace-replay.js +107 -0
  356. package/dist/src/evals/scheduler-trace-replay.js.map +1 -0
  357. package/dist/src/evals/service-types.js +2 -0
  358. package/dist/src/evals/service-types.js.map +1 -0
  359. package/dist/src/evals/service.js +331 -347
  360. package/dist/src/evals/service.js.map +1 -1
  361. package/dist/src/evals/task-resources.js +209 -0
  362. package/dist/src/evals/task-resources.js.map +1 -0
  363. package/dist/src/evals/trial-environment-evidence.js +87 -0
  364. package/dist/src/evals/trial-environment-evidence.js.map +1 -0
  365. package/dist/src/evals/trial-execution-evidence.js +13 -0
  366. package/dist/src/evals/trial-execution-evidence.js.map +1 -0
  367. package/dist/src/evals/trial-import-identity.js +31 -0
  368. package/dist/src/evals/trial-import-identity.js.map +1 -0
  369. package/dist/src/evals/trial-import.js +183 -188
  370. package/dist/src/evals/trial-import.js.map +1 -1
  371. package/dist/src/evals/trial-publication-recovery.js +144 -0
  372. package/dist/src/evals/trial-publication-recovery.js.map +1 -0
  373. package/dist/src/evals/trial-publication.js +25 -0
  374. package/dist/src/evals/trial-publication.js.map +1 -0
  375. package/dist/src/evals/trial-reference-validation.js +49 -0
  376. package/dist/src/evals/trial-reference-validation.js.map +1 -0
  377. package/dist/src/evals/verifier-artifacts.js +317 -0
  378. package/dist/src/evals/verifier-artifacts.js.map +1 -0
  379. package/dist/src/evals/verifier-eligibility.js +25 -0
  380. package/dist/src/evals/verifier-eligibility.js.map +1 -0
  381. package/dist/src/evals/verifier-only-rerun.js +210 -0
  382. package/dist/src/evals/verifier-only-rerun.js.map +1 -0
  383. package/dist/src/evals/verifier-runtime.js +61 -0
  384. package/dist/src/evals/verifier-runtime.js.map +1 -0
  385. package/dist/src/evals/verifier-score-artifacts.js +119 -0
  386. package/dist/src/evals/verifier-score-artifacts.js.map +1 -0
  387. package/dist/src/evals/work-item-artifacts.js +23 -0
  388. package/dist/src/evals/work-item-artifacts.js.map +1 -0
  389. package/dist/src/foundation/config.js +17 -0
  390. package/dist/src/foundation/config.js.map +1 -1
  391. package/dist/src/foundation/contained-file.js +85 -0
  392. package/dist/src/foundation/contained-file.js.map +1 -0
  393. package/dist/src/foundation/credential-redaction.js +162 -0
  394. package/dist/src/foundation/credential-redaction.js.map +1 -0
  395. package/dist/src/foundation/environment-image-references.js +59 -0
  396. package/dist/src/foundation/environment-image-references.js.map +1 -0
  397. package/dist/src/foundation/executable.js +45 -5
  398. package/dist/src/foundation/executable.js.map +1 -1
  399. package/dist/src/foundation/fs.js +2 -1
  400. package/dist/src/foundation/fs.js.map +1 -1
  401. package/dist/src/foundation/index.js +6 -2
  402. package/dist/src/foundation/index.js.map +1 -1
  403. package/dist/src/foundation/process-identity.js +69 -0
  404. package/dist/src/foundation/process-identity.js.map +1 -0
  405. package/dist/src/images/context.js +47 -0
  406. package/dist/src/images/context.js.map +1 -0
  407. package/dist/src/images/docker-buildkit.js +137 -0
  408. package/dist/src/images/docker-buildkit.js.map +1 -0
  409. package/dist/src/images/dockerfile.js +67 -0
  410. package/dist/src/images/dockerfile.js.map +1 -0
  411. package/dist/src/images/gc.js +261 -0
  412. package/dist/src/images/gc.js.map +1 -0
  413. package/dist/src/images/index.js +10 -0
  414. package/dist/src/images/index.js.map +1 -0
  415. package/dist/src/images/manifest.js +109 -0
  416. package/dist/src/images/manifest.js.map +1 -0
  417. package/dist/src/images/ownership.js +5 -0
  418. package/dist/src/images/ownership.js.map +1 -0
  419. package/dist/src/images/records.js +32 -0
  420. package/dist/src/images/records.js.map +1 -0
  421. package/dist/src/images/registry.js +154 -0
  422. package/dist/src/images/registry.js.map +1 -0
  423. package/dist/src/images/service.js +224 -0
  424. package/dist/src/images/service.js.map +1 -0
  425. package/dist/src/model-access/capture.js +255 -0
  426. package/dist/src/model-access/capture.js.map +1 -0
  427. package/dist/src/model-access/index.js +5 -0
  428. package/dist/src/model-access/index.js.map +1 -0
  429. package/dist/src/model-access/policy.js +49 -0
  430. package/dist/src/model-access/policy.js.map +1 -0
  431. package/dist/src/model-access/proxy.js +336 -0
  432. package/dist/src/model-access/proxy.js.map +1 -0
  433. package/dist/src/model-access/records.js +167 -0
  434. package/dist/src/model-access/records.js.map +1 -0
  435. package/dist/src/runs/adapter-process.js +12 -0
  436. package/dist/src/runs/adapter-process.js.map +1 -0
  437. package/dist/src/runs/bundle.js +450 -0
  438. package/dist/src/runs/bundle.js.map +1 -0
  439. package/dist/src/runs/events.js +12 -1
  440. package/dist/src/runs/events.js.map +1 -1
  441. package/dist/src/runs/executor.js +46 -45
  442. package/dist/src/runs/executor.js.map +1 -1
  443. package/dist/src/runs/finalizer.js +17 -0
  444. package/dist/src/runs/finalizer.js.map +1 -1
  445. package/dist/src/runs/index.js +8 -1
  446. package/dist/src/runs/index.js.map +1 -1
  447. package/dist/src/runs/manifest.js +5 -5
  448. package/dist/src/runs/manifest.js.map +1 -1
  449. package/dist/src/runs/phase-bundle.js +86 -0
  450. package/dist/src/runs/phase-bundle.js.map +1 -0
  451. package/dist/src/runs/phase-cancellation.js +107 -0
  452. package/dist/src/runs/phase-cancellation.js.map +1 -0
  453. package/dist/src/runs/phase-group.js +113 -0
  454. package/dist/src/runs/phase-group.js.map +1 -0
  455. package/dist/src/runs/query.js +1 -1
  456. package/dist/src/runs/query.js.map +1 -1
  457. package/dist/src/runs/records.js +3 -0
  458. package/dist/src/runs/records.js.map +1 -1
  459. package/dist/src/runs/request.js +10 -3
  460. package/dist/src/runs/request.js.map +1 -1
  461. package/dist/src/runs/training-candidate.js +161 -0
  462. package/dist/src/runs/training-candidate.js.map +1 -0
  463. package/dist/src/runs/verifier-evidence-redaction.js +61 -0
  464. package/dist/src/runs/verifier-evidence-redaction.js.map +1 -0
  465. package/dist/src/runs/verifier-evidence.js +459 -0
  466. package/dist/src/runs/verifier-evidence.js.map +1 -0
  467. package/dist/src/runs/verifier-structured-evidence.js +83 -0
  468. package/dist/src/runs/verifier-structured-evidence.js.map +1 -0
  469. package/dist/src/trajectories/analysis.js +288 -0
  470. package/dist/src/trajectories/analysis.js.map +1 -0
  471. package/dist/src/trajectories/chunk-projection.js +148 -0
  472. package/dist/src/trajectories/chunk-projection.js.map +1 -0
  473. package/dist/src/trajectories/content-projection.js +389 -0
  474. package/dist/src/trajectories/content-projection.js.map +1 -0
  475. package/dist/src/trajectories/dsh-chunk-contract.js +176 -0
  476. package/dist/src/trajectories/dsh-chunk-contract.js.map +1 -0
  477. package/dist/src/trajectories/dsh-contract.js +284 -0
  478. package/dist/src/trajectories/dsh-contract.js.map +1 -0
  479. package/dist/src/trajectories/events-chunk-drill.js +95 -0
  480. package/dist/src/trajectories/events-chunk-drill.js.map +1 -0
  481. package/dist/src/trajectories/events-page.js +367 -0
  482. package/dist/src/trajectories/events-page.js.map +1 -0
  483. package/dist/src/trajectories/index.js +3 -0
  484. package/dist/src/trajectories/index.js.map +1 -1
  485. package/dist/src/trajectories/provider-capture.js +15 -24
  486. package/dist/src/trajectories/provider-capture.js.map +1 -1
  487. package/dist/src/trajectories/providers/deepseek.js +3 -3
  488. package/dist/src/trajectories/providers/deepseek.js.map +1 -1
  489. package/dist/src/trajectories/request-attempt.js +47 -0
  490. package/dist/src/trajectories/request-attempt.js.map +1 -0
  491. package/dist/src/trajectories/stream-reader.js +311 -0
  492. package/dist/src/trajectories/stream-reader.js.map +1 -0
  493. package/dist/src/trajectories/surface-fold.js +126 -0
  494. package/dist/src/trajectories/surface-fold.js.map +1 -0
  495. package/dist/src/workers/index.js +3 -0
  496. package/dist/src/workers/index.js.map +1 -0
  497. package/dist/src/workers/remote-harbor-work-spec.js +97 -0
  498. package/dist/src/workers/remote-harbor-work-spec.js.map +1 -0
  499. package/dist/src/workers/remote-harbor-worker.js +223 -0
  500. package/dist/src/workers/remote-harbor-worker.js.map +1 -0
  501. package/docs/schemas/benchmark-hook-request-v1.schema.json +65 -0
  502. package/docs/schemas/benchmark-hook-response-v1.schema.json +75 -0
  503. package/docs/schemas/benchmark-hook-v1.schema.json +33 -0
  504. package/docs/schemas/benchmark-lock-v1.schema.json +351 -0
  505. package/docs/schemas/benchmark-metric-v1.schema.json +37 -0
  506. package/docs/schemas/benchmark-package-v1.schema.json +167 -0
  507. package/docs/schemas/benchmark-phase-group.schema.json +41 -0
  508. package/docs/schemas/benchmark-profile-v1.schema.json +141 -0
  509. package/docs/schemas/benchmark-task-v1.schema.json +384 -0
  510. package/docs/schemas/benchmark-tool-result-v1.schema.json +30 -0
  511. package/docs/schemas/environment-build-index.schema.json +13 -0
  512. package/docs/schemas/environment-build-record.schema.json +25 -0
  513. package/docs/schemas/environment-image-manifest.schema.json +62 -0
  514. package/docs/schemas/error.schema.json +22 -0
  515. package/docs/schemas/eval-control.schema.json +36 -0
  516. package/docs/schemas/eval-execution-plan.schema.json +216 -0
  517. package/docs/schemas/eval-progress.schema.json +1 -15
  518. package/docs/schemas/eval-rerun-result.schema.json +58 -0
  519. package/docs/schemas/eval-rerun-state.schema.json +61 -0
  520. package/docs/schemas/eval-rerun-submission.schema.json +60 -0
  521. package/docs/schemas/eval-result.schema.json +32 -13
  522. package/docs/schemas/eval-submission-input.schema.json +14 -0
  523. package/docs/schemas/eval-submission.schema.json +68 -0
  524. package/docs/schemas/eval-trial-publication.schema.json +15 -0
  525. package/docs/schemas/eval-trial-reference.schema.json +129 -0
  526. package/docs/schemas/execution-evidence.schema.json +96 -0
  527. package/docs/schemas/execution-lease.schema.json +50 -0
  528. package/docs/schemas/execution-provider-status.schema.json +65 -0
  529. package/docs/schemas/execution-worker.schema.json +53 -0
  530. package/docs/schemas/harbor-process-exit.schema.json +14 -0
  531. package/docs/schemas/interaction-capture-ref.schema.json +45 -0
  532. package/docs/schemas/local-provider-execution.schema.json +33 -0
  533. package/docs/schemas/model-interaction.schema.json +41 -0
  534. package/docs/schemas/regrade-assessment-reference.schema.json +11 -0
  535. package/docs/schemas/remote-result-envelope.schema.json +32 -0
  536. package/docs/schemas/remote-tree-envelope.schema.json +25 -0
  537. package/docs/schemas/remote-work-offer.schema.json +58 -0
  538. package/docs/schemas/remote-work-receipt.schema.json +50 -0
  539. package/docs/schemas/remote-worker-artifact.schema.json +17 -0
  540. package/docs/schemas/remote-worker-event.schema.json +18 -0
  541. package/docs/schemas/remote-worker-heartbeat.schema.json +37 -0
  542. package/docs/schemas/remote-worker-registration.schema.json +52 -0
  543. package/docs/schemas/result-bundle-index.schema.json +133 -0
  544. package/docs/schemas/run-context.schema.json +15 -0
  545. package/docs/schemas/run-environment-images.schema.json +39 -0
  546. package/docs/schemas/run-manifest.schema.json +4 -0
  547. package/docs/schemas/run-request.schema.json +4 -0
  548. package/docs/schemas/training-data-candidate.schema.json +46 -0
  549. package/docs/schemas/trajectory-analysis.schema.json +258 -0
  550. package/docs/schemas/trajectory-events-page.schema.json +40 -0
  551. package/docs/schemas/verifier-assessment.schema.json +77 -0
  552. package/docs/schemas/verifier-evidence.schema.json +276 -0
  553. package/integrations/harbor/hitch_benchmark.py +252 -0
  554. package/integrations/harbor/hitch_candidate_recycle.py +262 -0
  555. package/integrations/harbor/hitch_harbor_agent.py +698 -568
  556. package/integrations/harbor/hitch_harbor_environment.py +469 -0
  557. package/integrations/harbor/hitch_harbor_task_resources.py +275 -0
  558. package/integrations/harbor/hitch_harbor_verifier.py +83 -2
  559. package/integrations/harbor/hitch_phase_supervisor.py +452 -0
  560. package/integrations/harbor/hitch_tool_client.mjs +91 -0
  561. package/integrations/model-call/cli.js +57 -0
  562. package/package.json +30 -6
@@ -0,0 +1,36 @@
1
+ {
2
+ "$schema": "https://json-schema.org/draft/2020-12/schema",
3
+ "$id": "https://agent-hitch.local/schemas/eval-control.schema.json",
4
+ "title": "Hitch control-plane eval state V1",
5
+ "type": "object",
6
+ "additionalProperties": false,
7
+ "required": ["schema_version", "eval_id", "generation", "state", "requested_parallelism", "admitted_parallelism", "active_leases", "queued_work_items", "terminal_work_items", "created_at", "updated_at"],
8
+ "properties": {
9
+ "schema_version": { "const": "1" },
10
+ "eval_id": { "type": "string", "pattern": "^eval_[a-f0-9]{32}$" },
11
+ "generation": { "type": "integer", "minimum": 0 },
12
+ "state": { "enum": ["queued", "planning", "preparing", "running", "finalizing", "cancelling", "succeeded", "failed", "cancelled"] },
13
+ "requested_parallelism": { "type": "integer", "minimum": 1 },
14
+ "admitted_parallelism": { "type": "integer", "minimum": 0 },
15
+ "active_leases": { "$ref": "#/$defs/leases" },
16
+ "queued_work_items": { "$ref": "#/$defs/workItems" },
17
+ "terminal_work_items": { "$ref": "#/$defs/workItems" },
18
+ "allocation_id": { "type": "string", "pattern": "^allocation_[a-f0-9]{32}$" },
19
+ "cancel_requested_at": { "type": "string", "format": "date-time" },
20
+ "error": {
21
+ "type": "object",
22
+ "additionalProperties": false,
23
+ "required": ["code", "message"],
24
+ "properties": {
25
+ "code": { "type": "string", "minLength": 1, "maxLength": 256 },
26
+ "message": { "type": "string", "maxLength": 4096 }
27
+ }
28
+ },
29
+ "created_at": { "type": "string", "format": "date-time" },
30
+ "updated_at": { "type": "string", "format": "date-time" }
31
+ },
32
+ "$defs": {
33
+ "leases": { "type": "array", "items": { "type": "string", "pattern": "^lease_[a-f0-9]{32}$" }, "uniqueItems": true },
34
+ "workItems": { "type": "array", "items": { "type": "string", "pattern": "^work_[a-f0-9]{32}$" }, "uniqueItems": true }
35
+ }
36
+ }
@@ -0,0 +1,216 @@
1
+ {
2
+ "$schema": "https://json-schema.org/draft/2020-12/schema",
3
+ "$id": "https://agent-hitch.local/schemas/eval-execution-plan.schema.json",
4
+ "title": "Hitch eval execution plan V1",
5
+ "type": "object",
6
+ "additionalProperties": false,
7
+ "required": [
8
+ "schema_version", "planner", "eval_id", "membership", "candidate_identity", "benchmark",
9
+ "provider", "max_parallelism", "default_trial_resources", "slots", "work_items", "retry_policy", "created_at"
10
+ ],
11
+ "properties": {
12
+ "schema_version": { "const": "1" },
13
+ "planner": { "const": "hitch-local-v1" },
14
+ "eval_id": { "$ref": "#/$defs/evalId" },
15
+ "membership": { "enum": ["known", "opaque"] },
16
+ "candidate_identity": { "$ref": "#/$defs/sha256" },
17
+ "benchmark": {
18
+ "type": "object",
19
+ "additionalProperties": false,
20
+ "required": ["id", "revision", "verifier_identity"],
21
+ "properties": {
22
+ "id": { "type": "string", "minLength": 1 },
23
+ "revision": { "type": "string", "minLength": 1 },
24
+ "verifier_identity": { "$ref": "#/$defs/sha256" }
25
+ }
26
+ },
27
+ "provider": { "type": "string", "minLength": 1 },
28
+ "model_capture": { "$ref": "#/$defs/modelCapture" },
29
+ "max_parallelism": { "type": "integer", "minimum": 1 },
30
+ "default_trial_resources": { "$ref": "#/$defs/resources" },
31
+ "task_resources": {
32
+ "type": "array",
33
+ "items": { "$ref": "#/$defs/taskResources" }
34
+ },
35
+ "image_fallbacks": {
36
+ "type": "array",
37
+ "items": { "$ref": "#/$defs/imageFallback" }
38
+ },
39
+ "slots": {
40
+ "type": "array",
41
+ "items": {
42
+ "type": "object",
43
+ "additionalProperties": false,
44
+ "required": ["schema_version", "slot_id", "eval_id", "task_id", "attempt", "candidate_identity", "state", "physical_execution"],
45
+ "properties": {
46
+ "schema_version": { "const": "1" },
47
+ "slot_id": { "type": "string", "pattern": "^slot_[a-f0-9]{32}$" },
48
+ "eval_id": { "$ref": "#/$defs/evalId" },
49
+ "task_id": { "type": "string", "minLength": 1 },
50
+ "task_digest": { "$ref": "#/$defs/sha256" },
51
+ "attempt": { "type": "integer", "minimum": 1 },
52
+ "candidate_identity": { "$ref": "#/$defs/sha256" },
53
+ "state": { "enum": ["pending", "blocked", "ready", "leased", "running", "collecting", "succeeded", "invalid", "failed", "cancelled"] },
54
+ "physical_execution": { "type": "integer", "minimum": 1 },
55
+ "authoritative_run_id": { "type": "string", "pattern": "^run_[a-f0-9]{32}$" },
56
+ "invalid_reason": { "type": "string", "minLength": 1 }
57
+ }
58
+ }
59
+ },
60
+ "work_items": {
61
+ "type": "array",
62
+ "items": {
63
+ "type": "object",
64
+ "additionalProperties": false,
65
+ "required": [
66
+ "schema_version", "work_id", "eval_id", "backend", "logical_attempt", "task_ids", "slots",
67
+ "opaque_membership", "requested_parallelism", "reservation", "provider"
68
+ ],
69
+ "properties": {
70
+ "schema_version": { "const": "1" },
71
+ "work_id": { "type": "string", "pattern": "^work_[a-f0-9]{32}$" },
72
+ "eval_id": { "$ref": "#/$defs/evalId" },
73
+ "backend": { "const": "harbor" },
74
+ "logical_attempt": { "type": ["integer", "null"], "minimum": 1 },
75
+ "task_ids": { "type": "array", "items": { "type": "string", "minLength": 1 }, "uniqueItems": true },
76
+ "slots": { "type": "array", "items": { "type": "string", "pattern": "^slot_[a-f0-9]{32}$" }, "uniqueItems": true },
77
+ "opaque_membership": { "type": "boolean" },
78
+ "requested_parallelism": { "type": "integer", "minimum": 1 },
79
+ "reservation": { "$ref": "#/$defs/resources" },
80
+ "provider": { "type": "string", "minLength": 1 },
81
+ "image_refs": {
82
+ "type": "array",
83
+ "items": { "$ref": "#/$defs/imageUse" }
84
+ }
85
+ }
86
+ }
87
+ },
88
+ "retry_policy": {
89
+ "type": "object",
90
+ "additionalProperties": false,
91
+ "required": ["infrastructure_retries", "infrastructure_retry_backoff_ms", "verifier_execution", "candidate_rerun_on_verifier_failure"],
92
+ "properties": {
93
+ "infrastructure_retries": { "type": "integer", "minimum": 0 },
94
+ "infrastructure_retry_backoff_ms": { "type": "number", "minimum": 0 },
95
+ "verifier_execution": { "const": "same-trial-verifier-only" },
96
+ "candidate_rerun_on_verifier_failure": { "const": false }
97
+ }
98
+ },
99
+ "created_at": { "type": "string", "format": "date-time" }
100
+ },
101
+ "$defs": {
102
+ "evalId": { "type": "string", "pattern": "^eval_[a-f0-9]{32}$" },
103
+ "sha256": { "type": "string", "pattern": "^sha256:[a-f0-9]{64}$" },
104
+ "modelCapture": {
105
+ "type": "object",
106
+ "additionalProperties": false,
107
+ "required": ["requested_mode", "effective_mode", "required"],
108
+ "properties": {
109
+ "requested_mode": { "enum": ["off", "native", "proxy", "hybrid"] },
110
+ "effective_mode": { "enum": ["off", "native", "proxy", "hybrid"] },
111
+ "required": { "type": "boolean" },
112
+ "topology": { "enum": ["host-side", "in-sandbox"] },
113
+ "degraded_reason": { "type": "string", "minLength": 1 }
114
+ }
115
+ },
116
+ "resources": {
117
+ "type": "object",
118
+ "additionalProperties": false,
119
+ "required": ["cpu_millis", "memory_bytes", "container_slots", "build_slots"],
120
+ "properties": {
121
+ "cpu_millis": { "type": "integer", "minimum": 0 },
122
+ "memory_bytes": { "type": "integer", "minimum": 0 },
123
+ "container_slots": { "type": "integer", "minimum": 0 },
124
+ "build_slots": { "type": "integer", "minimum": 0 },
125
+ "gpu_count": { "type": "integer", "minimum": 0 },
126
+ "ephemeral_disk_bytes": { "type": "integer", "minimum": 0 }
127
+ }
128
+ },
129
+ "resourceField": {
130
+ "type": "object",
131
+ "additionalProperties": false,
132
+ "required": ["value", "source", "estimated"],
133
+ "properties": {
134
+ "value": { "type": "integer", "minimum": 0 },
135
+ "source": { "enum": ["task", "compose", "submission-default", "operator-default", "provider-policy", "derived-components"] },
136
+ "estimated": { "type": "boolean" }
137
+ }
138
+ },
139
+ "imageUse": {
140
+ "type": "object",
141
+ "additionalProperties": false,
142
+ "required": ["task_ids", "image_id", "requested_reference", "reference", "manifest_digest", "platform", "resolution", "cache_hit"],
143
+ "properties": {
144
+ "task_ids": { "type": "array", "minItems": 1, "uniqueItems": true, "items": { "type": "string", "minLength": 1 } },
145
+ "image_id": { "$ref": "#/$defs/sha256" },
146
+ "requested_reference": { "type": "string", "minLength": 1, "maxLength": 1024 },
147
+ "reference": { "type": "string", "minLength": 1, "maxLength": 1024 },
148
+ "manifest_digest": { "$ref": "#/$defs/sha256" },
149
+ "platform": { "type": "string", "minLength": 1 },
150
+ "resolution": { "enum": ["registry", "prebuilt", "backend-build"] },
151
+ "cache_hit": { "type": "boolean" }
152
+ }
153
+ },
154
+ "imageFallback": {
155
+ "type": "object",
156
+ "additionalProperties": false,
157
+ "required": ["task_id", "source", "service", "code"],
158
+ "properties": {
159
+ "task_id": { "type": "string", "minLength": 1 },
160
+ "source": { "enum": ["task", "verifier", "compose"] },
161
+ "service": { "type": "string", "minLength": 1 },
162
+ "code": { "enum": ["backend-build", "dynamic-image", "policy-backend", "resolver-unavailable", "resolution-failed"] }
163
+ }
164
+ },
165
+ "taskResources": {
166
+ "type": "object",
167
+ "additionalProperties": false,
168
+ "required": ["task_id", "reservation", "main_limits", "fields", "components", "diagnostics"],
169
+ "properties": {
170
+ "task_id": { "type": "string", "minLength": 1 },
171
+ "reservation": { "$ref": "#/$defs/resources" },
172
+ "main_limits": { "$ref": "#/$defs/resources" },
173
+ "fields": {
174
+ "type": "object",
175
+ "additionalProperties": false,
176
+ "required": ["cpu_millis", "memory_bytes", "container_slots", "build_slots"],
177
+ "properties": {
178
+ "cpu_millis": { "$ref": "#/$defs/resourceField" },
179
+ "memory_bytes": { "$ref": "#/$defs/resourceField" },
180
+ "container_slots": { "$ref": "#/$defs/resourceField" },
181
+ "build_slots": { "$ref": "#/$defs/resourceField" },
182
+ "gpu_count": { "$ref": "#/$defs/resourceField" },
183
+ "ephemeral_disk_bytes": { "$ref": "#/$defs/resourceField" }
184
+ }
185
+ },
186
+ "components": {
187
+ "type": "array",
188
+ "minItems": 1,
189
+ "items": {
190
+ "type": "object",
191
+ "additionalProperties": false,
192
+ "required": ["name", "role", "replicas", "resources", "fields"],
193
+ "properties": {
194
+ "name": { "type": "string", "minLength": 1 },
195
+ "role": { "enum": ["main", "task-sidecar", "verifier", "provider-sidecar"] },
196
+ "replicas": { "type": "integer", "minimum": 1 },
197
+ "resources": { "$ref": "#/$defs/resources" },
198
+ "fields": {
199
+ "type": "object",
200
+ "additionalProperties": false,
201
+ "required": ["cpu_millis", "memory_bytes"],
202
+ "properties": {
203
+ "cpu_millis": { "$ref": "#/$defs/resourceField" },
204
+ "memory_bytes": { "$ref": "#/$defs/resourceField" },
205
+ "gpu_count": { "$ref": "#/$defs/resourceField" },
206
+ "ephemeral_disk_bytes": { "$ref": "#/$defs/resourceField" }
207
+ }
208
+ }
209
+ }
210
+ }
211
+ },
212
+ "diagnostics": { "type": "array", "items": { "type": "string", "minLength": 1 }, "uniqueItems": true }
213
+ }
214
+ }
215
+ }
216
+ }
@@ -29,21 +29,7 @@
29
29
  "planned_trials": { "type": ["integer", "null"], "minimum": 0 },
30
30
  "trials": {
31
31
  "type": "array",
32
- "items": {
33
- "type": "object",
34
- "additionalProperties": false,
35
- "required": ["trial_id", "run_id", "task_id", "attempt", "observation_status"],
36
- "properties": {
37
- "trial_id": { "type": "string", "minLength": 1 },
38
- "run_id": { "type": "string", "pattern": "^run_[a-f0-9]{32}$" },
39
- "task_id": { "type": "string", "minLength": 1 },
40
- "attempt": { "type": "integer", "minimum": 1 },
41
- "observation_status": { "enum": ["valid", "invalid"] },
42
- "reward": { "type": "number" },
43
- "verifier_result_ref": { "type": "string", "minLength": 1 },
44
- "invalid_reason": { "type": "string", "minLength": 1 }
45
- }
46
- }
32
+ "items": { "$ref": "eval-trial-reference.schema.json" }
47
33
  },
48
34
  "summary": {
49
35
  "type": "object",
@@ -0,0 +1,58 @@
1
+ {
2
+ "$schema": "https://json-schema.org/draft/2020-12/schema",
3
+ "$id": "https://agent-hitch.local/schemas/eval-rerun-result.schema.json",
4
+ "title": "Hitch eval rerun result V1",
5
+ "type": "object",
6
+ "additionalProperties": false,
7
+ "required": [
8
+ "schema_version", "kind", "rerun_id", "rerun_type", "semantics", "eval_id", "status",
9
+ "selected_tasks", "repaired_tasks", "remaining_invalid_tasks", "selected_trials", "repaired_trials",
10
+ "remaining_invalid_trials", "eval_status", "started_at", "completed_at"
11
+ ],
12
+ "properties": {
13
+ "schema_version": { "const": "1" },
14
+ "kind": { "const": "eval-rerun" },
15
+ "rerun_id": { "type": "string", "pattern": "^rerun_[a-f0-9]{32}$" },
16
+ "rerun_type": { "enum": ["candidate-restart", "candidate-resume", "trajectory-replay", "verifier-only", "collect-only"] },
17
+ "semantics": { "$ref": "eval-rerun-submission.schema.json#/$defs/semantics" },
18
+ "eval_id": { "type": "string", "pattern": "^eval_[a-f0-9]{32}$" },
19
+ "status": { "const": "completed" },
20
+ "selected_tasks": { "type": "array", "items": { "type": "string" } },
21
+ "repaired_tasks": { "type": "array", "items": { "type": "string" } },
22
+ "remaining_invalid_tasks": { "type": "array", "items": { "type": "string" } },
23
+ "selected_trials": { "type": "array", "items": { "$ref": "#/$defs/trial" } },
24
+ "repaired_trials": { "type": "array", "items": { "$ref": "#/$defs/trial" } },
25
+ "remaining_invalid_trials": { "type": "array", "items": { "$ref": "#/$defs/trial" } },
26
+ "sources": { "type": "array", "items": { "$ref": "#/$defs/source" } },
27
+ "eval_status": { "enum": ["succeeded", "failed"] },
28
+ "started_at": { "type": "string", "format": "date-time" },
29
+ "completed_at": { "type": "string", "format": "date-time" }
30
+ },
31
+ "$defs": {
32
+ "trial": {
33
+ "type": "object",
34
+ "additionalProperties": false,
35
+ "required": ["task_id", "attempt"],
36
+ "properties": {
37
+ "task_id": { "type": "string", "minLength": 1 },
38
+ "attempt": { "type": "integer", "minimum": 1 }
39
+ }
40
+ },
41
+ "source": {
42
+ "type": "object",
43
+ "additionalProperties": false,
44
+ "required": ["source_trial_id", "source_work_id", "source_backend_directory"],
45
+ "oneOf": [
46
+ { "required": ["source_run_id"], "not": { "required": ["source_run_group_id"] } },
47
+ { "required": ["source_run_group_id"], "not": { "required": ["source_run_id"] } }
48
+ ],
49
+ "properties": {
50
+ "source_trial_id": { "type": "string", "minLength": 1 },
51
+ "source_run_id": { "type": "string", "pattern": "^run_[a-f0-9]{32}$" },
52
+ "source_run_group_id": { "type": "string", "pattern": "^run_group_[a-f0-9]{32}$" },
53
+ "source_work_id": { "type": "string", "pattern": "^work_[a-f0-9]{32}$" },
54
+ "source_backend_directory": { "type": "string", "pattern": "^harbor/work-items/work_[a-f0-9]{32}/epoch-[0-9]{6}$" }
55
+ }
56
+ }
57
+ }
58
+ }
@@ -0,0 +1,61 @@
1
+ {
2
+ "$schema": "https://json-schema.org/draft/2020-12/schema",
3
+ "$id": "https://agent-hitch.local/schemas/eval-rerun-state.schema.json",
4
+ "title": "Hitch eval rerun state V1",
5
+ "type": "object",
6
+ "additionalProperties": false,
7
+ "required": ["schema_version", "rerun_id", "eval_id", "rerun_type", "semantics", "status", "tasks", "repaired_tasks", "updated_at"],
8
+ "properties": {
9
+ "schema_version": { "const": "1" },
10
+ "rerun_id": { "type": "string", "pattern": "^rerun_[a-f0-9]{32}$" },
11
+ "eval_id": { "type": "string", "pattern": "^eval_[a-f0-9]{32}$" },
12
+ "rerun_type": { "enum": ["candidate-restart", "candidate-resume", "trajectory-replay", "verifier-only", "collect-only"] },
13
+ "semantics": { "$ref": "eval-rerun-submission.schema.json#/$defs/semantics" },
14
+ "status": { "enum": ["queued", "running", "completed", "failed", "cancelled"] },
15
+ "tasks": { "type": "array", "items": { "type": "string" } },
16
+ "trials": { "type": "array", "items": { "$ref": "#/$defs/trial" } },
17
+ "repaired_tasks": { "type": "array", "items": { "type": "string" } },
18
+ "repaired_trials": { "type": "array", "items": { "$ref": "#/$defs/trial" } },
19
+ "remaining_invalid_tasks": { "type": "array", "items": { "type": "string" } },
20
+ "remaining_invalid_trials": { "type": "array", "items": { "$ref": "#/$defs/trial" } },
21
+ "sources": { "type": "array", "items": { "$ref": "#/$defs/source" } },
22
+ "eval_status": { "enum": ["succeeded", "failed"] },
23
+ "admitted_parallelism": { "type": "integer", "minimum": 1 },
24
+ "allocation_id": { "type": "string", "pattern": "^allocation_[a-f0-9]{32}$" },
25
+ "error": {
26
+ "type": "object",
27
+ "additionalProperties": false,
28
+ "required": ["code"],
29
+ "properties": {
30
+ "code": { "type": "string", "minLength": 1 },
31
+ "message": { "type": "string", "maxLength": 4096 }
32
+ }
33
+ },
34
+ "submitted_at": { "type": "string", "format": "date-time" },
35
+ "started_at": { "type": "string", "format": "date-time" },
36
+ "completed_at": { "type": "string", "format": "date-time" },
37
+ "updated_at": { "type": "string", "format": "date-time" }
38
+ },
39
+ "$defs": {
40
+ "trial": {
41
+ "type": "object",
42
+ "additionalProperties": false,
43
+ "required": ["task_id", "attempt"],
44
+ "properties": {
45
+ "task_id": { "type": "string", "minLength": 1 },
46
+ "attempt": { "type": "integer", "minimum": 1 }
47
+ }
48
+ },
49
+ "source": {
50
+ "type": "object",
51
+ "additionalProperties": false,
52
+ "required": ["source_trial_id", "source_run_id", "source_work_id", "source_backend_directory"],
53
+ "properties": {
54
+ "source_trial_id": { "type": "string", "minLength": 1 },
55
+ "source_run_id": { "type": "string", "pattern": "^run_[a-f0-9]{32}$" },
56
+ "source_work_id": { "type": "string", "pattern": "^work_[a-f0-9]{32}$" },
57
+ "source_backend_directory": { "type": "string", "pattern": "^harbor/work-items/work_[a-f0-9]{32}/epoch-[0-9]{6}$" }
58
+ }
59
+ }
60
+ }
61
+ }
@@ -0,0 +1,60 @@
1
+ {
2
+ "$schema": "https://json-schema.org/draft/2020-12/schema",
3
+ "$id": "https://agent-hitch.local/schemas/eval-rerun-submission.schema.json",
4
+ "title": "Hitch eval rerun submission V1",
5
+ "type": "object",
6
+ "additionalProperties": false,
7
+ "required": ["schema_version", "rerun_id", "eval_id", "rerun_type", "semantics", "selector", "submitted_at"],
8
+ "properties": {
9
+ "schema_version": { "const": "1" },
10
+ "rerun_id": { "type": "string", "pattern": "^rerun_[a-f0-9]{32}$" },
11
+ "eval_id": { "type": "string", "pattern": "^eval_[a-f0-9]{32}$" },
12
+ "rerun_type": { "$ref": "#/$defs/rerunType" },
13
+ "verifier_runtime_id": { "type": "string", "pattern": "^sha256:[a-f0-9]{64}$" },
14
+ "semantics": { "$ref": "#/$defs/semantics" },
15
+ "selector": { "$ref": "#/$defs/selector" },
16
+ "submitted_at": { "type": "string", "format": "date-time" }
17
+ },
18
+ "allOf": [{ "if": { "required": ["verifier_runtime_id"] }, "then": { "properties": { "rerun_type": { "const": "verifier-only" } } } }],
19
+ "$defs": {
20
+ "rerunType": {
21
+ "enum": ["candidate-restart", "candidate-resume", "trajectory-replay", "verifier-only", "collect-only"]
22
+ },
23
+ "semantics": {
24
+ "type": "object",
25
+ "additionalProperties": false,
26
+ "required": ["candidate_action", "conversation_source", "sandbox_source", "candidate_executes"],
27
+ "properties": {
28
+ "candidate_action": { "enum": ["restart", "resume", "replay", "none"] },
29
+ "conversation_source": { "enum": ["original-instruction", "native-session", "canonical-trajectory", "none"] },
30
+ "sandbox_source": { "enum": ["clean", "checkpoint", "retained", "recorded-artifacts", "none"] },
31
+ "candidate_executes": { "type": "boolean" }
32
+ }
33
+ },
34
+ "selector": {
35
+ "oneOf": [
36
+ {
37
+ "type": "object",
38
+ "additionalProperties": false,
39
+ "required": ["mode"],
40
+ "properties": { "mode": { "const": "invalid" } }
41
+ },
42
+ {
43
+ "type": "object",
44
+ "additionalProperties": false,
45
+ "required": ["mode", "task_names"],
46
+ "properties": {
47
+ "mode": { "const": "tasks" },
48
+ "task_names": {
49
+ "type": "array",
50
+ "minItems": 1,
51
+ "maxItems": 10000,
52
+ "uniqueItems": true,
53
+ "items": { "type": "string", "minLength": 1, "maxLength": 1024 }
54
+ }
55
+ }
56
+ }
57
+ ]
58
+ }
59
+ }
60
+ }
@@ -37,28 +37,47 @@
37
37
  "harness_id": { "type": "string", "pattern": "^[a-z][a-z0-9-]*$" },
38
38
  "revision_identity": { "type": "string", "pattern": "^sha256:[0-9a-f]{64}$" },
39
39
  "platform": { "type": "string", "pattern": "^(?:linux|darwin|win32)-[a-z0-9_]+$" },
40
- "source_type": { "enum": ["npm", "git"] }
40
+ "node_version": { "type": "string", "pattern": "^v[0-9]+\\.[0-9]+\\.[0-9]+$" },
41
+ "source_type": { "enum": ["npm", "git"] },
42
+ "storage": { "enum": ["host-artifact-store-v1", "harbor-artifact-cache-v2"] }
41
43
  },
42
44
  "additionalProperties": false
43
45
  },
44
- "trials": {
46
+ "prepared_artifacts": {
45
47
  "type": "array",
48
+ "minItems": 1,
46
49
  "items": {
47
50
  "type": "object",
48
- "additionalProperties": false,
49
- "required": ["trial_id", "run_id", "task_id", "attempt", "observation_status"],
51
+ "required": ["artifact_id", "artifact_integrity", "entrypoint_integrity", "harness_id", "revision_identity", "platform", "node_version", "source_type", "task_ids", "runtime_contract"],
50
52
  "properties": {
51
- "trial_id": { "type": "string", "minLength": 1 },
52
- "run_id": { "type": "string", "pattern": "^run_[a-f0-9]{32}$" },
53
- "task_id": { "type": "string", "minLength": 1 },
54
- "attempt": { "type": "integer", "minimum": 1 },
55
- "observation_status": { "enum": ["valid", "invalid"] },
56
- "reward": { "type": "number" },
57
- "verifier_result_ref": { "type": "string", "minLength": 1 },
58
- "invalid_reason": { "type": "string", "minLength": 1 }
59
- }
53
+ "artifact_id": { "type": "string", "pattern": "^sha256:[0-9a-f]{64}$" },
54
+ "artifact_integrity": { "type": "string", "pattern": "^sha256:[0-9a-f]{64}$" },
55
+ "entrypoint_integrity": { "type": "string", "pattern": "^sha256:[0-9a-f]{64}$" },
56
+ "harness_id": { "type": "string", "pattern": "^[a-z][a-z0-9-]*$" },
57
+ "revision_identity": { "type": "string", "pattern": "^sha256:[0-9a-f]{64}$" },
58
+ "platform": { "type": "string", "pattern": "^(?:linux|darwin|win32)-[a-z0-9_]+$" },
59
+ "node_version": { "type": "string", "pattern": "^v[0-9]+\\.[0-9]+\\.[0-9]+$" },
60
+ "source_type": { "enum": ["npm", "git"] },
61
+ "storage": { "enum": ["host-artifact-store-v1", "harbor-artifact-cache-v2"] },
62
+ "task_ids": { "type": "array", "items": { "type": "string", "minLength": 1 }, "uniqueItems": true },
63
+ "runtime_contract": {
64
+ "type": "object",
65
+ "required": ["docker_platform", "artifact_platform", "node_version"],
66
+ "properties": {
67
+ "docker_platform": { "enum": ["linux/amd64", "linux/arm64"] },
68
+ "artifact_platform": { "enum": ["linux-x64", "linux-arm64"] },
69
+ "node_version": { "type": "string", "pattern": "^v[0-9]+\\.[0-9]+\\.[0-9]+$" }
70
+ },
71
+ "additionalProperties": false
72
+ }
73
+ },
74
+ "additionalProperties": false
60
75
  }
61
76
  },
77
+ "trials": {
78
+ "type": "array",
79
+ "items": { "$ref": "eval-trial-reference.schema.json" }
80
+ },
62
81
  "local_source_transport": {
63
82
  "type": "object",
64
83
  "required": ["kind", "resolution_identity", "commit", "tree", "payload_sha256", "payload_bytes"],
@@ -0,0 +1,14 @@
1
+ {
2
+ "$schema": "https://json-schema.org/draft/2020-12/schema",
3
+ "$id": "https://agent-hitch.local/schemas/eval-submission-input.schema.json",
4
+ "title": "Hitch control-plane eval submission input V1",
5
+ "type": "object",
6
+ "additionalProperties": false,
7
+ "required": ["request"],
8
+ "properties": {
9
+ "schema_version": { "const": "1" },
10
+ "request": { "$ref": "eval-request.schema.json" },
11
+ "execution": { "$ref": "eval-submission.schema.json#/$defs/execution" },
12
+ "idempotency_key": { "type": "string", "minLength": 1, "maxLength": 256, "pattern": "^[!-~]+$" }
13
+ }
14
+ }
@@ -0,0 +1,68 @@
1
+ {
2
+ "$schema": "https://json-schema.org/draft/2020-12/schema",
3
+ "$id": "https://agent-hitch.local/schemas/eval-submission.schema.json",
4
+ "title": "Hitch control-plane eval submission V1",
5
+ "type": "object",
6
+ "additionalProperties": false,
7
+ "required": ["schema_version", "eval_id", "request", "submission_digest", "submitted_at"],
8
+ "properties": {
9
+ "schema_version": { "const": "1" },
10
+ "eval_id": { "type": "string", "pattern": "^eval_[a-f0-9]{32}$" },
11
+ "request": { "$ref": "eval-request.schema.json" },
12
+ "execution": { "$ref": "#/$defs/execution" },
13
+ "submission_digest": { "type": "string", "pattern": "^sha256:[0-9a-f]{64}$" },
14
+ "idempotency_key_hash": { "type": "string", "pattern": "^sha256:[0-9a-f]{64}$" },
15
+ "submitted_at": { "type": "string", "format": "date-time" }
16
+ },
17
+ "$defs": {
18
+ "resources": {
19
+ "type": "object",
20
+ "additionalProperties": false,
21
+ "required": ["cpu_millis", "memory_bytes", "container_slots", "build_slots"],
22
+ "properties": {
23
+ "cpu_millis": { "type": "integer", "minimum": 0 },
24
+ "memory_bytes": { "type": "integer", "minimum": 0 },
25
+ "container_slots": { "type": "integer", "minimum": 0 },
26
+ "build_slots": { "type": "integer", "minimum": 0 },
27
+ "gpu_count": { "type": "integer", "minimum": 0 },
28
+ "ephemeral_disk_bytes": { "type": "integer", "minimum": 0 }
29
+ }
30
+ },
31
+ "execution": {
32
+ "type": "object",
33
+ "additionalProperties": false,
34
+ "required": ["provider", "max_parallelism", "resources", "build", "model_capture"],
35
+ "properties": {
36
+ "provider": { "type": "string", "minLength": 1 },
37
+ "max_parallelism": { "type": "integer", "minimum": 1 },
38
+ "resources": {
39
+ "type": "object",
40
+ "additionalProperties": false,
41
+ "required": ["default_trial"],
42
+ "properties": {
43
+ "default_trial": { "$ref": "#/$defs/resources" },
44
+ "setup": { "$ref": "#/$defs/resources" }
45
+ }
46
+ },
47
+ "build": {
48
+ "type": "object",
49
+ "additionalProperties": false,
50
+ "required": ["mode"],
51
+ "properties": {
52
+ "mode": { "enum": ["backend", "prebuild-preferred", "prebuild-required"] },
53
+ "remote_cache": { "type": "string", "minLength": 1 }
54
+ }
55
+ },
56
+ "model_capture": {
57
+ "type": "object",
58
+ "additionalProperties": false,
59
+ "required": ["mode", "required"],
60
+ "properties": {
61
+ "mode": { "enum": ["off", "native", "proxy", "hybrid"] },
62
+ "required": { "type": "boolean" }
63
+ }
64
+ }
65
+ }
66
+ }
67
+ }
68
+ }
@@ -0,0 +1,15 @@
1
+ {
2
+ "$schema": "https://json-schema.org/draft/2020-12/schema",
3
+ "$id": "https://agent-hitch.local/schemas/eval-trial-publication.schema.json",
4
+ "title": "Hitch Eval Trial Publication v1",
5
+ "type": "object",
6
+ "additionalProperties": false,
7
+ "required": ["schema_version", "eval_id", "mode", "trial", "created_at"],
8
+ "properties": {
9
+ "schema_version": { "const": "1" },
10
+ "eval_id": { "type": "string", "pattern": "^eval_[a-f0-9]{32}$" },
11
+ "mode": { "enum": ["settle", "replace-invalid"] },
12
+ "trial": { "$ref": "eval-trial-reference.schema.json" },
13
+ "created_at": { "type": "string", "format": "date-time" }
14
+ }
15
+ }