agent-hitch 0.2.7 → 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 (287) hide show
  1. package/README.md +16 -2
  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-eval-scheduler-throughput.js +47 -0
  8. package/dist/scripts/canary-eval-scheduler-throughput.js.map +1 -0
  9. package/dist/scripts/check-architecture.js +3 -2
  10. package/dist/scripts/check-architecture.js.map +1 -1
  11. package/dist/src/adapters/catalog.js +2 -1
  12. package/dist/src/adapters/catalog.js.map +1 -1
  13. package/dist/src/adapters/providers/codex-auth.js +29 -0
  14. package/dist/src/adapters/providers/codex-auth.js.map +1 -0
  15. package/dist/src/adapters/providers/codex.js +23 -6
  16. package/dist/src/adapters/providers/codex.js.map +1 -1
  17. package/dist/src/adapters/providers/model-call.js +26 -0
  18. package/dist/src/adapters/providers/model-call.js.map +1 -0
  19. package/dist/src/artifacts/index.js +1 -1
  20. package/dist/src/artifacts/index.js.map +1 -1
  21. package/dist/src/artifacts/preparer.js +7 -2
  22. package/dist/src/artifacts/preparer.js.map +1 -1
  23. package/dist/src/backends/harbor/agent-budget.js +31 -0
  24. package/dist/src/backends/harbor/agent-budget.js.map +1 -0
  25. package/dist/src/backends/harbor/backend.js +6 -4
  26. package/dist/src/backends/harbor/backend.js.map +1 -1
  27. package/dist/src/backends/harbor/index.js +1 -0
  28. package/dist/src/backends/harbor/index.js.map +1 -1
  29. package/dist/src/backends/harbor/regrade.js +77 -0
  30. package/dist/src/backends/harbor/regrade.js.map +1 -0
  31. package/dist/src/backends/index.js +1 -0
  32. package/dist/src/backends/index.js.map +1 -1
  33. package/dist/src/benchmarks/index.js +3 -0
  34. package/dist/src/benchmarks/index.js.map +1 -0
  35. package/dist/src/benchmarks/loader.js +271 -0
  36. package/dist/src/benchmarks/loader.js.map +1 -0
  37. package/dist/src/benchmarks/metrics.js +16 -0
  38. package/dist/src/benchmarks/metrics.js.map +1 -0
  39. package/dist/src/benchmarks/toml.js +13 -0
  40. package/dist/src/benchmarks/toml.js.map +1 -0
  41. package/dist/src/benchmarks/validation.js +242 -0
  42. package/dist/src/benchmarks/validation.js.map +1 -0
  43. package/dist/src/cli/arguments.js +6 -4
  44. package/dist/src/cli/arguments.js.map +1 -1
  45. package/dist/src/cli/commands/benchmark.js +22 -0
  46. package/dist/src/cli/commands/benchmark.js.map +1 -0
  47. package/dist/src/cli/commands/capabilities.js +18 -0
  48. package/dist/src/cli/commands/capabilities.js.map +1 -0
  49. package/dist/src/cli/commands/eval.js +15 -3
  50. package/dist/src/cli/commands/eval.js.map +1 -1
  51. package/dist/src/cli/commands/run.js +37 -15
  52. package/dist/src/cli/commands/run.js.map +1 -1
  53. package/dist/src/cli/commands/trajectory.js +96 -5
  54. package/dist/src/cli/commands/trajectory.js.map +1 -1
  55. package/dist/src/cli/commands/verifier.js +36 -0
  56. package/dist/src/cli/commands/verifier.js.map +1 -0
  57. package/dist/src/cli/main.js +6 -0
  58. package/dist/src/cli/main.js.map +1 -1
  59. package/dist/src/cli/output.js +14 -2
  60. package/dist/src/cli/output.js.map +1 -1
  61. package/dist/src/control-plane/eval-control-work.js +5 -0
  62. package/dist/src/control-plane/eval-control-work.js.map +1 -1
  63. package/dist/src/control-plane/eval-scheduler.js +4 -1
  64. package/dist/src/control-plane/eval-scheduler.js.map +1 -1
  65. package/dist/src/control-plane/index.js +1 -1
  66. package/dist/src/control-plane/index.js.map +1 -1
  67. package/dist/src/control-plane/remote-result-transport.js +3 -0
  68. package/dist/src/control-plane/remote-result-transport.js.map +1 -1
  69. package/dist/src/control-plane/remote-work-coordinator.js +1 -0
  70. package/dist/src/control-plane/remote-work-coordinator.js.map +1 -1
  71. package/dist/src/control-plane/remote-work-item.js +62 -0
  72. package/dist/src/control-plane/remote-work-item.js.map +1 -0
  73. package/dist/src/control-plane/remote-work-recovery.js +44 -5
  74. package/dist/src/control-plane/remote-work-recovery.js.map +1 -1
  75. package/dist/src/control-plane/remote-worker-protocol.js +2 -23
  76. package/dist/src/control-plane/remote-worker-protocol.js.map +1 -1
  77. package/dist/src/control-plane/rerun-scheduler.js +51 -14
  78. package/dist/src/control-plane/rerun-scheduler.js.map +1 -1
  79. package/dist/src/control-plane/rerun-submission.js +10 -4
  80. package/dist/src/control-plane/rerun-submission.js.map +1 -1
  81. package/dist/src/control-plane/work-admission.js +2 -0
  82. package/dist/src/control-plane/work-admission.js.map +1 -1
  83. package/dist/src/control-plane/work-dispatcher.js +9 -1
  84. package/dist/src/control-plane/work-dispatcher.js.map +1 -1
  85. package/dist/src/controller-runtime/hash.js +7 -2
  86. package/dist/src/controller-runtime/hash.js.map +1 -1
  87. package/dist/src/daemon/auth.js +5 -0
  88. package/dist/src/daemon/auth.js.map +1 -1
  89. package/dist/src/domain/benchmarks.js +2 -0
  90. package/dist/src/domain/benchmarks.js.map +1 -0
  91. package/dist/src/domain/index.js +4 -0
  92. package/dist/src/domain/index.js.map +1 -1
  93. package/dist/src/domain/runs.js.map +1 -1
  94. package/dist/src/domain/validation.js +11 -4
  95. package/dist/src/domain/validation.js.map +1 -1
  96. package/dist/src/domain/verifier-evidence-validation.js +214 -0
  97. package/dist/src/domain/verifier-evidence-validation.js.map +1 -0
  98. package/dist/src/domain/verifier-evidence.js +3 -0
  99. package/dist/src/domain/verifier-evidence.js.map +1 -0
  100. package/dist/src/domain/verifier-score-contract.js +232 -0
  101. package/dist/src/domain/verifier-score-contract.js.map +1 -0
  102. package/dist/src/evals/benchmark-adapter-manifest.js +219 -0
  103. package/dist/src/evals/benchmark-adapter-manifest.js.map +1 -0
  104. package/dist/src/evals/benchmark-candidate.js +36 -0
  105. package/dist/src/evals/benchmark-candidate.js.map +1 -0
  106. package/dist/src/evals/benchmark-run.js +166 -0
  107. package/dist/src/evals/benchmark-run.js.map +1 -0
  108. package/dist/src/evals/collect-only-rerun.js +2 -1
  109. package/dist/src/evals/collect-only-rerun.js.map +1 -1
  110. package/dist/src/evals/directory.js +1 -1
  111. package/dist/src/evals/directory.js.map +1 -1
  112. package/dist/src/evals/duration-estimator.js +126 -0
  113. package/dist/src/evals/duration-estimator.js.map +1 -0
  114. package/dist/src/evals/eval-lifecycle-events.js +13 -0
  115. package/dist/src/evals/eval-lifecycle-events.js.map +1 -1
  116. package/dist/src/evals/eval-logical-plan.js +51 -0
  117. package/dist/src/evals/eval-logical-plan.js.map +1 -0
  118. package/dist/src/evals/eval-result-builder.js +81 -0
  119. package/dist/src/evals/eval-result-builder.js.map +1 -0
  120. package/dist/src/evals/evolution-baseline.js +191 -0
  121. package/dist/src/evals/evolution-baseline.js.map +1 -0
  122. package/dist/src/evals/execution-plan.js +39 -3
  123. package/dist/src/evals/execution-plan.js.map +1 -1
  124. package/dist/src/evals/failure-classifier.js +61 -0
  125. package/dist/src/evals/failure-classifier.js.map +1 -0
  126. package/dist/src/evals/harbor-artifact-builder.js +6 -2
  127. package/dist/src/evals/harbor-artifact-builder.js.map +1 -1
  128. package/dist/src/evals/harbor-bridge-error.js +41 -1
  129. package/dist/src/evals/harbor-bridge-error.js.map +1 -1
  130. package/dist/src/evals/harbor-node-runtime.js +146 -0
  131. package/dist/src/evals/harbor-node-runtime.js.map +1 -0
  132. package/dist/src/evals/index.js +15 -0
  133. package/dist/src/evals/index.js.map +1 -1
  134. package/dist/src/evals/infrastructure-retry.js +109 -22
  135. package/dist/src/evals/infrastructure-retry.js.map +1 -1
  136. package/dist/src/evals/native-phase-evidence.js +440 -0
  137. package/dist/src/evals/native-phase-evidence.js.map +1 -0
  138. package/dist/src/evals/physical-retry-work.js +31 -0
  139. package/dist/src/evals/physical-retry-work.js.map +1 -0
  140. package/dist/src/evals/planned-execution-support.js +22 -10
  141. package/dist/src/evals/planned-execution-support.js.map +1 -1
  142. package/dist/src/evals/planned-execution.js +120 -112
  143. package/dist/src/evals/planned-execution.js.map +1 -1
  144. package/dist/src/evals/planned-progress-publisher.js +62 -0
  145. package/dist/src/evals/planned-progress-publisher.js.map +1 -0
  146. package/dist/src/evals/planned-retry-execution.js +174 -0
  147. package/dist/src/evals/planned-retry-execution.js.map +1 -0
  148. package/dist/src/evals/planned-retry-lifecycle.js +4 -2
  149. package/dist/src/evals/planned-retry-lifecycle.js.map +1 -1
  150. package/dist/src/evals/preparation-rerun.js +225 -0
  151. package/dist/src/evals/preparation-rerun.js.map +1 -0
  152. package/dist/src/evals/progress.js +48 -8
  153. package/dist/src/evals/progress.js.map +1 -1
  154. package/dist/src/evals/recovery.js +54 -4
  155. package/dist/src/evals/recovery.js.map +1 -1
  156. package/dist/src/evals/regrade-evidence.js +77 -0
  157. package/dist/src/evals/regrade-evidence.js.map +1 -0
  158. package/dist/src/evals/remote-infrastructure-retry.js +84 -18
  159. package/dist/src/evals/remote-infrastructure-retry.js.map +1 -1
  160. package/dist/src/evals/request.js +21 -3
  161. package/dist/src/evals/request.js.map +1 -1
  162. package/dist/src/evals/rerun-inputs.js +19 -0
  163. package/dist/src/evals/rerun-inputs.js.map +1 -1
  164. package/dist/src/evals/rerun-types.js +2 -6
  165. package/dist/src/evals/rerun-types.js.map +1 -1
  166. package/dist/src/evals/rerun.js +27 -20
  167. package/dist/src/evals/rerun.js.map +1 -1
  168. package/dist/src/evals/result-helpers.js +13 -0
  169. package/dist/src/evals/result-helpers.js.map +1 -1
  170. package/dist/src/evals/retry-backoff.js +20 -0
  171. package/dist/src/evals/retry-backoff.js.map +1 -0
  172. package/dist/src/evals/retry-state.js +195 -0
  173. package/dist/src/evals/retry-state.js.map +1 -0
  174. package/dist/src/evals/scheduler-metrics.js +95 -0
  175. package/dist/src/evals/scheduler-metrics.js.map +1 -0
  176. package/dist/src/evals/scheduler-trace-replay.js +107 -0
  177. package/dist/src/evals/scheduler-trace-replay.js.map +1 -0
  178. package/dist/src/evals/service.js +87 -138
  179. package/dist/src/evals/service.js.map +1 -1
  180. package/dist/src/evals/trial-import.js +47 -50
  181. package/dist/src/evals/trial-import.js.map +1 -1
  182. package/dist/src/evals/trial-publication-recovery.js +45 -5
  183. package/dist/src/evals/trial-publication-recovery.js.map +1 -1
  184. package/dist/src/evals/trial-reference-validation.js +49 -0
  185. package/dist/src/evals/trial-reference-validation.js.map +1 -0
  186. package/dist/src/evals/verifier-artifacts.js +317 -0
  187. package/dist/src/evals/verifier-artifacts.js.map +1 -0
  188. package/dist/src/evals/verifier-eligibility.js +25 -0
  189. package/dist/src/evals/verifier-eligibility.js.map +1 -0
  190. package/dist/src/evals/verifier-only-rerun.js +210 -0
  191. package/dist/src/evals/verifier-only-rerun.js.map +1 -0
  192. package/dist/src/evals/verifier-runtime.js +61 -0
  193. package/dist/src/evals/verifier-runtime.js.map +1 -0
  194. package/dist/src/evals/verifier-score-artifacts.js +119 -0
  195. package/dist/src/evals/verifier-score-artifacts.js.map +1 -0
  196. package/dist/src/foundation/contained-file.js +85 -0
  197. package/dist/src/foundation/contained-file.js.map +1 -0
  198. package/dist/src/foundation/credential-redaction.js +26 -1
  199. package/dist/src/foundation/credential-redaction.js.map +1 -1
  200. package/dist/src/foundation/fs.js +2 -1
  201. package/dist/src/foundation/fs.js.map +1 -1
  202. package/dist/src/foundation/index.js +2 -1
  203. package/dist/src/foundation/index.js.map +1 -1
  204. package/dist/src/runs/adapter-process.js +12 -0
  205. package/dist/src/runs/adapter-process.js.map +1 -0
  206. package/dist/src/runs/executor.js +24 -30
  207. package/dist/src/runs/executor.js.map +1 -1
  208. package/dist/src/runs/finalizer.js +17 -0
  209. package/dist/src/runs/finalizer.js.map +1 -1
  210. package/dist/src/runs/index.js +5 -0
  211. package/dist/src/runs/index.js.map +1 -1
  212. package/dist/src/runs/phase-bundle.js +86 -0
  213. package/dist/src/runs/phase-bundle.js.map +1 -0
  214. package/dist/src/runs/phase-cancellation.js +107 -0
  215. package/dist/src/runs/phase-cancellation.js.map +1 -0
  216. package/dist/src/runs/phase-group.js +113 -0
  217. package/dist/src/runs/phase-group.js.map +1 -0
  218. package/dist/src/runs/query.js +1 -1
  219. package/dist/src/runs/query.js.map +1 -1
  220. package/dist/src/runs/records.js +3 -0
  221. package/dist/src/runs/records.js.map +1 -1
  222. package/dist/src/runs/request.js +4 -2
  223. package/dist/src/runs/request.js.map +1 -1
  224. package/dist/src/runs/verifier-evidence-redaction.js +61 -0
  225. package/dist/src/runs/verifier-evidence-redaction.js.map +1 -0
  226. package/dist/src/runs/verifier-evidence.js +459 -0
  227. package/dist/src/runs/verifier-evidence.js.map +1 -0
  228. package/dist/src/runs/verifier-structured-evidence.js +83 -0
  229. package/dist/src/runs/verifier-structured-evidence.js.map +1 -0
  230. package/dist/src/trajectories/analysis.js +288 -0
  231. package/dist/src/trajectories/analysis.js.map +1 -0
  232. package/dist/src/trajectories/chunk-projection.js +148 -0
  233. package/dist/src/trajectories/chunk-projection.js.map +1 -0
  234. package/dist/src/trajectories/content-projection.js +389 -0
  235. package/dist/src/trajectories/content-projection.js.map +1 -0
  236. package/dist/src/trajectories/dsh-chunk-contract.js +176 -0
  237. package/dist/src/trajectories/dsh-chunk-contract.js.map +1 -0
  238. package/dist/src/trajectories/dsh-contract.js +284 -0
  239. package/dist/src/trajectories/dsh-contract.js.map +1 -0
  240. package/dist/src/trajectories/events-chunk-drill.js +95 -0
  241. package/dist/src/trajectories/events-chunk-drill.js.map +1 -0
  242. package/dist/src/trajectories/events-page.js +367 -0
  243. package/dist/src/trajectories/events-page.js.map +1 -0
  244. package/dist/src/trajectories/index.js +3 -0
  245. package/dist/src/trajectories/index.js.map +1 -1
  246. package/dist/src/trajectories/request-attempt.js +47 -0
  247. package/dist/src/trajectories/request-attempt.js.map +1 -0
  248. package/dist/src/trajectories/stream-reader.js +311 -0
  249. package/dist/src/trajectories/stream-reader.js.map +1 -0
  250. package/dist/src/trajectories/surface-fold.js +126 -0
  251. package/dist/src/trajectories/surface-fold.js.map +1 -0
  252. package/dist/src/workers/remote-harbor-worker.js +3 -0
  253. package/dist/src/workers/remote-harbor-worker.js.map +1 -1
  254. package/docs/schemas/benchmark-hook-request-v1.schema.json +65 -0
  255. package/docs/schemas/benchmark-hook-response-v1.schema.json +75 -0
  256. package/docs/schemas/benchmark-hook-v1.schema.json +33 -0
  257. package/docs/schemas/benchmark-lock-v1.schema.json +351 -0
  258. package/docs/schemas/benchmark-metric-v1.schema.json +37 -0
  259. package/docs/schemas/benchmark-package-v1.schema.json +167 -0
  260. package/docs/schemas/benchmark-phase-group.schema.json +41 -0
  261. package/docs/schemas/benchmark-profile-v1.schema.json +141 -0
  262. package/docs/schemas/benchmark-task-v1.schema.json +384 -0
  263. package/docs/schemas/benchmark-tool-result-v1.schema.json +30 -0
  264. package/docs/schemas/error.schema.json +22 -0
  265. package/docs/schemas/eval-progress.schema.json +1 -15
  266. package/docs/schemas/eval-rerun-result.schema.json +6 -1
  267. package/docs/schemas/eval-rerun-submission.schema.json +3 -1
  268. package/docs/schemas/eval-result.schema.json +1 -15
  269. package/docs/schemas/eval-trial-publication.schema.json +1 -25
  270. package/docs/schemas/eval-trial-reference.schema.json +129 -0
  271. package/docs/schemas/regrade-assessment-reference.schema.json +11 -0
  272. package/docs/schemas/run-context.schema.json +15 -0
  273. package/docs/schemas/run-manifest.schema.json +4 -0
  274. package/docs/schemas/run-request.schema.json +4 -0
  275. package/docs/schemas/trajectory-analysis.schema.json +258 -0
  276. package/docs/schemas/trajectory-events-page.schema.json +40 -0
  277. package/docs/schemas/verifier-assessment.schema.json +77 -0
  278. package/docs/schemas/verifier-evidence.schema.json +276 -0
  279. package/integrations/harbor/hitch_benchmark.py +252 -0
  280. package/integrations/harbor/hitch_candidate_recycle.py +262 -0
  281. package/integrations/harbor/hitch_harbor_agent.py +541 -43
  282. package/integrations/harbor/hitch_harbor_environment.py +138 -0
  283. package/integrations/harbor/hitch_harbor_verifier.py +60 -2
  284. package/integrations/harbor/hitch_phase_supervisor.py +452 -0
  285. package/integrations/harbor/hitch_tool_client.mjs +91 -0
  286. package/integrations/model-call/cli.js +57 -0
  287. package/package.json +17 -6
@@ -9,31 +9,7 @@
9
9
  "schema_version": { "const": "1" },
10
10
  "eval_id": { "type": "string", "pattern": "^eval_[a-f0-9]{32}$" },
11
11
  "mode": { "enum": ["settle", "replace-invalid"] },
12
- "trial": {
13
- "type": "object",
14
- "additionalProperties": false,
15
- "required": ["trial_id", "run_id", "task_id", "attempt", "observation_status"],
16
- "properties": {
17
- "trial_id": { "type": "string", "minLength": 1 },
18
- "run_id": { "type": "string", "pattern": "^run_[a-f0-9]{32}$" },
19
- "task_id": { "type": "string", "minLength": 1 },
20
- "attempt": { "type": "integer", "minimum": 1 },
21
- "observation_status": { "enum": ["valid", "invalid"] },
22
- "reward": { "type": "number" },
23
- "verifier_result_ref": { "type": "string", "minLength": 1 },
24
- "invalid_reason": { "type": "string", "minLength": 1 }
25
- },
26
- "allOf": [
27
- {
28
- "if": { "properties": { "observation_status": { "const": "valid" } } },
29
- "then": { "required": ["reward"] }
30
- },
31
- {
32
- "if": { "properties": { "observation_status": { "const": "invalid" } } },
33
- "then": { "required": ["invalid_reason"] }
34
- }
35
- ]
36
- },
12
+ "trial": { "$ref": "eval-trial-reference.schema.json" },
37
13
  "created_at": { "type": "string", "format": "date-time" }
38
14
  }
39
15
  }
@@ -0,0 +1,129 @@
1
+ {
2
+ "$schema": "https://json-schema.org/draft/2020-12/schema",
3
+ "$id": "https://agent-hitch.local/schemas/eval-trial-reference.schema.json",
4
+ "type": "object",
5
+ "additionalProperties": false,
6
+ "required": [
7
+ "trial_id",
8
+ "task_id",
9
+ "attempt",
10
+ "observation_status"
11
+ ],
12
+ "properties": {
13
+ "trial_id": {
14
+ "type": "string",
15
+ "minLength": 1
16
+ },
17
+ "run_id": {
18
+ "type": "string",
19
+ "pattern": "^run_[a-f0-9]{32}$"
20
+ },
21
+ "task_id": {
22
+ "type": "string",
23
+ "minLength": 1
24
+ },
25
+ "attempt": {
26
+ "type": "integer",
27
+ "minimum": 1
28
+ },
29
+ "observation_status": {
30
+ "enum": [
31
+ "valid",
32
+ "invalid"
33
+ ]
34
+ },
35
+ "reward": {
36
+ "type": "number"
37
+ },
38
+ "scores": {
39
+ "$ref": "verifier-evidence.schema.json#/$defs/scores"
40
+ },
41
+ "verifier_result_ref": {
42
+ "type": "string",
43
+ "minLength": 1
44
+ },
45
+ "invalid_reason": {
46
+ "type": "string",
47
+ "minLength": 1
48
+ },
49
+ "assessment": {
50
+ "$ref": "regrade-assessment-reference.schema.json"
51
+ },
52
+ "run_group": {
53
+ "type": "object",
54
+ "required": [
55
+ "run_group_id",
56
+ "digest"
57
+ ],
58
+ "properties": {
59
+ "run_group_id": {
60
+ "type": "string",
61
+ "pattern": "^run_group_[a-f0-9]{32}$"
62
+ },
63
+ "digest": {
64
+ "type": "string",
65
+ "pattern": "^sha256:[a-f0-9]{64}$"
66
+ }
67
+ },
68
+ "additionalProperties": false
69
+ }
70
+ },
71
+ "oneOf": [
72
+ {
73
+ "required": [
74
+ "run_id"
75
+ ],
76
+ "not": {
77
+ "required": [
78
+ "run_group"
79
+ ]
80
+ }
81
+ },
82
+ {
83
+ "required": [
84
+ "run_group",
85
+ "assessment"
86
+ ],
87
+ "not": {
88
+ "required": [
89
+ "run_id"
90
+ ]
91
+ }
92
+ }
93
+ ],
94
+ "allOf": [
95
+ {
96
+ "if": {
97
+ "properties": {
98
+ "observation_status": {
99
+ "const": "valid"
100
+ }
101
+ }
102
+ },
103
+ "then": {
104
+ "required": [
105
+ "reward"
106
+ ]
107
+ }
108
+ },
109
+ {
110
+ "if": {
111
+ "properties": {
112
+ "observation_status": {
113
+ "const": "invalid"
114
+ }
115
+ }
116
+ },
117
+ "then": {
118
+ "required": [
119
+ "invalid_reason"
120
+ ],
121
+ "not": {
122
+ "required": [
123
+ "scores"
124
+ ]
125
+ }
126
+ }
127
+ }
128
+ ]
129
+ }
@@ -0,0 +1,11 @@
1
+ {
2
+ "$schema": "https://json-schema.org/draft/2020-12/schema",
3
+ "title": "Hitch immutable verifier-only assessment reference",
4
+ "type": "object",
5
+ "additionalProperties": false,
6
+ "required": ["id", "digest"],
7
+ "properties": {
8
+ "id": { "type": "string", "pattern": "^assessment_[a-f0-9]{32}$" },
9
+ "digest": { "type": "string", "pattern": "^sha256:[a-f0-9]{64}$" }
10
+ }
11
+ }
@@ -34,6 +34,21 @@
34
34
  "task_digest": { "type": "string", "pattern": "^sha256:[0-9a-f]{64}$" },
35
35
  "verifier_identity": { "type": "string", "pattern": "^sha256:[0-9a-f]{64}$" }
36
36
  }
37
+ },
38
+ {
39
+ "type": "object",
40
+ "additionalProperties": false,
41
+ "required": ["kind", "benchmark_id", "benchmark_revision", "task_id", "task_digest", "verifier_identity", "run_group_id", "phase_index"],
42
+ "properties": {
43
+ "kind": { "const": "benchmark_phase" },
44
+ "benchmark_id": { "type": "string", "minLength": 1 },
45
+ "benchmark_revision": { "type": "string", "minLength": 1, "not": { "pattern": "^[Ll][Aa][Tt][Ee][Ss][Tt]$" } },
46
+ "task_id": { "type": "string", "minLength": 1 },
47
+ "task_digest": { "type": "string", "pattern": "^sha256:[0-9a-f]{64}$" },
48
+ "verifier_identity": { "type": "string", "pattern": "^sha256:[0-9a-f]{64}$" },
49
+ "run_group_id": { "type": "string", "pattern": "^run_group_[a-f0-9]{32}$" },
50
+ "phase_index": { "type": "integer", "minimum": 1, "maximum": 9007199254740991 }
51
+ }
37
52
  }
38
53
  ]
39
54
  }
@@ -5,6 +5,10 @@
5
5
  "type": "object",
6
6
  "additionalProperties": true,
7
7
  "required": ["schema_version", "run_id", "context", "status", "harness", "model", "protocol", "request_ref", "resolution_ref", "created_at"],
8
+ "allOf": [{
9
+ "if": { "properties": { "context": { "properties": { "kind": { "const": "benchmark_phase" } }, "required": ["kind"] } }, "required": ["context"] },
10
+ "then": { "required": ["parent"], "not": { "required": ["observation"] } }
11
+ }],
8
12
  "properties": {
9
13
  "schema_version": { "const": "1" },
10
14
  "run_id": { "type": "string", "pattern": "^run_[a-f0-9]{32}$" },
@@ -5,6 +5,10 @@
5
5
  "type": "object",
6
6
  "additionalProperties": false,
7
7
  "required": ["prompt"],
8
+ "allOf": [{
9
+ "if": { "properties": { "context": { "properties": { "kind": { "const": "benchmark_phase" } }, "required": ["kind"] } }, "required": ["context"] },
10
+ "then": { "required": ["parent"], "properties": { "defer_benchmark_observation": { "const": false } } }
11
+ }],
8
12
  "oneOf": [
9
13
  {
10
14
  "required": ["harness_ref"],
@@ -0,0 +1,258 @@
1
+ {
2
+ "$schema": "https://json-schema.org/draft/2020-12/schema",
3
+ "$id": "https://agent-hitch.local/schemas/trajectory-analysis.schema.json",
4
+ "title": "Hitch bounded trajectory analysis",
5
+ "type": "object",
6
+ "additionalProperties": false,
7
+ "required": ["schema_version", "kind", "run_id", "source", "header", "surface", "events", "chunk_summaries", "omitted_event_types", "coverage"],
8
+ "properties": {
9
+ "schema_version": { "const": "1" },
10
+ "kind": { "const": "trajectory-analysis" },
11
+ "run_id": { "$ref": "#/$defs/run_id" },
12
+ "source": {
13
+ "type": "object",
14
+ "additionalProperties": false,
15
+ "required": ["fidelity", "session_id", "canonical_sha256", "canonical_bytes", "event_count", "event_types"],
16
+ "properties": {
17
+ "fidelity": { "enum": ["provider_native", "normalized", "minimal"] },
18
+ "provider": { "type": "string", "minLength": 1 },
19
+ "session_id": { "type": "string", "minLength": 1 },
20
+ "canonical_sha256": { "$ref": "#/$defs/sha256" },
21
+ "canonical_bytes": { "type": "integer", "minimum": 0 },
22
+ "event_count": { "type": "integer", "minimum": 0 },
23
+ "event_types": { "$ref": "#/$defs/counts" }
24
+ }
25
+ },
26
+ "header": true,
27
+ "surface": {
28
+ "type": "object",
29
+ "additionalProperties": false,
30
+ "required": ["fidelity", "nodes", "current_node_seqs", "replacements", "request_boundaries", "request_headers"],
31
+ "properties": {
32
+ "fidelity": { "enum": ["exact", "normalized", "partial"] },
33
+ "nodes": {
34
+ "type": "array",
35
+ "items": {
36
+ "type": "object",
37
+ "additionalProperties": false,
38
+ "required": ["seq", "event_type", "surface_op", "message"],
39
+ "properties": {
40
+ "seq": { "$ref": "#/$defs/seq" },
41
+ "event_type": { "enum": ["user/message", "assistant/message", "tool/result"] },
42
+ "surface_op": { "$ref": "#/$defs/surface_op" },
43
+ "message": true
44
+ }
45
+ }
46
+ },
47
+ "current_node_seqs": { "type": "array", "items": { "$ref": "#/$defs/seq" } },
48
+ "replacements": {
49
+ "type": "array",
50
+ "items": {
51
+ "type": "object",
52
+ "additionalProperties": false,
53
+ "required": ["seq", "start", "end", "shadowed_seqs"],
54
+ "properties": {
55
+ "seq": { "$ref": "#/$defs/seq" },
56
+ "start": { "$ref": "#/$defs/seq" },
57
+ "end": { "$ref": "#/$defs/seq" },
58
+ "shadowed_seqs": { "type": "array", "items": { "$ref": "#/$defs/seq" } }
59
+ }
60
+ }
61
+ },
62
+ "request_boundaries": {
63
+ "type": "array",
64
+ "items": {
65
+ "type": "object",
66
+ "additionalProperties": false,
67
+ "required": ["turn", "step", "attempt", "boundary_seq", "surface_revision"],
68
+ "properties": {
69
+ "turn": { "$ref": "#/$defs/seq" },
70
+ "step": { "$ref": "#/$defs/seq" },
71
+ "attempt": { "$ref": "#/$defs/seq" },
72
+ "retry_id": { "$ref": "#/$defs/projected_string" },
73
+ "boundary_seq": { "$ref": "#/$defs/seq" },
74
+ "surface_revision": { "$ref": "#/$defs/seq" },
75
+ "request_header_seq": { "$ref": "#/$defs/seq" }
76
+ }
77
+ }
78
+ },
79
+ "request_headers": {
80
+ "type": "array",
81
+ "items": {
82
+ "type": "object",
83
+ "additionalProperties": false,
84
+ "required": ["seq", "header"],
85
+ "properties": {
86
+ "seq": { "$ref": "#/$defs/seq" },
87
+ "header": true
88
+ }
89
+ }
90
+ }
91
+ }
92
+ },
93
+ "events": { "type": "array", "items": { "$ref": "#/$defs/diagnostic_event" } },
94
+ "chunk_summaries": {
95
+ "type": "array",
96
+ "items": {
97
+ "type": "object",
98
+ "additionalProperties": false,
99
+ "required": ["turn", "step", "attempt", "first_seq", "last_seq", "count", "types", "model_boundary_seq"],
100
+ "properties": {
101
+ "turn": { "$ref": "#/$defs/seq" },
102
+ "step": { "$ref": "#/$defs/seq" },
103
+ "attempt": { "$ref": "#/$defs/seq" },
104
+ "retry_id": { "$ref": "#/$defs/projected_string" },
105
+ "first_seq": { "$ref": "#/$defs/seq" },
106
+ "last_seq": { "$ref": "#/$defs/seq" },
107
+ "count": { "type": "integer", "minimum": 1 },
108
+ "types": { "$ref": "#/$defs/counts" },
109
+ "model_boundary_seq": { "$ref": "#/$defs/seq" },
110
+ "usage": true,
111
+ "finish_reason": true,
112
+ "partial": {
113
+ "type": "object",
114
+ "additionalProperties": false,
115
+ "required": ["status", "source_seq_count"],
116
+ "oneOf": [
117
+ { "required": ["content"] },
118
+ { "required": ["streams"] }
119
+ ],
120
+ "properties": {
121
+ "status": { "const": "incomplete" },
122
+ "content": { "$ref": "#/$defs/content_excerpt" },
123
+ "streams": {
124
+ "type": "array",
125
+ "minItems": 2,
126
+ "items": {
127
+ "type": "object",
128
+ "additionalProperties": false,
129
+ "required": ["block_index", "block_start_seq", "kind", "content", "source_seq_count"],
130
+ "properties": {
131
+ "block_index": { "$ref": "#/$defs/seq" },
132
+ "block_start_seq": { "$ref": "#/$defs/seq" },
133
+ "kind": { "enum": ["text", "reasoning", "tool_arguments"] },
134
+ "content": { "$ref": "#/$defs/content_excerpt" },
135
+ "source_seq_count": { "type": "integer", "minimum": 0 }
136
+ }
137
+ }
138
+ },
139
+ "source_seq_count": { "type": "integer", "minimum": 0 }
140
+ }
141
+ }
142
+ }
143
+ }
144
+ },
145
+ "omitted_event_types": { "$ref": "#/$defs/counts" },
146
+ "coverage": {
147
+ "type": "object",
148
+ "additionalProperties": false,
149
+ "required": ["surface", "chunks", "content", "child_sessions"],
150
+ "properties": {
151
+ "surface": { "enum": ["complete", "partial"] },
152
+ "chunks": { "enum": ["coalesced", "omitted", "partial"] },
153
+ "content": { "enum": ["complete", "excerpted", "partial"] },
154
+ "child_sessions": { "enum": ["complete", "partial", "none", "unavailable"] }
155
+ }
156
+ },
157
+ "redactions": { "$ref": "#/$defs/redactions" }
158
+ },
159
+ "$defs": {
160
+ "run_id": { "type": "string", "pattern": "^run_[a-f0-9]{32}$" },
161
+ "sha256": { "type": "string", "pattern": "^sha256:[0-9a-f]{64}$" },
162
+ "seq": { "type": "integer", "minimum": 0 },
163
+ "surface_op": {
164
+ "oneOf": [
165
+ { "const": "append" },
166
+ {
167
+ "type": "object",
168
+ "additionalProperties": false,
169
+ "required": ["op", "start", "end"],
170
+ "properties": {
171
+ "op": { "const": "replace" },
172
+ "start": { "$ref": "#/$defs/seq" },
173
+ "end": { "$ref": "#/$defs/seq" }
174
+ }
175
+ }
176
+ ]
177
+ },
178
+ "counts": { "type": "object", "additionalProperties": { "type": "integer", "minimum": 0 } },
179
+ "projected_string": {
180
+ "oneOf": [
181
+ { "type": "string" },
182
+ { "$ref": "#/$defs/content_excerpt" }
183
+ ]
184
+ },
185
+ "diagnostic_event": {
186
+ "oneOf": [
187
+ {
188
+ "type": "object",
189
+ "additionalProperties": false,
190
+ "required": ["type", "seq", "time", "data"],
191
+ "properties": {
192
+ "type": { "type": "string", "minLength": 1, "maxLength": 1024 },
193
+ "seq": { "$ref": "#/$defs/seq" },
194
+ "time": { "type": "number" },
195
+ "data": true,
196
+ "ignorable": { "const": true },
197
+ "surfaceOp": { "$ref": "#/$defs/surface_op" },
198
+ "source_event_seqs_summary": {
199
+ "type": "object",
200
+ "additionalProperties": false,
201
+ "required": ["count"],
202
+ "properties": {
203
+ "count": { "type": "integer", "minimum": 0 },
204
+ "first": { "$ref": "#/$defs/seq" },
205
+ "last": { "$ref": "#/$defs/seq" }
206
+ }
207
+ }
208
+ }
209
+ },
210
+ {
211
+ "type": "object",
212
+ "additionalProperties": false,
213
+ "required": ["type", "seq", "time", "event_excerpt"],
214
+ "properties": {
215
+ "type": { "type": "string", "minLength": 1, "maxLength": 1024 },
216
+ "seq": { "$ref": "#/$defs/seq" },
217
+ "time": { "type": "number" },
218
+ "event_excerpt": { "$ref": "#/$defs/content_excerpt" }
219
+ }
220
+ }
221
+ ]
222
+ },
223
+ "content_excerpt": {
224
+ "type": "object",
225
+ "additionalProperties": false,
226
+ "required": ["preview", "bytes", "sha256", "truncated", "source"],
227
+ "properties": {
228
+ "preview": { "type": "string" },
229
+ "tail": { "type": "string" },
230
+ "bytes": { "type": "integer", "minimum": 0 },
231
+ "sha256": { "$ref": "#/$defs/sha256" },
232
+ "truncated": { "type": "boolean" },
233
+ "source": {
234
+ "type": "object",
235
+ "additionalProperties": false,
236
+ "required": ["run_id", "seq", "field"],
237
+ "properties": {
238
+ "run_id": { "$ref": "#/$defs/run_id" },
239
+ "seq": { "$ref": "#/$defs/seq" },
240
+ "field": { "type": "string", "minLength": 1, "maxLength": 1024 }
241
+ }
242
+ }
243
+ }
244
+ },
245
+ "redactions": {
246
+ "type": "array",
247
+ "items": {
248
+ "type": "object",
249
+ "additionalProperties": false,
250
+ "required": ["rule_id", "count"],
251
+ "properties": {
252
+ "rule_id": { "type": "string", "minLength": 1 },
253
+ "count": { "type": "integer", "minimum": 1 }
254
+ }
255
+ }
256
+ }
257
+ }
258
+ }
@@ -0,0 +1,40 @@
1
+ {
2
+ "$schema": "https://json-schema.org/draft/2020-12/schema",
3
+ "$id": "https://agent-hitch.local/schemas/trajectory-events-page.schema.json",
4
+ "title": "Hitch source-filtered trajectory event page",
5
+ "type": "object",
6
+ "additionalProperties": false,
7
+ "required": ["schema_version", "kind", "run_id", "canonical_sha256", "filter", "events", "total_matches", "eof"],
8
+ "properties": {
9
+ "schema_version": { "const": "1" },
10
+ "kind": { "const": "trajectory-events-page" },
11
+ "run_id": { "type": "string", "pattern": "^run_[a-f0-9]{32}$" },
12
+ "canonical_sha256": { "type": "string", "pattern": "^sha256:[0-9a-f]{64}$" },
13
+ "filter": {
14
+ "type": "object",
15
+ "additionalProperties": false,
16
+ "properties": {
17
+ "types": { "type": "array", "minItems": 1, "uniqueItems": true, "items": { "type": "string", "minLength": 1, "maxLength": 1024 } },
18
+ "seq_start": { "type": "integer", "minimum": 0 },
19
+ "seq_end": { "type": "integer", "minimum": 0 },
20
+ "field": { "type": "string", "minLength": 1, "maxLength": 1024 }
21
+ }
22
+ },
23
+ "events": { "type": "array" },
24
+ "total_matches": { "type": "integer", "minimum": 0 },
25
+ "next_cursor": { "type": "string", "minLength": 1 },
26
+ "eof": { "type": "boolean" },
27
+ "redactions": {
28
+ "type": "array",
29
+ "items": {
30
+ "type": "object",
31
+ "additionalProperties": false,
32
+ "required": ["rule_id", "count"],
33
+ "properties": {
34
+ "rule_id": { "type": "string", "minLength": 1 },
35
+ "count": { "type": "integer", "minimum": 1 }
36
+ }
37
+ }
38
+ }
39
+ }
40
+ }
@@ -0,0 +1,77 @@
1
+ {
2
+ "$schema": "https://json-schema.org/draft/2020-12/schema",
3
+ "title": "Hitch artifact-only verifier assessment v1",
4
+ "type": "object",
5
+ "additionalProperties": false,
6
+ "required": ["schema_version", "kind", "candidate_executes", "eval_id", "task_id", "attempt", "rerun_id", "source", "controller_runtime_id", "backend", "observation", "evidence_digest", "completed_at"],
7
+ "properties": {
8
+ "schema_version": { "const": "1" },
9
+ "kind": { "const": "verifier-only-assessment" },
10
+ "candidate_executes": { "const": false },
11
+ "eval_id": { "type": "string", "pattern": "^eval_[a-f0-9]{32}$" },
12
+ "task_id": { "type": "string", "minLength": 1 },
13
+ "attempt": { "type": "integer", "minimum": 1 },
14
+ "rerun_id": { "type": "string", "pattern": "^rerun_[a-f0-9]{32}$" },
15
+ "controller_runtime_id": { "$ref": "#/$defs/digest" },
16
+ "runtime_repair": {
17
+ "type": "object", "additionalProperties": false,
18
+ "required": ["schema_version", "kind", "source_runtime_id", "replacement_runtime_id", "path", "source_sha256", "replacement_sha256", "unchanged_file_count"],
19
+ "properties": {
20
+ "schema_version": { "const": "1" }, "kind": { "const": "verifier-environment-runtime-repair" },
21
+ "source_runtime_id": { "$ref": "#/$defs/digest" }, "replacement_runtime_id": { "$ref": "#/$defs/digest" },
22
+ "path": { "const": "integrations/harbor/hitch_harbor_environment.py" },
23
+ "source_sha256": { "$ref": "#/$defs/digest" }, "replacement_sha256": { "$ref": "#/$defs/digest" },
24
+ "unchanged_file_count": { "type": "integer", "minimum": 0 }
25
+ }
26
+ },
27
+ "evidence_digest": { "$ref": "#/$defs/digest" },
28
+ "completed_at": { "type": "string", "format": "date-time" },
29
+ "source": {
30
+ "type": "object",
31
+ "additionalProperties": false,
32
+ "required": ["trial_id", "run_id", "work_id", "backend_directory", "bundle_index_digest", "trial_tree_digest", "artifacts_digest", "task_tree_digest", "artifact_capture_at"],
33
+ "properties": {
34
+ "trial_id": { "type": "string", "minLength": 1 },
35
+ "run_id": { "type": "string", "pattern": "^run_[a-f0-9]{32}$" },
36
+ "controller_runtime_id": { "$ref": "#/$defs/digest" },
37
+ "work_id": { "type": "string", "pattern": "^work_[a-f0-9]{32}$" },
38
+ "backend_directory": { "type": "string", "minLength": 1 },
39
+ "bundle_index_digest": { "$ref": "#/$defs/digest" },
40
+ "trial_tree_digest": { "$ref": "#/$defs/digest" },
41
+ "artifacts_digest": { "$ref": "#/$defs/digest" },
42
+ "task_tree_digest": { "$ref": "#/$defs/digest" },
43
+ "artifact_capture_at": { "type": "string", "format": "date-time" }
44
+ }
45
+ },
46
+ "backend": {
47
+ "type": "object",
48
+ "additionalProperties": false,
49
+ "required": ["name", "version", "identity", "process_exit_code", "signal"],
50
+ "properties": {
51
+ "name": { "const": "harbor" },
52
+ "version": { "const": "0.21.0" },
53
+ "identity": { "type": "string" },
54
+ "process_exit_code": { "type": ["integer", "null"] },
55
+ "signal": { "type": ["string", "null"] }
56
+ }
57
+ },
58
+ "observation": {
59
+ "type": "object",
60
+ "additionalProperties": false,
61
+ "required": ["status"],
62
+ "properties": {
63
+ "status": { "enum": ["valid", "invalid"] },
64
+ "reward": { "type": "number" },
65
+ "invalid_reason": { "type": "string", "minLength": 1 },
66
+ "verifier_result_ref": { "type": "string", "minLength": 1 }
67
+ },
68
+ "allOf": [{
69
+ "if": { "properties": { "status": { "const": "valid" } } },
70
+ "then": { "required": ["reward", "verifier_result_ref"], "not": { "required": ["invalid_reason"] } },
71
+ "else": { "required": ["invalid_reason"] }
72
+ }]
73
+ }
74
+ },
75
+ "allOf": [{ "if": { "required": ["runtime_repair"] }, "then": { "properties": { "source": { "required": ["controller_runtime_id"] } } } }],
76
+ "$defs": { "digest": { "type": "string", "pattern": "^sha256:[a-f0-9]{64}$" } }
77
+ }