@claude-flow/cli 3.10.46 → 3.11.0

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (551) hide show
  1. package/.claude/agents/analysis/analyze-code-quality.md +178 -178
  2. package/.claude/agents/analysis/code-analyzer.md +209 -209
  3. package/.claude/agents/analysis/code-review/analyze-code-quality.md +178 -178
  4. package/.claude/agents/architecture/arch-system-design.md +156 -156
  5. package/.claude/agents/architecture/system-design/arch-system-design.md +154 -154
  6. package/.claude/agents/browser/browser-agent.yaml +182 -182
  7. package/.claude/agents/consensus/byzantine-coordinator.md +62 -62
  8. package/.claude/agents/consensus/crdt-synchronizer.md +996 -996
  9. package/.claude/agents/consensus/gossip-coordinator.md +62 -62
  10. package/.claude/agents/consensus/performance-benchmarker.md +850 -850
  11. package/.claude/agents/consensus/quorum-manager.md +822 -822
  12. package/.claude/agents/consensus/raft-manager.md +62 -62
  13. package/.claude/agents/consensus/security-manager.md +621 -621
  14. package/.claude/agents/core/planner.md +374 -374
  15. package/.claude/agents/custom/test-long-runner.md +44 -44
  16. package/.claude/agents/data/data-ml-model.md +444 -444
  17. package/.claude/agents/data/ml/data-ml-model.md +192 -192
  18. package/.claude/agents/development/backend/dev-backend-api.md +141 -141
  19. package/.claude/agents/development/dev-backend-api.md +344 -344
  20. package/.claude/agents/devops/ci-cd/ops-cicd-github.md +163 -163
  21. package/.claude/agents/devops/ops-cicd-github.md +164 -164
  22. package/.claude/agents/documentation/api-docs/docs-api-openapi.md +173 -173
  23. package/.claude/agents/documentation/docs-api-openapi.md +354 -354
  24. package/.claude/agents/flow-nexus/app-store.md +87 -87
  25. package/.claude/agents/flow-nexus/authentication.md +68 -68
  26. package/.claude/agents/flow-nexus/challenges.md +80 -80
  27. package/.claude/agents/flow-nexus/neural-network.md +87 -87
  28. package/.claude/agents/flow-nexus/payments.md +82 -82
  29. package/.claude/agents/flow-nexus/sandbox.md +75 -75
  30. package/.claude/agents/flow-nexus/swarm.md +75 -75
  31. package/.claude/agents/flow-nexus/user-tools.md +95 -95
  32. package/.claude/agents/flow-nexus/workflow.md +83 -83
  33. package/.claude/agents/github/code-review-swarm.md +377 -377
  34. package/.claude/agents/github/github-modes.md +172 -172
  35. package/.claude/agents/github/issue-tracker.md +575 -575
  36. package/.claude/agents/github/multi-repo-swarm.md +552 -552
  37. package/.claude/agents/github/pr-manager.md +437 -437
  38. package/.claude/agents/github/project-board-sync.md +508 -508
  39. package/.claude/agents/github/release-manager.md +604 -604
  40. package/.claude/agents/github/release-swarm.md +582 -582
  41. package/.claude/agents/github/repo-architect.md +397 -397
  42. package/.claude/agents/github/swarm-issue.md +572 -572
  43. package/.claude/agents/github/swarm-pr.md +427 -427
  44. package/.claude/agents/github/sync-coordinator.md +451 -451
  45. package/.claude/agents/github/workflow-automation.md +902 -902
  46. package/.claude/agents/goal/agent.md +815 -815
  47. package/.claude/agents/optimization/benchmark-suite.md +664 -664
  48. package/.claude/agents/optimization/load-balancer.md +430 -430
  49. package/.claude/agents/optimization/performance-monitor.md +671 -671
  50. package/.claude/agents/optimization/resource-allocator.md +673 -673
  51. package/.claude/agents/optimization/topology-optimizer.md +807 -807
  52. package/.claude/agents/payments/agentic-payments.md +126 -126
  53. package/.claude/agents/sona/sona-learning-optimizer.md +74 -74
  54. package/.claude/agents/sparc/architecture.md +698 -698
  55. package/.claude/agents/sparc/pseudocode.md +519 -519
  56. package/.claude/agents/sparc/refinement.md +801 -801
  57. package/.claude/agents/sparc/specification.md +477 -477
  58. package/.claude/agents/specialized/mobile/spec-mobile-react-native.md +224 -224
  59. package/.claude/agents/specialized/spec-mobile-react-native.md +226 -226
  60. package/.claude/agents/sublinear/consensus-coordinator.md +337 -337
  61. package/.claude/agents/sublinear/matrix-optimizer.md +184 -184
  62. package/.claude/agents/sublinear/pagerank-analyzer.md +298 -298
  63. package/.claude/agents/sublinear/performance-optimizer.md +367 -367
  64. package/.claude/agents/sublinear/trading-predictor.md +245 -245
  65. package/.claude/agents/swarm/adaptive-coordinator.md +1126 -1126
  66. package/.claude/agents/swarm/hierarchical-coordinator.md +709 -709
  67. package/.claude/agents/swarm/mesh-coordinator.md +962 -962
  68. package/.claude/agents/templates/automation-smart-agent.md +204 -204
  69. package/.claude/agents/templates/base-template-generator.md +289 -289
  70. package/.claude/agents/templates/coordinator-swarm-init.md +89 -89
  71. package/.claude/agents/templates/github-pr-manager.md +176 -176
  72. package/.claude/agents/templates/implementer-sparc-coder.md +258 -258
  73. package/.claude/agents/templates/memory-coordinator.md +186 -186
  74. package/.claude/agents/templates/orchestrator-task.md +138 -138
  75. package/.claude/agents/templates/performance-analyzer.md +198 -198
  76. package/.claude/agents/templates/sparc-coordinator.md +513 -513
  77. package/.claude/agents/testing/production-validator.md +394 -394
  78. package/.claude/agents/testing/tdd-london-swarm.md +243 -243
  79. package/.claude/agents/v3/aidefence-guardian.md +282 -282
  80. package/.claude/agents/v3/claims-authorizer.md +208 -208
  81. package/.claude/agents/v3/collective-intelligence-coordinator.md +993 -993
  82. package/.claude/agents/v3/ddd-domain-expert.md +220 -220
  83. package/.claude/agents/v3/injection-analyst.md +236 -236
  84. package/.claude/agents/v3/performance-engineer.md +1233 -1233
  85. package/.claude/agents/v3/pii-detector.md +151 -151
  86. package/.claude/agents/v3/reasoningbank-learner.md +213 -213
  87. package/.claude/agents/v3/security-architect-aidefence.md +410 -410
  88. package/.claude/agents/v3/security-architect.md +867 -867
  89. package/.claude/agents/v3/swarm-memory-manager.md +157 -157
  90. package/.claude/agents/v3/v3-integration-architect.md +205 -205
  91. package/.claude/commands/agents/README.md +50 -50
  92. package/.claude/commands/agents/agent-capabilities.md +140 -140
  93. package/.claude/commands/agents/agent-coordination.md +28 -28
  94. package/.claude/commands/agents/agent-spawning.md +28 -28
  95. package/.claude/commands/agents/agent-types.md +216 -216
  96. package/.claude/commands/agents/health.md +139 -139
  97. package/.claude/commands/agents/list.md +100 -100
  98. package/.claude/commands/agents/logs.md +130 -130
  99. package/.claude/commands/agents/metrics.md +122 -122
  100. package/.claude/commands/agents/pool.md +127 -127
  101. package/.claude/commands/agents/spawn.md +140 -140
  102. package/.claude/commands/agents/status.md +115 -115
  103. package/.claude/commands/agents/stop.md +102 -102
  104. package/.claude/commands/analysis/COMMAND_COMPLIANCE_REPORT.md +53 -53
  105. package/.claude/commands/analysis/README.md +9 -9
  106. package/.claude/commands/analysis/bottleneck-detect.md +162 -162
  107. package/.claude/commands/analysis/performance-bottlenecks.md +58 -58
  108. package/.claude/commands/analysis/performance-report.md +25 -25
  109. package/.claude/commands/analysis/token-efficiency.md +44 -44
  110. package/.claude/commands/analysis/token-usage.md +25 -25
  111. package/.claude/commands/automation/README.md +9 -9
  112. package/.claude/commands/automation/auto-agent.md +122 -122
  113. package/.claude/commands/automation/self-healing.md +105 -105
  114. package/.claude/commands/automation/session-memory.md +89 -89
  115. package/.claude/commands/automation/smart-agents.md +72 -72
  116. package/.claude/commands/automation/smart-spawn.md +25 -25
  117. package/.claude/commands/automation/workflow-select.md +25 -25
  118. package/.claude/commands/claude-flow-help.md +103 -103
  119. package/.claude/commands/claude-flow-memory.md +107 -107
  120. package/.claude/commands/claude-flow-swarm.md +205 -205
  121. package/.claude/commands/coordination/README.md +9 -9
  122. package/.claude/commands/coordination/agent-spawn.md +25 -25
  123. package/.claude/commands/coordination/init.md +44 -44
  124. package/.claude/commands/coordination/orchestrate.md +43 -43
  125. package/.claude/commands/coordination/spawn.md +45 -45
  126. package/.claude/commands/coordination/swarm-init.md +85 -85
  127. package/.claude/commands/coordination/task-orchestrate.md +25 -25
  128. package/.claude/commands/github/README.md +11 -11
  129. package/.claude/commands/github/code-review-swarm.md +513 -513
  130. package/.claude/commands/github/code-review.md +25 -25
  131. package/.claude/commands/github/github-modes.md +146 -146
  132. package/.claude/commands/github/github-swarm.md +121 -121
  133. package/.claude/commands/github/issue-tracker.md +291 -291
  134. package/.claude/commands/github/issue-triage.md +25 -25
  135. package/.claude/commands/github/multi-repo-swarm.md +518 -518
  136. package/.claude/commands/github/pr-enhance.md +26 -26
  137. package/.claude/commands/github/pr-manager.md +169 -169
  138. package/.claude/commands/github/project-board-sync.md +470 -470
  139. package/.claude/commands/github/release-manager.md +339 -339
  140. package/.claude/commands/github/release-swarm.md +543 -543
  141. package/.claude/commands/github/repo-analyze.md +25 -25
  142. package/.claude/commands/github/repo-architect.md +366 -366
  143. package/.claude/commands/github/swarm-issue.md +484 -484
  144. package/.claude/commands/github/swarm-pr.md +287 -287
  145. package/.claude/commands/github/sync-coordinator.md +302 -302
  146. package/.claude/commands/github/workflow-automation.md +441 -441
  147. package/.claude/commands/hive-mind/README.md +17 -17
  148. package/.claude/commands/hive-mind/hive-mind-consensus.md +8 -8
  149. package/.claude/commands/hive-mind/hive-mind-init.md +18 -18
  150. package/.claude/commands/hive-mind/hive-mind-memory.md +8 -8
  151. package/.claude/commands/hive-mind/hive-mind-metrics.md +8 -8
  152. package/.claude/commands/hive-mind/hive-mind-resume.md +8 -8
  153. package/.claude/commands/hive-mind/hive-mind-sessions.md +8 -8
  154. package/.claude/commands/hive-mind/hive-mind-spawn.md +21 -21
  155. package/.claude/commands/hive-mind/hive-mind-status.md +8 -8
  156. package/.claude/commands/hive-mind/hive-mind-stop.md +8 -8
  157. package/.claude/commands/hive-mind/hive-mind-wizard.md +8 -8
  158. package/.claude/commands/hive-mind/hive-mind.md +27 -27
  159. package/.claude/commands/hooks/README.md +11 -11
  160. package/.claude/commands/hooks/overview.md +57 -57
  161. package/.claude/commands/hooks/post-edit.md +117 -117
  162. package/.claude/commands/hooks/post-task.md +112 -112
  163. package/.claude/commands/hooks/pre-edit.md +113 -113
  164. package/.claude/commands/hooks/pre-task.md +111 -111
  165. package/.claude/commands/hooks/session-end.md +118 -118
  166. package/.claude/commands/hooks/setup.md +102 -102
  167. package/.claude/commands/memory/README.md +9 -9
  168. package/.claude/commands/memory/memory-persist.md +25 -25
  169. package/.claude/commands/memory/memory-search.md +25 -25
  170. package/.claude/commands/memory/memory-usage.md +25 -25
  171. package/.claude/commands/memory/neural.md +47 -47
  172. package/.claude/commands/monitoring/README.md +9 -9
  173. package/.claude/commands/monitoring/agent-metrics.md +25 -25
  174. package/.claude/commands/monitoring/agents.md +44 -44
  175. package/.claude/commands/monitoring/real-time-view.md +25 -25
  176. package/.claude/commands/monitoring/status.md +46 -46
  177. package/.claude/commands/monitoring/swarm-monitor.md +25 -25
  178. package/.claude/commands/optimization/README.md +9 -9
  179. package/.claude/commands/optimization/auto-topology.md +61 -61
  180. package/.claude/commands/optimization/cache-manage.md +25 -25
  181. package/.claude/commands/optimization/parallel-execute.md +25 -25
  182. package/.claude/commands/optimization/parallel-execution.md +49 -49
  183. package/.claude/commands/optimization/topology-optimize.md +25 -25
  184. package/.claude/commands/pair/README.md +260 -260
  185. package/.claude/commands/pair/commands.md +545 -545
  186. package/.claude/commands/pair/config.md +509 -509
  187. package/.claude/commands/pair/examples.md +511 -511
  188. package/.claude/commands/pair/modes.md +347 -347
  189. package/.claude/commands/pair/session.md +406 -406
  190. package/.claude/commands/pair/start.md +208 -208
  191. package/.claude/commands/sparc/analyzer.md +51 -51
  192. package/.claude/commands/sparc/architect.md +53 -53
  193. package/.claude/commands/sparc/ask.md +97 -97
  194. package/.claude/commands/sparc/batch-executor.md +54 -54
  195. package/.claude/commands/sparc/code.md +89 -89
  196. package/.claude/commands/sparc/coder.md +54 -54
  197. package/.claude/commands/sparc/debug.md +83 -83
  198. package/.claude/commands/sparc/debugger.md +54 -54
  199. package/.claude/commands/sparc/designer.md +53 -53
  200. package/.claude/commands/sparc/devops.md +109 -109
  201. package/.claude/commands/sparc/docs-writer.md +80 -80
  202. package/.claude/commands/sparc/documenter.md +54 -54
  203. package/.claude/commands/sparc/innovator.md +54 -54
  204. package/.claude/commands/sparc/integration.md +83 -83
  205. package/.claude/commands/sparc/mcp.md +117 -117
  206. package/.claude/commands/sparc/memory-manager.md +54 -54
  207. package/.claude/commands/sparc/optimizer.md +54 -54
  208. package/.claude/commands/sparc/orchestrator.md +131 -131
  209. package/.claude/commands/sparc/post-deployment-monitoring-mode.md +83 -83
  210. package/.claude/commands/sparc/refinement-optimization-mode.md +83 -83
  211. package/.claude/commands/sparc/researcher.md +54 -54
  212. package/.claude/commands/sparc/reviewer.md +54 -54
  213. package/.claude/commands/sparc/security-review.md +80 -80
  214. package/.claude/commands/sparc/sparc-modes.md +174 -174
  215. package/.claude/commands/sparc/sparc.md +111 -111
  216. package/.claude/commands/sparc/spec-pseudocode.md +80 -80
  217. package/.claude/commands/sparc/supabase-admin.md +348 -348
  218. package/.claude/commands/sparc/swarm-coordinator.md +54 -54
  219. package/.claude/commands/sparc/tdd.md +54 -54
  220. package/.claude/commands/sparc/tester.md +54 -54
  221. package/.claude/commands/sparc/tutorial.md +79 -79
  222. package/.claude/commands/sparc/workflow-manager.md +54 -54
  223. package/.claude/commands/sparc.md +166 -166
  224. package/.claude/commands/stream-chain/pipeline.md +120 -120
  225. package/.claude/commands/stream-chain/run.md +69 -69
  226. package/.claude/commands/swarm/README.md +15 -15
  227. package/.claude/commands/swarm/analysis.md +95 -95
  228. package/.claude/commands/swarm/development.md +96 -96
  229. package/.claude/commands/swarm/examples.md +168 -168
  230. package/.claude/commands/swarm/maintenance.md +102 -102
  231. package/.claude/commands/swarm/optimization.md +117 -117
  232. package/.claude/commands/swarm/research.md +136 -136
  233. package/.claude/commands/swarm/swarm-analysis.md +8 -8
  234. package/.claude/commands/swarm/swarm-background.md +8 -8
  235. package/.claude/commands/swarm/swarm-init.md +19 -19
  236. package/.claude/commands/swarm/swarm-modes.md +8 -8
  237. package/.claude/commands/swarm/swarm-monitor.md +8 -8
  238. package/.claude/commands/swarm/swarm-spawn.md +19 -19
  239. package/.claude/commands/swarm/swarm-status.md +8 -8
  240. package/.claude/commands/swarm/swarm-strategies.md +8 -8
  241. package/.claude/commands/swarm/swarm.md +87 -87
  242. package/.claude/commands/swarm/testing.md +131 -131
  243. package/.claude/commands/training/README.md +9 -9
  244. package/.claude/commands/training/model-update.md +25 -25
  245. package/.claude/commands/training/neural-patterns.md +107 -107
  246. package/.claude/commands/training/neural-train.md +75 -75
  247. package/.claude/commands/training/pattern-learn.md +25 -25
  248. package/.claude/commands/training/specialization.md +62 -62
  249. package/.claude/commands/truth/start.md +142 -142
  250. package/.claude/commands/verify/check.md +49 -49
  251. package/.claude/commands/verify/start.md +127 -127
  252. package/.claude/commands/workflows/README.md +9 -9
  253. package/.claude/commands/workflows/development.md +77 -77
  254. package/.claude/commands/workflows/research.md +62 -62
  255. package/.claude/commands/workflows/workflow-create.md +25 -25
  256. package/.claude/commands/workflows/workflow-execute.md +25 -25
  257. package/.claude/commands/workflows/workflow-export.md +25 -25
  258. package/.claude/helpers/README.md +96 -96
  259. package/.claude/helpers/adr-compliance.sh +186 -186
  260. package/.claude/helpers/auto-commit.sh +178 -178
  261. package/.claude/helpers/auto-memory-hook.mjs +368 -368
  262. package/.claude/helpers/checkpoint-manager.sh +251 -251
  263. package/.claude/helpers/daemon-manager.sh +252 -252
  264. package/.claude/helpers/ddd-tracker.sh +144 -144
  265. package/.claude/helpers/github-safe.js +156 -156
  266. package/.claude/helpers/github-setup.sh +45 -45
  267. package/.claude/helpers/guidance-hook.sh +13 -13
  268. package/.claude/helpers/guidance-hooks.sh +102 -102
  269. package/.claude/helpers/health-monitor.sh +108 -108
  270. package/.claude/helpers/hook-handler.cjs +286 -286
  271. package/.claude/helpers/intelligence.cjs +1031 -1031
  272. package/.claude/helpers/learning-hooks.sh +329 -329
  273. package/.claude/helpers/learning-optimizer.sh +127 -127
  274. package/.claude/helpers/learning-service.mjs +1144 -1144
  275. package/.claude/helpers/memory.js +83 -83
  276. package/.claude/helpers/metrics-db.mjs +488 -488
  277. package/.claude/helpers/pattern-consolidator.sh +86 -86
  278. package/.claude/helpers/perf-worker.sh +160 -160
  279. package/.claude/helpers/post-commit +16 -16
  280. package/.claude/helpers/pre-commit +26 -26
  281. package/.claude/helpers/quick-start.sh +19 -19
  282. package/.claude/helpers/router.js +105 -105
  283. package/.claude/helpers/security-scanner.sh +127 -127
  284. package/.claude/helpers/session.js +157 -157
  285. package/.claude/helpers/setup-mcp.sh +18 -18
  286. package/.claude/helpers/standard-checkpoint-hooks.sh +189 -189
  287. package/.claude/helpers/statusline-hook.sh +21 -21
  288. package/.claude/helpers/statusline.cjs +604 -604
  289. package/.claude/helpers/statusline.js +352 -352
  290. package/.claude/helpers/swarm-comms.sh +353 -353
  291. package/.claude/helpers/swarm-hooks.sh +761 -761
  292. package/.claude/helpers/swarm-monitor.sh +210 -210
  293. package/.claude/helpers/sync-v3-metrics.sh +245 -245
  294. package/.claude/helpers/update-v3-progress.sh +165 -165
  295. package/.claude/helpers/v3-quick-status.sh +57 -57
  296. package/.claude/helpers/v3.sh +110 -110
  297. package/.claude/helpers/validate-v3-config.sh +215 -215
  298. package/.claude/helpers/worker-manager.sh +170 -170
  299. package/.claude/settings.json +182 -182
  300. package/.claude/skills/agentdb-advanced/SKILL.md +550 -550
  301. package/.claude/skills/agentdb-learning/SKILL.md +545 -545
  302. package/.claude/skills/agentdb-memory-patterns/SKILL.md +339 -339
  303. package/.claude/skills/agentdb-optimization/SKILL.md +509 -509
  304. package/.claude/skills/agentdb-vector-search/SKILL.md +339 -339
  305. package/.claude/skills/browser/SKILL.md +204 -204
  306. package/.claude/skills/dual-mode/README.md +71 -71
  307. package/.claude/skills/dual-mode/dual-collect.md +103 -103
  308. package/.claude/skills/dual-mode/dual-coordinate.md +85 -85
  309. package/.claude/skills/dual-mode/dual-spawn.md +81 -81
  310. package/.claude/skills/flow-nexus-neural/SKILL.md +727 -727
  311. package/.claude/skills/flow-nexus-platform/SKILL.md +1154 -1154
  312. package/.claude/skills/flow-nexus-swarm/SKILL.md +604 -604
  313. package/.claude/skills/github-code-review/SKILL.md +1125 -1125
  314. package/.claude/skills/github-multi-repo/SKILL.md +862 -862
  315. package/.claude/skills/github-project-management/SKILL.md +1262 -1262
  316. package/.claude/skills/github-release-management/SKILL.md +1064 -1064
  317. package/.claude/skills/github-workflow-automation/SKILL.md +1047 -1047
  318. package/.claude/skills/hooks-automation/SKILL.md +1201 -1201
  319. package/.claude/skills/pair-programming/SKILL.md +1202 -1202
  320. package/.claude/skills/reasoningbank-agentdb/SKILL.md +446 -446
  321. package/.claude/skills/reasoningbank-intelligence/SKILL.md +201 -201
  322. package/.claude/skills/skill-builder/SKILL.md +910 -910
  323. package/.claude/skills/sparc-methodology/SKILL.md +1106 -1106
  324. package/.claude/skills/stream-chain/SKILL.md +560 -560
  325. package/.claude/skills/swarm-advanced/SKILL.md +970 -970
  326. package/.claude/skills/swarm-orchestration/SKILL.md +179 -179
  327. package/.claude/skills/v3-cli-modernization/SKILL.md +871 -871
  328. package/.claude/skills/v3-core-implementation/SKILL.md +796 -796
  329. package/.claude/skills/v3-ddd-architecture/SKILL.md +441 -441
  330. package/.claude/skills/v3-integration-deep/SKILL.md +240 -240
  331. package/.claude/skills/v3-mcp-optimization/SKILL.md +776 -776
  332. package/.claude/skills/v3-memory-unification/SKILL.md +173 -173
  333. package/.claude/skills/v3-performance-optimization/SKILL.md +389 -389
  334. package/.claude/skills/v3-security-overhaul/SKILL.md +81 -81
  335. package/.claude/skills/v3-swarm-coordination/SKILL.md +339 -339
  336. package/.claude/skills/verification-quality/SKILL.md +691 -691
  337. package/README.md +412 -412
  338. package/bin/cli.js +314 -314
  339. package/bin/mcp-server.js +224 -224
  340. package/bin/preinstall.cjs +2 -2
  341. package/dist/src/benchmarks/capability-tasks.d.ts +126 -0
  342. package/dist/src/benchmarks/capability-tasks.d.ts.map +1 -0
  343. package/dist/src/benchmarks/capability-tasks.js +147 -0
  344. package/dist/src/benchmarks/capability-tasks.js.map +1 -0
  345. package/dist/src/benchmarks/gaia-agent-contrastive.d.ts +53 -0
  346. package/dist/src/benchmarks/gaia-agent-contrastive.d.ts.map +1 -0
  347. package/dist/src/benchmarks/gaia-agent-contrastive.js +177 -0
  348. package/dist/src/benchmarks/gaia-agent-contrastive.js.map +1 -0
  349. package/dist/src/benchmarks/gaia-agent-gemini.d.ts +51 -0
  350. package/dist/src/benchmarks/gaia-agent-gemini.d.ts.map +1 -0
  351. package/dist/src/benchmarks/gaia-agent-gemini.js +352 -0
  352. package/dist/src/benchmarks/gaia-agent-gemini.js.map +1 -0
  353. package/dist/src/benchmarks/gaia-claude-p.d.ts +105 -0
  354. package/dist/src/benchmarks/gaia-claude-p.d.ts.map +1 -0
  355. package/dist/src/benchmarks/gaia-claude-p.js +269 -0
  356. package/dist/src/benchmarks/gaia-claude-p.js.map +1 -0
  357. package/dist/src/benchmarks/gaia-claude-p.smoke.d.ts +20 -0
  358. package/dist/src/benchmarks/gaia-claude-p.smoke.d.ts.map +1 -0
  359. package/dist/src/benchmarks/gaia-claude-p.smoke.js +168 -0
  360. package/dist/src/benchmarks/gaia-claude-p.smoke.js.map +1 -0
  361. package/dist/src/benchmarks/gaia-codeagent-runner.py +556 -0
  362. package/dist/src/benchmarks/gaia-codeagent.d.ts +129 -0
  363. package/dist/src/benchmarks/gaia-codeagent.d.ts.map +1 -0
  364. package/dist/src/benchmarks/gaia-codeagent.js +590 -0
  365. package/dist/src/benchmarks/gaia-codeagent.js.map +1 -0
  366. package/dist/src/benchmarks/gaia-codeagent.smoke.d.ts +33 -0
  367. package/dist/src/benchmarks/gaia-codeagent.smoke.d.ts.map +1 -0
  368. package/dist/src/benchmarks/gaia-codeagent.smoke.js +241 -0
  369. package/dist/src/benchmarks/gaia-codeagent.smoke.js.map +1 -0
  370. package/dist/src/benchmarks/gaia-critic.js +24 -24
  371. package/dist/src/benchmarks/gaia-dag-pilot.d.ts +22 -0
  372. package/dist/src/benchmarks/gaia-dag-pilot.d.ts.map +1 -0
  373. package/dist/src/benchmarks/gaia-dag-pilot.js +158 -0
  374. package/dist/src/benchmarks/gaia-dag-pilot.js.map +1 -0
  375. package/dist/src/benchmarks/gaia-dag.d.ts +110 -0
  376. package/dist/src/benchmarks/gaia-dag.d.ts.map +1 -0
  377. package/dist/src/benchmarks/gaia-dag.js +658 -0
  378. package/dist/src/benchmarks/gaia-dag.js.map +1 -0
  379. package/dist/src/benchmarks/gaia-dag.smoke.d.ts +21 -0
  380. package/dist/src/benchmarks/gaia-dag.smoke.d.ts.map +1 -0
  381. package/dist/src/benchmarks/gaia-dag.smoke.js +290 -0
  382. package/dist/src/benchmarks/gaia-dag.smoke.js.map +1 -0
  383. package/dist/src/benchmarks/gaia-ensemble-pilot.d.ts +17 -0
  384. package/dist/src/benchmarks/gaia-ensemble-pilot.d.ts.map +1 -0
  385. package/dist/src/benchmarks/gaia-ensemble-pilot.js +106 -0
  386. package/dist/src/benchmarks/gaia-ensemble-pilot.js.map +1 -0
  387. package/dist/src/benchmarks/gaia-ensemble.d.ts +103 -0
  388. package/dist/src/benchmarks/gaia-ensemble.d.ts.map +1 -0
  389. package/dist/src/benchmarks/gaia-ensemble.js +462 -0
  390. package/dist/src/benchmarks/gaia-ensemble.js.map +1 -0
  391. package/dist/src/benchmarks/gaia-gemini-pilot.d.ts +17 -0
  392. package/dist/src/benchmarks/gaia-gemini-pilot.d.ts.map +1 -0
  393. package/dist/src/benchmarks/gaia-gemini-pilot.js +168 -0
  394. package/dist/src/benchmarks/gaia-gemini-pilot.js.map +1 -0
  395. package/dist/src/benchmarks/gaia-mode-router.d.ts +67 -0
  396. package/dist/src/benchmarks/gaia-mode-router.d.ts.map +1 -0
  397. package/dist/src/benchmarks/gaia-mode-router.js +126 -0
  398. package/dist/src/benchmarks/gaia-mode-router.js.map +1 -0
  399. package/dist/src/benchmarks/gaia-mode-router.smoke.d.ts +15 -0
  400. package/dist/src/benchmarks/gaia-mode-router.smoke.d.ts.map +1 -0
  401. package/dist/src/benchmarks/gaia-mode-router.smoke.js +131 -0
  402. package/dist/src/benchmarks/gaia-mode-router.smoke.js.map +1 -0
  403. package/dist/src/benchmarks/gaia-tools/grounded_query.smoke.d.ts +24 -0
  404. package/dist/src/benchmarks/gaia-tools/grounded_query.smoke.d.ts.map +1 -0
  405. package/dist/src/benchmarks/gaia-tools/grounded_query.smoke.js +276 -0
  406. package/dist/src/benchmarks/gaia-tools/grounded_query.smoke.js.map +1 -0
  407. package/dist/src/benchmarks/gaia-tools/image_describe.d.ts +45 -0
  408. package/dist/src/benchmarks/gaia-tools/image_describe.d.ts.map +1 -0
  409. package/dist/src/benchmarks/gaia-tools/image_describe.js +222 -0
  410. package/dist/src/benchmarks/gaia-tools/image_describe.js.map +1 -0
  411. package/dist/src/benchmarks/gaia-tools/image_describe.smoke.d.ts +24 -0
  412. package/dist/src/benchmarks/gaia-tools/image_describe.smoke.d.ts.map +1 -0
  413. package/dist/src/benchmarks/gaia-tools/image_describe.smoke.js +169 -0
  414. package/dist/src/benchmarks/gaia-tools/image_describe.smoke.js.map +1 -0
  415. package/dist/src/benchmarks/gaia-tools/pdf_read.d.ts +27 -0
  416. package/dist/src/benchmarks/gaia-tools/pdf_read.d.ts.map +1 -0
  417. package/dist/src/benchmarks/gaia-tools/pdf_read.js +226 -0
  418. package/dist/src/benchmarks/gaia-tools/pdf_read.js.map +1 -0
  419. package/dist/src/benchmarks/gaia-tools/python_exec.d.ts +31 -0
  420. package/dist/src/benchmarks/gaia-tools/python_exec.d.ts.map +1 -0
  421. package/dist/src/benchmarks/gaia-tools/python_exec.js +94 -0
  422. package/dist/src/benchmarks/gaia-tools/python_exec.js.map +1 -0
  423. package/dist/src/benchmarks/gaia-tools/python_exec.smoke.d.ts +13 -0
  424. package/dist/src/benchmarks/gaia-tools/python_exec.smoke.d.ts.map +1 -0
  425. package/dist/src/benchmarks/gaia-tools/python_exec.smoke.js +98 -0
  426. package/dist/src/benchmarks/gaia-tools/python_exec.smoke.js.map +1 -0
  427. package/dist/src/benchmarks/gaia-tools/visit_webpage.d.ts +52 -0
  428. package/dist/src/benchmarks/gaia-tools/visit_webpage.d.ts.map +1 -0
  429. package/dist/src/benchmarks/gaia-tools/visit_webpage.js +304 -0
  430. package/dist/src/benchmarks/gaia-tools/visit_webpage.js.map +1 -0
  431. package/dist/src/benchmarks/gaia-tools/web_browse.d.ts +61 -0
  432. package/dist/src/benchmarks/gaia-tools/web_browse.d.ts.map +1 -0
  433. package/dist/src/benchmarks/gaia-tools/web_browse.js +217 -0
  434. package/dist/src/benchmarks/gaia-tools/web_browse.js.map +1 -0
  435. package/dist/src/benchmarks/gaia-tools/web_browse.smoke.d.ts +26 -0
  436. package/dist/src/benchmarks/gaia-tools/web_browse.smoke.d.ts.map +1 -0
  437. package/dist/src/benchmarks/gaia-tools/web_browse.smoke.js +152 -0
  438. package/dist/src/benchmarks/gaia-tools/web_browse.smoke.js.map +1 -0
  439. package/dist/src/benchmarks/gaia-tools/web_search.smoke.d.ts +24 -0
  440. package/dist/src/benchmarks/gaia-tools/web_search.smoke.d.ts.map +1 -0
  441. package/dist/src/benchmarks/gaia-tools/web_search.smoke.js +311 -0
  442. package/dist/src/benchmarks/gaia-tools/web_search.smoke.js.map +1 -0
  443. package/dist/src/benchmarks/gaia-visit-webpage.smoke.d.ts +25 -0
  444. package/dist/src/benchmarks/gaia-visit-webpage.smoke.d.ts.map +1 -0
  445. package/dist/src/benchmarks/gaia-visit-webpage.smoke.js +304 -0
  446. package/dist/src/benchmarks/gaia-visit-webpage.smoke.js.map +1 -0
  447. package/dist/src/commands/completions.js +409 -409
  448. package/dist/src/commands/daemon.js +44 -44
  449. package/dist/src/commands/embeddings.js +26 -26
  450. package/dist/src/commands/hive-mind.js +97 -97
  451. package/dist/src/commands/hooks.js +9 -9
  452. package/dist/src/commands/neural.d.ts.map +1 -1
  453. package/dist/src/commands/neural.js +2351 -1
  454. package/dist/src/commands/neural.js.map +1 -1
  455. package/dist/src/commands/performance-capability.d.ts +30 -0
  456. package/dist/src/commands/performance-capability.d.ts.map +1 -0
  457. package/dist/src/commands/performance-capability.js +349 -0
  458. package/dist/src/commands/performance-capability.js.map +1 -0
  459. package/dist/src/commands/ruvector/backup.js +23 -23
  460. package/dist/src/commands/ruvector/benchmark.js +31 -31
  461. package/dist/src/commands/ruvector/import.js +14 -14
  462. package/dist/src/commands/ruvector/init.js +115 -115
  463. package/dist/src/commands/ruvector/migrate.js +99 -99
  464. package/dist/src/commands/ruvector/optimize.js +51 -51
  465. package/dist/src/commands/ruvector/setup.js +624 -624
  466. package/dist/src/commands/ruvector/status.js +38 -38
  467. package/dist/src/init/claudemd-generator.js +231 -231
  468. package/dist/src/init/executor.js +453 -453
  469. package/dist/src/init/helpers-generator.js +727 -727
  470. package/dist/src/init/statusline-generator.js +717 -717
  471. package/dist/src/mcp-tools/agent-execute-core.d.ts +23 -1
  472. package/dist/src/mcp-tools/agent-execute-core.d.ts.map +1 -1
  473. package/dist/src/mcp-tools/agent-execute-core.js +136 -9
  474. package/dist/src/mcp-tools/agent-execute-core.js.map +1 -1
  475. package/dist/src/mcp-tools/agent-tools.d.ts.map +1 -1
  476. package/dist/src/mcp-tools/agent-tools.js +48 -4
  477. package/dist/src/mcp-tools/agent-tools.js.map +1 -1
  478. package/dist/src/mcp-tools/agentdb-tools.js +15 -15
  479. package/dist/src/mcp-tools/hooks-tools.d.ts.map +1 -1
  480. package/dist/src/mcp-tools/hooks-tools.js +196 -0
  481. package/dist/src/mcp-tools/hooks-tools.js.map +1 -1
  482. package/dist/src/memory/graph-edge-writer.js +18 -18
  483. package/dist/src/memory/memory-bridge.js +84 -84
  484. package/dist/src/memory/memory-initializer.js +393 -393
  485. package/dist/src/memory/rabitq-index.js +5 -5
  486. package/dist/src/runtime/headless.js +28 -28
  487. package/dist/src/ruvector/enhanced-model-router.d.ts +5 -0
  488. package/dist/src/ruvector/enhanced-model-router.d.ts.map +1 -1
  489. package/dist/src/ruvector/enhanced-model-router.js +18 -1
  490. package/dist/src/ruvector/enhanced-model-router.js.map +1 -1
  491. package/dist/src/ruvector/model-prices.d.ts +50 -0
  492. package/dist/src/ruvector/model-prices.d.ts.map +1 -0
  493. package/dist/src/ruvector/model-prices.js +72 -0
  494. package/dist/src/ruvector/model-prices.js.map +1 -0
  495. package/dist/src/ruvector/model-router.d.ts +147 -8
  496. package/dist/src/ruvector/model-router.d.ts.map +1 -1
  497. package/dist/src/ruvector/model-router.js +418 -18
  498. package/dist/src/ruvector/model-router.js.map +1 -1
  499. package/dist/src/ruvector/neural-router.d.ts +172 -39
  500. package/dist/src/ruvector/neural-router.d.ts.map +1 -1
  501. package/dist/src/ruvector/neural-router.js +780 -91
  502. package/dist/src/ruvector/neural-router.js.map +1 -1
  503. package/dist/src/ruvector/router-calibrator.d.ts +65 -0
  504. package/dist/src/ruvector/router-calibrator.d.ts.map +1 -0
  505. package/dist/src/ruvector/router-calibrator.js +120 -0
  506. package/dist/src/ruvector/router-calibrator.js.map +1 -0
  507. package/dist/src/ruvector/router-trajectory.d.ts +137 -45
  508. package/dist/src/ruvector/router-trajectory.d.ts.map +1 -1
  509. package/dist/src/ruvector/router-trajectory.js +226 -63
  510. package/dist/src/ruvector/router-trajectory.js.map +1 -1
  511. package/dist/src/ruvector/task-embedder.d.ts +56 -0
  512. package/dist/src/ruvector/task-embedder.d.ts.map +1 -0
  513. package/dist/src/ruvector/task-embedder.js +237 -0
  514. package/dist/src/ruvector/task-embedder.js.map +1 -0
  515. package/dist/src/services/headless-worker-executor.js +84 -84
  516. package/dist/src/transfer/deploy-seraphine.js +23 -23
  517. package/dist/tsconfig.tsbuildinfo +1 -1
  518. package/package.json +136 -134
  519. package/scripts/beir-bootstrap-significance.mjs +162 -162
  520. package/scripts/bench-rvagent.mjs +312 -312
  521. package/scripts/benchmark-codemods.mjs +127 -127
  522. package/scripts/benchmark-cross-repo.mjs +206 -206
  523. package/scripts/benchmark-pretrained-retrieval.mjs +305 -305
  524. package/scripts/benchmark-self-learning.mjs +241 -241
  525. package/scripts/benchmark-trajectory-mrr.mjs +191 -191
  526. package/scripts/deploy-ipfs-node.sh +153 -153
  527. package/scripts/grid-search-retrieval.mjs +216 -216
  528. package/scripts/postinstall.cjs +153 -153
  529. package/scripts/pretrain-from-github.mjs +358 -358
  530. package/scripts/publish-registry.ts +345 -345
  531. package/scripts/publish.sh +57 -57
  532. package/scripts/run-beir-bge.mjs +324 -324
  533. package/scripts/run-beir-hybrid.mjs +261 -261
  534. package/scripts/run-beir-lucene-bm25.mjs +154 -154
  535. package/scripts/run-beir-nfcorpus.mjs +231 -231
  536. package/scripts/run-beir-rrf-ablation.mjs +221 -221
  537. package/scripts/setup-ipfs-registry.md +366 -366
  538. package/dist/src/ruvector/flash-attention.d.ts +0 -195
  539. package/dist/src/ruvector/flash-attention.d.ts.map +0 -1
  540. package/dist/src/ruvector/flash-attention.js +0 -643
  541. package/dist/src/ruvector/flash-attention.js.map +0 -1
  542. package/dist/src/ruvector/moe-router.d.ts +0 -206
  543. package/dist/src/ruvector/moe-router.d.ts.map +0 -1
  544. package/dist/src/ruvector/moe-router.js +0 -626
  545. package/dist/src/ruvector/moe-router.js.map +0 -1
  546. package/dist/src/services/event-stream.d.ts +0 -25
  547. package/dist/src/services/event-stream.js +0 -27
  548. package/dist/src/services/loop-worker-runner.d.ts +0 -16
  549. package/dist/src/services/loop-worker-runner.js +0 -34
  550. package/dist/src/services/runtime-capabilities.d.ts +0 -22
  551. package/dist/src/services/runtime-capabilities.js +0 -45
@@ -0,0 +1,276 @@
1
+ /**
2
+ * Smoke tests for grounded_query.ts — ADR-135 / iter-33
3
+ *
4
+ * All tests are mock-based using groundedQueryTestHooks for ESM-safe injection.
5
+ * NO live API calls are made.
6
+ *
7
+ * Coverage:
8
+ * 1. resolveGoogleAIApiKey — returns key from env var
9
+ * 2. resolveGoogleAIApiKey — throws when env var absent and gcloud unavailable
10
+ * 3. parseGeminiResponse — well-formed response → correct GroundedQueryResult shape
11
+ * 4. parseGeminiResponse — empty grounding metadata → grounded=false, empty sources
12
+ * 5. parseGeminiResponse — deduplicates sources by URI
13
+ * 6. parseGeminiResponse — cost calculation uses token counts
14
+ * 7. GroundedQueryTool.execute — happy path: answer + sources in output
15
+ * 8. GroundedQueryTool.execute — HTTP 4xx surfaces useful error message
16
+ * 9. GroundedQueryTool.execute — ungrounded response includes note in output
17
+ * 10. GroundedQueryTool.execute — rejects empty query
18
+ * 11. GroundedQueryTool.execute — max_tokens clamped to 8192
19
+ * 12. GroundedQueryTool.execute — uses injected apiKey from constructor
20
+ *
21
+ * Run with: npx tsx grounded_query.smoke.ts (from gaia-tools directory)
22
+ */
23
+ import * as assert from 'node:assert/strict';
24
+ import { GroundedQueryTool, groundedQueryTestHooks, resolveGoogleAIApiKey, parseGeminiResponse, createGroundedQueryTool, } from './grounded_query.js';
25
+ // ---------------------------------------------------------------------------
26
+ // Minimal test harness (identical to web_search.smoke.ts — no external deps)
27
+ // ---------------------------------------------------------------------------
28
+ let passed = 0;
29
+ let failed = 0;
30
+ async function test(name, fn) {
31
+ try {
32
+ await fn();
33
+ console.log(` PASS ${name}`);
34
+ passed++;
35
+ }
36
+ catch (err) {
37
+ console.error(` FAIL ${name}`);
38
+ console.error(` ${err.message}`);
39
+ failed++;
40
+ }
41
+ }
42
+ function clearHooks() {
43
+ delete groundedQueryTestHooks.resolveKey;
44
+ delete groundedQueryTestHooks.fetchResponse;
45
+ }
46
+ // ---------------------------------------------------------------------------
47
+ // Shared stub data
48
+ // ---------------------------------------------------------------------------
49
+ const STUB_RESPONSE_GROUNDED = {
50
+ candidates: [
51
+ {
52
+ content: { parts: [{ text: 'Mercedes Sosa released 5 studio albums in that period.' }] },
53
+ groundingMetadata: {
54
+ groundingChunks: [
55
+ { web: { title: 'Mercedes Sosa discography', uri: 'https://example.com/mercedes-sosa' } },
56
+ { web: { title: 'AllMusic Mercedes Sosa', uri: 'https://allmusic.com/sosa' } },
57
+ { web: { title: 'Wikipedia Mercedes Sosa', uri: 'https://en.wikipedia.org/wiki/Mercedes_Sosa' } },
58
+ { web: { title: 'Duplicate entry', uri: 'https://example.com/mercedes-sosa' } }, // duplicate URI
59
+ ],
60
+ webSearchQueries: ['Mercedes Sosa studio albums count'],
61
+ },
62
+ },
63
+ ],
64
+ usageMetadata: { promptTokenCount: 200, candidatesTokenCount: 50 },
65
+ };
66
+ const STUB_RESPONSE_UNGROUNDED = {
67
+ candidates: [
68
+ {
69
+ content: { parts: [{ text: 'I cannot verify this without current data.' }] },
70
+ // No groundingMetadata
71
+ },
72
+ ],
73
+ usageMetadata: { promptTokenCount: 100, candidatesTokenCount: 30 },
74
+ };
75
+ const STUB_RESPONSE_EMPTY_GROUNDING = {
76
+ candidates: [
77
+ {
78
+ content: { parts: [{ text: 'Some answer.' }] },
79
+ groundingMetadata: {
80
+ groundingChunks: [],
81
+ webSearchQueries: [],
82
+ },
83
+ },
84
+ ],
85
+ usageMetadata: { promptTokenCount: 50, candidatesTokenCount: 20 },
86
+ };
87
+ // ---------------------------------------------------------------------------
88
+ // Test suite
89
+ // ---------------------------------------------------------------------------
90
+ async function runSuite() {
91
+ console.log('\ngrounded_query.smoke.ts — ADR-135 / iter-33 Gemini Grounding\n');
92
+ // -------------------------------------------------------------------------
93
+ // 1. resolveGoogleAIApiKey — env var happy path
94
+ // -------------------------------------------------------------------------
95
+ await test('resolveGoogleAIApiKey: returns key from env var', async () => {
96
+ const savedKey = process.env['GOOGLE_AI_API_KEY'];
97
+ process.env['GOOGLE_AI_API_KEY'] = 'test-gemini-key-123';
98
+ try {
99
+ const key = await resolveGoogleAIApiKey();
100
+ assert.equal(key, 'test-gemini-key-123');
101
+ }
102
+ finally {
103
+ if (savedKey === undefined)
104
+ delete process.env['GOOGLE_AI_API_KEY'];
105
+ else
106
+ process.env['GOOGLE_AI_API_KEY'] = savedKey;
107
+ }
108
+ });
109
+ // -------------------------------------------------------------------------
110
+ // 2. resolveGoogleAIApiKey — throws when missing (tested via tool hook path
111
+ // to avoid depending on whether GOOGLE_AI_API_KEY is set in CI or dev env)
112
+ // -------------------------------------------------------------------------
113
+ await test('resolveGoogleAIApiKey: throws with guidance when no key available', async () => {
114
+ // Use the resolveKey hook to simulate the case where no key is configured.
115
+ // This avoids a brittle env-var delete that breaks when the key IS present.
116
+ groundedQueryTestHooks.resolveKey = async () => {
117
+ throw new Error('grounded_query: No Google AI API key found.\n' +
118
+ 'Set GOOGLE_AI_API_KEY env var, or ensure `gcloud` is authenticated.\n' +
119
+ 'Get a free key at: https://aistudio.google.com/apikey');
120
+ };
121
+ const tool = new GroundedQueryTool();
122
+ try {
123
+ await assert.rejects(() => tool.execute({ query: 'test missing key' }), /GOOGLE_AI_API_KEY/);
124
+ }
125
+ finally {
126
+ clearHooks();
127
+ }
128
+ });
129
+ // -------------------------------------------------------------------------
130
+ // 3. parseGeminiResponse — well-formed grounded response
131
+ // -------------------------------------------------------------------------
132
+ await test('parseGeminiResponse: maps answer + sources + queries correctly', async () => {
133
+ const result = parseGeminiResponse(STUB_RESPONSE_GROUNDED);
134
+ assert.equal(result.model, 'gemini-2.5-flash');
135
+ assert.ok(result.answer.includes('5 studio albums'));
136
+ assert.equal(result.grounded, true);
137
+ // 4 chunks but one duplicate URI → 3 unique sources
138
+ assert.equal(result.sources.length, 3, `expected 3 unique sources, got ${result.sources.length}`);
139
+ assert.equal(result.search_queries_used.length, 1);
140
+ assert.equal(result.search_queries_used[0], 'Mercedes Sosa studio albums count');
141
+ });
142
+ // -------------------------------------------------------------------------
143
+ // 4. parseGeminiResponse — empty grounding → grounded=false
144
+ // -------------------------------------------------------------------------
145
+ await test('parseGeminiResponse: empty grounding metadata → grounded=false', async () => {
146
+ const result = parseGeminiResponse(STUB_RESPONSE_EMPTY_GROUNDING);
147
+ assert.equal(result.grounded, false);
148
+ assert.equal(result.sources.length, 0);
149
+ assert.equal(result.search_queries_used.length, 0);
150
+ });
151
+ // -------------------------------------------------------------------------
152
+ // 5. parseGeminiResponse — deduplicates sources by URI (already tested in #3
153
+ // but let's verify explicitly)
154
+ // -------------------------------------------------------------------------
155
+ await test('parseGeminiResponse: deduplicates sources by URI', async () => {
156
+ const result = parseGeminiResponse(STUB_RESPONSE_GROUNDED);
157
+ const uris = result.sources.map((s) => s.uri);
158
+ const uniqueUris = new Set(uris);
159
+ assert.equal(uris.length, uniqueUris.size, 'duplicate URIs found in sources');
160
+ });
161
+ // -------------------------------------------------------------------------
162
+ // 6. parseGeminiResponse — cost calculation
163
+ // -------------------------------------------------------------------------
164
+ await test('parseGeminiResponse: cost_usd calculated from token counts', async () => {
165
+ const result = parseGeminiResponse(STUB_RESPONSE_GROUNDED);
166
+ // 200 input @ $0.075/M + 50 output @ $0.30/M
167
+ const expectedMin = (200 / 1_000_000) * 0.075 + (50 / 1_000_000) * 0.30;
168
+ const expectedMax = expectedMin * 1.01; // 1% tolerance for float arithmetic
169
+ assert.ok(result.cost_usd >= expectedMin && result.cost_usd <= expectedMax, `cost_usd=${result.cost_usd} not in [${expectedMin}, ${expectedMax}]`);
170
+ });
171
+ // -------------------------------------------------------------------------
172
+ // 7. GroundedQueryTool.execute — happy path
173
+ // -------------------------------------------------------------------------
174
+ await test('GroundedQueryTool.execute: happy path produces answer + sources in output', async () => {
175
+ groundedQueryTestHooks.resolveKey = async () => 'injected-key';
176
+ groundedQueryTestHooks.fetchResponse = async () => STUB_RESPONSE_GROUNDED;
177
+ const tool = new GroundedQueryTool();
178
+ try {
179
+ const output = await tool.execute({ query: 'How many albums did Mercedes Sosa release?' });
180
+ assert.ok(output.includes('gemini-2.5-flash'), 'model missing from output');
181
+ assert.ok(output.includes('5 studio albums'), 'answer text missing from output');
182
+ assert.ok(output.includes('Sources:'), 'Sources section missing from output');
183
+ assert.ok(output.includes('example.com/mercedes-sosa'), 'source URI missing from output');
184
+ }
185
+ finally {
186
+ clearHooks();
187
+ }
188
+ });
189
+ // -------------------------------------------------------------------------
190
+ // 8. GroundedQueryTool.execute — HTTP 4xx surfaces useful error
191
+ // -------------------------------------------------------------------------
192
+ await test('GroundedQueryTool.execute: HTTP error surfaces in thrown Error', async () => {
193
+ groundedQueryTestHooks.resolveKey = async () => 'bad-key';
194
+ groundedQueryTestHooks.fetchResponse = async () => {
195
+ throw new Error('grounded_query: Gemini HTTP 400: API key not valid');
196
+ };
197
+ const tool = new GroundedQueryTool();
198
+ try {
199
+ await assert.rejects(() => tool.execute({ query: 'test error' }), /Gemini HTTP 400/);
200
+ }
201
+ finally {
202
+ clearHooks();
203
+ }
204
+ });
205
+ // -------------------------------------------------------------------------
206
+ // 9. GroundedQueryTool.execute — ungrounded response includes note
207
+ // -------------------------------------------------------------------------
208
+ await test('GroundedQueryTool.execute: ungrounded response includes note in output', async () => {
209
+ groundedQueryTestHooks.resolveKey = async () => 'injected-key';
210
+ groundedQueryTestHooks.fetchResponse = async () => STUB_RESPONSE_UNGROUNDED;
211
+ const tool = new GroundedQueryTool();
212
+ try {
213
+ const output = await tool.execute({ query: 'some query' });
214
+ assert.ok(output.includes('unverified'), `expected unverified note in: ${output}`);
215
+ }
216
+ finally {
217
+ clearHooks();
218
+ }
219
+ });
220
+ // -------------------------------------------------------------------------
221
+ // 10. GroundedQueryTool.execute — rejects empty query
222
+ // -------------------------------------------------------------------------
223
+ await test('GroundedQueryTool.execute: rejects empty query', async () => {
224
+ const tool = new GroundedQueryTool();
225
+ await assert.rejects(() => tool.execute({ query: '' }), /query.*required/);
226
+ });
227
+ // -------------------------------------------------------------------------
228
+ // 11. GroundedQueryTool.execute — max_tokens clamped
229
+ // -------------------------------------------------------------------------
230
+ await test('GroundedQueryTool.execute: max_tokens clamped to 8192', async () => {
231
+ let capturedMaxTokens = 0;
232
+ groundedQueryTestHooks.resolveKey = async () => 'injected-key';
233
+ groundedQueryTestHooks.fetchResponse = async (_q, maxTokens) => {
234
+ capturedMaxTokens = maxTokens;
235
+ return STUB_RESPONSE_GROUNDED;
236
+ };
237
+ const tool = new GroundedQueryTool();
238
+ try {
239
+ await tool.execute({ query: 'test max tokens', max_tokens: 99_999 });
240
+ assert.equal(capturedMaxTokens, 8192, `expected 8192, got ${capturedMaxTokens}`);
241
+ }
242
+ finally {
243
+ clearHooks();
244
+ }
245
+ });
246
+ // -------------------------------------------------------------------------
247
+ // 12. GroundedQueryTool — uses injected apiKey from constructor
248
+ // -------------------------------------------------------------------------
249
+ await test('createGroundedQueryTool: uses apiKey passed to constructor', async () => {
250
+ let capturedKey = '';
251
+ groundedQueryTestHooks.fetchResponse = async (_q, _max, key) => {
252
+ capturedKey = key;
253
+ return STUB_RESPONSE_GROUNDED;
254
+ };
255
+ const tool = createGroundedQueryTool({ apiKey: 'constructor-key' });
256
+ try {
257
+ await tool.execute({ query: 'test constructor key' });
258
+ assert.equal(capturedKey, 'constructor-key', `expected 'constructor-key', got '${capturedKey}'`);
259
+ }
260
+ finally {
261
+ clearHooks();
262
+ }
263
+ });
264
+ // ---------------------------------------------------------------------------
265
+ // Summary
266
+ // ---------------------------------------------------------------------------
267
+ console.log(`\n${passed} passed, ${failed} failed\n`);
268
+ if (failed > 0) {
269
+ process.exit(1);
270
+ }
271
+ }
272
+ runSuite().catch((err) => {
273
+ console.error('Smoke test runner error:', err);
274
+ process.exit(1);
275
+ });
276
+ //# sourceMappingURL=grounded_query.smoke.js.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"grounded_query.smoke.js","sourceRoot":"","sources":["../../../../src/benchmarks/gaia-tools/grounded_query.smoke.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;GAqBG;AAEH,OAAO,KAAK,MAAM,MAAM,oBAAoB,CAAC;AAC7C,OAAO,EACL,iBAAiB,EACjB,sBAAsB,EACtB,qBAAqB,EACrB,mBAAmB,EACnB,uBAAuB,GACxB,MAAM,qBAAqB,CAAC;AAE7B,8EAA8E;AAC9E,6EAA6E;AAC7E,8EAA8E;AAE9E,IAAI,MAAM,GAAG,CAAC,CAAC;AACf,IAAI,MAAM,GAAG,CAAC,CAAC;AAEf,KAAK,UAAU,IAAI,CAAC,IAAY,EAAE,EAAuB;IACvD,IAAI,CAAC;QACH,MAAM,EAAE,EAAE,CAAC;QACX,OAAO,CAAC,GAAG,CAAC,WAAW,IAAI,EAAE,CAAC,CAAC;QAC/B,MAAM,EAAE,CAAC;IACX,CAAC;IAAC,OAAO,GAAG,EAAE,CAAC;QACb,OAAO,CAAC,KAAK,CAAC,WAAW,IAAI,EAAE,CAAC,CAAC;QACjC,OAAO,CAAC,KAAK,CAAC,WAAY,GAAa,CAAC,OAAO,EAAE,CAAC,CAAC;QACnD,MAAM,EAAE,CAAC;IACX,CAAC;AACH,CAAC;AAED,SAAS,UAAU;IACjB,OAAO,sBAAsB,CAAC,UAAU,CAAC;IACzC,OAAO,sBAAsB,CAAC,aAAa,CAAC;AAC9C,CAAC;AAED,8EAA8E;AAC9E,mBAAmB;AACnB,8EAA8E;AAE9E,MAAM,sBAAsB,GAAG;IAC7B,UAAU,EAAE;QACV;YACE,OAAO,EAAE,EAAE,KAAK,EAAE,CAAC,EAAE,IAAI,EAAE,wDAAwD,EAAE,CAAC,EAAE;YACxF,iBAAiB,EAAE;gBACjB,eAAe,EAAE;oBACf,EAAE,GAAG,EAAE,EAAE,KAAK,EAAE,2BAA2B,EAAE,GAAG,EAAE,mCAAmC,EAAE,EAAE;oBACzF,EAAE,GAAG,EAAE,EAAE,KAAK,EAAE,wBAAwB,EAAE,GAAG,EAAE,2BAA2B,EAAE,EAAE;oBAC9E,EAAE,GAAG,EAAE,EAAE,KAAK,EAAE,yBAAyB,EAAE,GAAG,EAAE,6CAA6C,EAAE,EAAE;oBACjG,EAAE,GAAG,EAAE,EAAE,KAAK,EAAE,iBAAiB,EAAE,GAAG,EAAE,mCAAmC,EAAE,EAAE,EAAE,gBAAgB;iBAClG;gBACD,gBAAgB,EAAE,CAAC,mCAAmC,CAAC;aACxD;SACF;KACF;IACD,aAAa,EAAE,EAAE,gBAAgB,EAAE,GAAG,EAAE,oBAAoB,EAAE,EAAE,EAAE;CACnE,CAAC;AAEF,MAAM,wBAAwB,GAAG;IAC/B,UAAU,EAAE;QACV;YACE,OAAO,EAAE,EAAE,KAAK,EAAE,CAAC,EAAE,IAAI,EAAE,4CAA4C,EAAE,CAAC,EAAE;YAC5E,uBAAuB;SACxB;KACF;IACD,aAAa,EAAE,EAAE,gBAAgB,EAAE,GAAG,EAAE,oBAAoB,EAAE,EAAE,EAAE;CACnE,CAAC;AAEF,MAAM,6BAA6B,GAAG;IACpC,UAAU,EAAE;QACV;YACE,OAAO,EAAE,EAAE,KAAK,EAAE,CAAC,EAAE,IAAI,EAAE,cAAc,EAAE,CAAC,EAAE;YAC9C,iBAAiB,EAAE;gBACjB,eAAe,EAAE,EAAE;gBACnB,gBAAgB,EAAE,EAAE;aACrB;SACF;KACF;IACD,aAAa,EAAE,EAAE,gBAAgB,EAAE,EAAE,EAAE,oBAAoB,EAAE,EAAE,EAAE;CAClE,CAAC;AAEF,8EAA8E;AAC9E,aAAa;AACb,8EAA8E;AAE9E,KAAK,UAAU,QAAQ;IACrB,OAAO,CAAC,GAAG,CAAC,kEAAkE,CAAC,CAAC;IAEhF,4EAA4E;IAC5E,gDAAgD;IAChD,4EAA4E;IAC5E,MAAM,IAAI,CAAC,iDAAiD,EAAE,KAAK,IAAI,EAAE;QACvE,MAAM,QAAQ,GAAG,OAAO,CAAC,GAAG,CAAC,mBAAmB,CAAC,CAAC;QAClD,OAAO,CAAC,GAAG,CAAC,mBAAmB,CAAC,GAAG,qBAAqB,CAAC;QACzD,IAAI,CAAC;YACH,MAAM,GAAG,GAAG,MAAM,qBAAqB,EAAE,CAAC;YAC1C,MAAM,CAAC,KAAK,CAAC,GAAG,EAAE,qBAAqB,CAAC,CAAC;QAC3C,CAAC;gBAAS,CAAC;YACT,IAAI,QAAQ,KAAK,SAAS;gBAAE,OAAO,OAAO,CAAC,GAAG,CAAC,mBAAmB,CAAC,CAAC;;gBAC/D,OAAO,CAAC,GAAG,CAAC,mBAAmB,CAAC,GAAG,QAAQ,CAAC;QACnD,CAAC;IACH,CAAC,CAAC,CAAC;IAEH,4EAA4E;IAC5E,4EAA4E;IAC5E,8EAA8E;IAC9E,4EAA4E;IAC5E,MAAM,IAAI,CAAC,mEAAmE,EAAE,KAAK,IAAI,EAAE;QACzF,2EAA2E;QAC3E,4EAA4E;QAC5E,sBAAsB,CAAC,UAAU,GAAG,KAAK,IAAI,EAAE;YAC7C,MAAM,IAAI,KAAK,CACb,+CAA+C;gBAC7C,uEAAuE;gBACvE,uDAAuD,CAC1D,CAAC;QACJ,CAAC,CAAC;QACF,MAAM,IAAI,GAAG,IAAI,iBAAiB,EAAE,CAAC;QACrC,IAAI,CAAC;YACH,MAAM,MAAM,CAAC,OAAO,CAClB,GAAG,EAAE,CAAC,IAAI,CAAC,OAAO,CAAC,EAAE,KAAK,EAAE,kBAAkB,EAAE,CAAC,EACjD,mBAAmB,CACpB,CAAC;QACJ,CAAC;gBAAS,CAAC;YACT,UAAU,EAAE,CAAC;QACf,CAAC;IACH,CAAC,CAAC,CAAC;IAEH,4EAA4E;IAC5E,yDAAyD;IACzD,4EAA4E;IAC5E,MAAM,IAAI,CAAC,gEAAgE,EAAE,KAAK,IAAI,EAAE;QACtF,MAAM,MAAM,GAAG,mBAAmB,CAAC,sBAAsB,CAAC,CAAC;QAC3D,MAAM,CAAC,KAAK,CAAC,MAAM,CAAC,KAAK,EAAE,kBAAkB,CAAC,CAAC;QAC/C,MAAM,CAAC,EAAE,CAAC,MAAM,CAAC,MAAM,CAAC,QAAQ,CAAC,iBAAiB,CAAC,CAAC,CAAC;QACrD,MAAM,CAAC,KAAK,CAAC,MAAM,CAAC,QAAQ,EAAE,IAAI,CAAC,CAAC;QACpC,oDAAoD;QACpD,MAAM,CAAC,KAAK,CAAC,MAAM,CAAC,OAAO,CAAC,MAAM,EAAE,CAAC,EAAE,kCAAkC,MAAM,CAAC,OAAO,CAAC,MAAM,EAAE,CAAC,CAAC;QAClG,MAAM,CAAC,KAAK,CAAC,MAAM,CAAC,mBAAmB,CAAC,MAAM,EAAE,CAAC,CAAC,CAAC;QACnD,MAAM,CAAC,KAAK,CAAC,MAAM,CAAC,mBAAmB,CAAC,CAAC,CAAC,EAAE,mCAAmC,CAAC,CAAC;IACnF,CAAC,CAAC,CAAC;IAEH,4EAA4E;IAC5E,4DAA4D;IAC5D,4EAA4E;IAC5E,MAAM,IAAI,CAAC,gEAAgE,EAAE,KAAK,IAAI,EAAE;QACtF,MAAM,MAAM,GAAG,mBAAmB,CAAC,6BAA6B,CAAC,CAAC;QAClE,MAAM,CAAC,KAAK,CAAC,MAAM,CAAC,QAAQ,EAAE,KAAK,CAAC,CAAC;QACrC,MAAM,CAAC,KAAK,CAAC,MAAM,CAAC,OAAO,CAAC,MAAM,EAAE,CAAC,CAAC,CAAC;QACvC,MAAM,CAAC,KAAK,CAAC,MAAM,CAAC,mBAAmB,CAAC,MAAM,EAAE,CAAC,CAAC,CAAC;IACrD,CAAC,CAAC,CAAC;IAEH,4EAA4E;IAC5E,6EAA6E;IAC7E,kCAAkC;IAClC,4EAA4E;IAC5E,MAAM,IAAI,CAAC,kDAAkD,EAAE,KAAK,IAAI,EAAE;QACxE,MAAM,MAAM,GAAG,mBAAmB,CAAC,sBAAsB,CAAC,CAAC;QAC3D,MAAM,IAAI,GAAG,MAAM,CAAC,OAAO,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC;QAC9C,MAAM,UAAU,GAAG,IAAI,GAAG,CAAC,IAAI,CAAC,CAAC;QACjC,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,MAAM,EAAE,UAAU,CAAC,IAAI,EAAE,iCAAiC,CAAC,CAAC;IAChF,CAAC,CAAC,CAAC;IAEH,4EAA4E;IAC5E,4CAA4C;IAC5C,4EAA4E;IAC5E,MAAM,IAAI,CAAC,4DAA4D,EAAE,KAAK,IAAI,EAAE;QAClF,MAAM,MAAM,GAAG,mBAAmB,CAAC,sBAAsB,CAAC,CAAC;QAC3D,6CAA6C;QAC7C,MAAM,WAAW,GAAG,CAAC,GAAG,GAAG,SAAS,CAAC,GAAG,KAAK,GAAG,CAAC,EAAE,GAAG,SAAS,CAAC,GAAG,IAAI,CAAC;QACxE,MAAM,WAAW,GAAG,WAAW,GAAG,IAAI,CAAC,CAAC,oCAAoC;QAC5E,MAAM,CAAC,EAAE,CACP,MAAM,CAAC,QAAQ,IAAI,WAAW,IAAI,MAAM,CAAC,QAAQ,IAAI,WAAW,EAChE,YAAY,MAAM,CAAC,QAAQ,YAAY,WAAW,KAAK,WAAW,GAAG,CACtE,CAAC;IACJ,CAAC,CAAC,CAAC;IAEH,4EAA4E;IAC5E,4CAA4C;IAC5C,4EAA4E;IAC5E,MAAM,IAAI,CAAC,2EAA2E,EAAE,KAAK,IAAI,EAAE;QACjG,sBAAsB,CAAC,UAAU,GAAG,KAAK,IAAI,EAAE,CAAC,cAAc,CAAC;QAC/D,sBAAsB,CAAC,aAAa,GAAG,KAAK,IAAI,EAAE,CAAC,sBAAsB,CAAC;QAE1E,MAAM,IAAI,GAAG,IAAI,iBAAiB,EAAE,CAAC;QACrC,IAAI,CAAC;YACH,MAAM,MAAM,GAAG,MAAM,IAAI,CAAC,OAAO,CAAC,EAAE,KAAK,EAAE,4CAA4C,EAAE,CAAC,CAAC;YAC3F,MAAM,CAAC,EAAE,CAAC,MAAM,CAAC,QAAQ,CAAC,kBAAkB,CAAC,EAAE,2BAA2B,CAAC,CAAC;YAC5E,MAAM,CAAC,EAAE,CAAC,MAAM,CAAC,QAAQ,CAAC,iBAAiB,CAAC,EAAE,iCAAiC,CAAC,CAAC;YACjF,MAAM,CAAC,EAAE,CAAC,MAAM,CAAC,QAAQ,CAAC,UAAU,CAAC,EAAE,qCAAqC,CAAC,CAAC;YAC9E,MAAM,CAAC,EAAE,CAAC,MAAM,CAAC,QAAQ,CAAC,2BAA2B,CAAC,EAAE,gCAAgC,CAAC,CAAC;QAC5F,CAAC;gBAAS,CAAC;YACT,UAAU,EAAE,CAAC;QACf,CAAC;IACH,CAAC,CAAC,CAAC;IAEH,4EAA4E;IAC5E,gEAAgE;IAChE,4EAA4E;IAC5E,MAAM,IAAI,CAAC,gEAAgE,EAAE,KAAK,IAAI,EAAE;QACtF,sBAAsB,CAAC,UAAU,GAAG,KAAK,IAAI,EAAE,CAAC,SAAS,CAAC;QAC1D,sBAAsB,CAAC,aAAa,GAAG,KAAK,IAAI,EAAE;YAChD,MAAM,IAAI,KAAK,CAAC,oDAAoD,CAAC,CAAC;QACxE,CAAC,CAAC;QAEF,MAAM,IAAI,GAAG,IAAI,iBAAiB,EAAE,CAAC;QACrC,IAAI,CAAC;YACH,MAAM,MAAM,CAAC,OAAO,CAClB,GAAG,EAAE,CAAC,IAAI,CAAC,OAAO,CAAC,EAAE,KAAK,EAAE,YAAY,EAAE,CAAC,EAC3C,iBAAiB,CAClB,CAAC;QACJ,CAAC;gBAAS,CAAC;YACT,UAAU,EAAE,CAAC;QACf,CAAC;IACH,CAAC,CAAC,CAAC;IAEH,4EAA4E;IAC5E,mEAAmE;IACnE,4EAA4E;IAC5E,MAAM,IAAI,CAAC,wEAAwE,EAAE,KAAK,IAAI,EAAE;QAC9F,sBAAsB,CAAC,UAAU,GAAG,KAAK,IAAI,EAAE,CAAC,cAAc,CAAC;QAC/D,sBAAsB,CAAC,aAAa,GAAG,KAAK,IAAI,EAAE,CAAC,wBAAwB,CAAC;QAE5E,MAAM,IAAI,GAAG,IAAI,iBAAiB,EAAE,CAAC;QACrC,IAAI,CAAC;YACH,MAAM,MAAM,GAAG,MAAM,IAAI,CAAC,OAAO,CAAC,EAAE,KAAK,EAAE,YAAY,EAAE,CAAC,CAAC;YAC3D,MAAM,CAAC,EAAE,CAAC,MAAM,CAAC,QAAQ,CAAC,YAAY,CAAC,EAAE,gCAAgC,MAAM,EAAE,CAAC,CAAC;QACrF,CAAC;gBAAS,CAAC;YACT,UAAU,EAAE,CAAC;QACf,CAAC;IACH,CAAC,CAAC,CAAC;IAEH,4EAA4E;IAC5E,sDAAsD;IACtD,4EAA4E;IAC5E,MAAM,IAAI,CAAC,gDAAgD,EAAE,KAAK,IAAI,EAAE;QACtE,MAAM,IAAI,GAAG,IAAI,iBAAiB,EAAE,CAAC;QACrC,MAAM,MAAM,CAAC,OAAO,CAClB,GAAG,EAAE,CAAC,IAAI,CAAC,OAAO,CAAC,EAAE,KAAK,EAAE,EAAE,EAAE,CAAC,EACjC,iBAAiB,CAClB,CAAC;IACJ,CAAC,CAAC,CAAC;IAEH,4EAA4E;IAC5E,qDAAqD;IACrD,4EAA4E;IAC5E,MAAM,IAAI,CAAC,uDAAuD,EAAE,KAAK,IAAI,EAAE;QAC7E,IAAI,iBAAiB,GAAG,CAAC,CAAC;QAC1B,sBAAsB,CAAC,UAAU,GAAG,KAAK,IAAI,EAAE,CAAC,cAAc,CAAC;QAC/D,sBAAsB,CAAC,aAAa,GAAG,KAAK,EAAE,EAAE,EAAE,SAAS,EAAE,EAAE;YAC7D,iBAAiB,GAAG,SAAS,CAAC;YAC9B,OAAO,sBAAsB,CAAC;QAChC,CAAC,CAAC;QAEF,MAAM,IAAI,GAAG,IAAI,iBAAiB,EAAE,CAAC;QACrC,IAAI,CAAC;YACH,MAAM,IAAI,CAAC,OAAO,CAAC,EAAE,KAAK,EAAE,iBAAiB,EAAE,UAAU,EAAE,MAAM,EAAE,CAAC,CAAC;YACrE,MAAM,CAAC,KAAK,CAAC,iBAAiB,EAAE,IAAI,EAAE,sBAAsB,iBAAiB,EAAE,CAAC,CAAC;QACnF,CAAC;gBAAS,CAAC;YACT,UAAU,EAAE,CAAC;QACf,CAAC;IACH,CAAC,CAAC,CAAC;IAEH,4EAA4E;IAC5E,gEAAgE;IAChE,4EAA4E;IAC5E,MAAM,IAAI,CAAC,4DAA4D,EAAE,KAAK,IAAI,EAAE;QAClF,IAAI,WAAW,GAAG,EAAE,CAAC;QACrB,sBAAsB,CAAC,aAAa,GAAG,KAAK,EAAE,EAAE,EAAE,IAAI,EAAE,GAAG,EAAE,EAAE;YAC7D,WAAW,GAAG,GAAG,CAAC;YAClB,OAAO,sBAAsB,CAAC;QAChC,CAAC,CAAC;QAEF,MAAM,IAAI,GAAG,uBAAuB,CAAC,EAAE,MAAM,EAAE,iBAAiB,EAAE,CAAC,CAAC;QACpE,IAAI,CAAC;YACH,MAAM,IAAI,CAAC,OAAO,CAAC,EAAE,KAAK,EAAE,sBAAsB,EAAE,CAAC,CAAC;YACtD,MAAM,CAAC,KAAK,CAAC,WAAW,EAAE,iBAAiB,EAAE,oCAAoC,WAAW,GAAG,CAAC,CAAC;QACnG,CAAC;gBAAS,CAAC;YACT,UAAU,EAAE,CAAC;QACf,CAAC;IACH,CAAC,CAAC,CAAC;IAEH,8EAA8E;IAC9E,UAAU;IACV,8EAA8E;IAC9E,OAAO,CAAC,GAAG,CAAC,KAAK,MAAM,YAAY,MAAM,WAAW,CAAC,CAAC;IACtD,IAAI,MAAM,GAAG,CAAC,EAAE,CAAC;QACf,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;IAClB,CAAC;AACH,CAAC;AAED,QAAQ,EAAE,CAAC,KAAK,CAAC,CAAC,GAAG,EAAE,EAAE;IACvB,OAAO,CAAC,KAAK,CAAC,0BAA0B,EAAE,GAAG,CAAC,CAAC;IAC/C,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;AAClB,CAAC,CAAC,CAAC"}
@@ -0,0 +1,45 @@
1
+ /**
2
+ * GAIA Tool: image_describe — ADR-133-PR5
3
+ *
4
+ * Describes an image at a given URL or local file path using Anthropic's
5
+ * claude-haiku-4-5 model with vision (image content blocks). Covers the
6
+ * subset of GAIA Level-1 questions that provide image attachments (graphs,
7
+ * screenshots, photos, diagrams).
8
+ *
9
+ * ============================================================
10
+ * DESIGN NOTES
11
+ * ============================================================
12
+ * - Uses the Anthropic Messages API directly via `fetch` (same approach as
13
+ * gaia-agent.ts / gaia-judge.ts) — no SDK dependency.
14
+ * - API key resolution order mirrors gaia-agent.ts:
15
+ * 1. `options.apiKey` (caller-supplied)
16
+ * 2. ANTHROPIC_API_KEY env var
17
+ * 3. gcloud secrets versions access latest --secret=ANTHROPIC_API_KEY
18
+ * - Model: claude-haiku-4-5 (cheapest vision-capable model, ~$0.001/call).
19
+ * - URL images: sent as { type: 'url', url } — Anthropic fetches the image.
20
+ * - Local files: read as Buffer, base64-encoded, MIME detected from extension.
21
+ * - execute() NEVER throws — returns a structured error string so the agent
22
+ * loop can forward it to Claude rather than crashing.
23
+ *
24
+ * ============================================================
25
+ * SUPPORTED IMAGE FORMATS
26
+ * ============================================================
27
+ * Anthropic vision accepts: JPEG, PNG, GIF, WebP.
28
+ * Unsupported formats return a descriptive error that Claude can relay.
29
+ *
30
+ * Refs: ADR-133, #2156
31
+ */
32
+ import { GaiaTool, ToolDefinition } from './types.js';
33
+ export declare class ImageDescribeTool implements GaiaTool {
34
+ readonly name = "image_describe";
35
+ private readonly apiKey?;
36
+ constructor(apiKey?: string);
37
+ readonly definition: ToolDefinition;
38
+ execute(input: Record<string, unknown>): Promise<string>;
39
+ }
40
+ export interface ImageDescribeToolOptions {
41
+ /** Anthropic API key (falls back to env / gcloud if omitted). */
42
+ apiKey?: string;
43
+ }
44
+ export declare function createImageDescribeTool(opts?: ImageDescribeToolOptions): ImageDescribeTool;
45
+ //# sourceMappingURL=image_describe.d.ts.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"image_describe.d.ts","sourceRoot":"","sources":["../../../../src/benchmarks/gaia-tools/image_describe.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GA8BG;AAKH,OAAO,EAAE,QAAQ,EAAE,cAAc,EAAE,MAAM,YAAY,CAAC;AAwJtD,qBAAa,iBAAkB,YAAW,QAAQ;IAChD,QAAQ,CAAC,IAAI,oBAAoB;IACjC,OAAO,CAAC,QAAQ,CAAC,MAAM,CAAC,CAAS;gBAErB,MAAM,CAAC,EAAE,MAAM;IAI3B,QAAQ,CAAC,UAAU,EAAE,cAAc,CA0BjC;IAEI,OAAO,CAAC,KAAK,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,GAAG,OAAO,CAAC,MAAM,CAAC;CA6D/D;AAMD,MAAM,WAAW,wBAAwB;IACvC,iEAAiE;IACjE,MAAM,CAAC,EAAE,MAAM,CAAC;CACjB;AAED,wBAAgB,uBAAuB,CAAC,IAAI,CAAC,EAAE,wBAAwB,GAAG,iBAAiB,CAE1F"}
@@ -0,0 +1,222 @@
1
+ /**
2
+ * GAIA Tool: image_describe — ADR-133-PR5
3
+ *
4
+ * Describes an image at a given URL or local file path using Anthropic's
5
+ * claude-haiku-4-5 model with vision (image content blocks). Covers the
6
+ * subset of GAIA Level-1 questions that provide image attachments (graphs,
7
+ * screenshots, photos, diagrams).
8
+ *
9
+ * ============================================================
10
+ * DESIGN NOTES
11
+ * ============================================================
12
+ * - Uses the Anthropic Messages API directly via `fetch` (same approach as
13
+ * gaia-agent.ts / gaia-judge.ts) — no SDK dependency.
14
+ * - API key resolution order mirrors gaia-agent.ts:
15
+ * 1. `options.apiKey` (caller-supplied)
16
+ * 2. ANTHROPIC_API_KEY env var
17
+ * 3. gcloud secrets versions access latest --secret=ANTHROPIC_API_KEY
18
+ * - Model: claude-haiku-4-5 (cheapest vision-capable model, ~$0.001/call).
19
+ * - URL images: sent as { type: 'url', url } — Anthropic fetches the image.
20
+ * - Local files: read as Buffer, base64-encoded, MIME detected from extension.
21
+ * - execute() NEVER throws — returns a structured error string so the agent
22
+ * loop can forward it to Claude rather than crashing.
23
+ *
24
+ * ============================================================
25
+ * SUPPORTED IMAGE FORMATS
26
+ * ============================================================
27
+ * Anthropic vision accepts: JPEG, PNG, GIF, WebP.
28
+ * Unsupported formats return a descriptive error that Claude can relay.
29
+ *
30
+ * Refs: ADR-133, #2156
31
+ */
32
+ import * as fs from 'node:fs';
33
+ import * as path from 'node:path';
34
+ import { execSync } from 'node:child_process';
35
+ // ---------------------------------------------------------------------------
36
+ // Constants
37
+ // ---------------------------------------------------------------------------
38
+ const VISION_MODEL = 'claude-haiku-4-5';
39
+ const ANTHROPIC_API_VERSION = '2023-06-01';
40
+ const ANTHROPIC_API_URL = 'https://api.anthropic.com/v1/messages';
41
+ const MAX_TOKENS = 512;
42
+ const DEFAULT_PROMPT = 'Describe this image in detail. Note any text, charts, diagrams, tables, ' +
43
+ 'numerical data, labels, axes, legends, or other content that would be ' +
44
+ 'useful to answer a factual question about what the image shows.';
45
+ const EXT_TO_MIME = {
46
+ '.jpg': 'image/jpeg',
47
+ '.jpeg': 'image/jpeg',
48
+ '.png': 'image/png',
49
+ '.gif': 'image/gif',
50
+ '.webp': 'image/webp',
51
+ };
52
+ function detectMime(filePath, buf) {
53
+ // Magic bytes first
54
+ if (buf.length >= 2 && buf[0] === 0xff && buf[1] === 0xd8)
55
+ return 'image/jpeg';
56
+ if (buf.length >= 4 &&
57
+ buf[0] === 0x89 &&
58
+ buf[1] === 0x50 &&
59
+ buf[2] === 0x4e &&
60
+ buf[3] === 0x47)
61
+ return 'image/png';
62
+ if (buf.length >= 4 &&
63
+ buf[0] === 0x47 &&
64
+ buf[1] === 0x49 &&
65
+ buf[2] === 0x46 &&
66
+ buf[3] === 0x38)
67
+ return 'image/gif';
68
+ if (buf.length >= 4 &&
69
+ buf[0] === 0x52 &&
70
+ buf[1] === 0x49 &&
71
+ buf[2] === 0x46 &&
72
+ buf[3] === 0x46)
73
+ return 'image/webp';
74
+ // Fall back to extension
75
+ const ext = path.extname(filePath).toLowerCase();
76
+ return EXT_TO_MIME[ext] ?? 'image/png';
77
+ }
78
+ // ---------------------------------------------------------------------------
79
+ // API key resolution (mirrors gaia-agent.ts)
80
+ // ---------------------------------------------------------------------------
81
+ function resolveAnthropicApiKey(suppliedKey) {
82
+ if (suppliedKey && suppliedKey.trim())
83
+ return suppliedKey.trim();
84
+ const envKey = process.env.ANTHROPIC_API_KEY;
85
+ if (envKey && envKey.trim())
86
+ return envKey.trim();
87
+ try {
88
+ const out = execSync('gcloud secrets versions access latest --secret=ANTHROPIC_API_KEY 2>/dev/null', { encoding: 'utf-8', timeout: 10_000 }).trim();
89
+ if (out)
90
+ return out;
91
+ }
92
+ catch {
93
+ /* fall through */
94
+ }
95
+ throw new Error('ANTHROPIC_API_KEY not found. Set the env var or store it in GCP Secret Manager ' +
96
+ 'under "ANTHROPIC_API_KEY".');
97
+ }
98
+ async function callVisionApi(imageBlock, prompt, apiKey) {
99
+ const body = JSON.stringify({
100
+ model: VISION_MODEL,
101
+ max_tokens: MAX_TOKENS,
102
+ messages: [
103
+ {
104
+ role: 'user',
105
+ content: [imageBlock, { type: 'text', text: prompt }],
106
+ },
107
+ ],
108
+ });
109
+ const resp = await fetch(ANTHROPIC_API_URL, {
110
+ method: 'POST',
111
+ headers: {
112
+ 'x-api-key': apiKey,
113
+ 'anthropic-version': ANTHROPIC_API_VERSION,
114
+ 'content-type': 'application/json',
115
+ },
116
+ body,
117
+ });
118
+ if (!resp.ok) {
119
+ const errText = await resp.text().catch(() => '(no body)');
120
+ throw new Error(`Anthropic API error ${resp.status}: ${errText}`);
121
+ }
122
+ const json = (await resp.json());
123
+ if (json.error)
124
+ throw new Error(`Anthropic API error: ${json.error.message}`);
125
+ const textBlock = json.content?.find((b) => b.type === 'text');
126
+ return textBlock?.text ?? '(no description returned)';
127
+ }
128
+ // ---------------------------------------------------------------------------
129
+ // GaiaTool implementation
130
+ // ---------------------------------------------------------------------------
131
+ export class ImageDescribeTool {
132
+ name = 'image_describe';
133
+ apiKey;
134
+ constructor(apiKey) {
135
+ this.apiKey = apiKey;
136
+ }
137
+ definition = {
138
+ name: 'image_describe',
139
+ description: 'Describe an image at a given URL or local absolute file path. ' +
140
+ 'Returns a detailed text description suitable as input to answer ' +
141
+ 'factual questions about the image content (charts, graphs, photos, ' +
142
+ 'screenshots, diagrams, text in images). ' +
143
+ `Uses ${VISION_MODEL} for cost-efficient vision (~$0.001/call). ` +
144
+ 'Supported formats: JPEG, PNG, GIF, WebP.',
145
+ input_schema: {
146
+ type: 'object',
147
+ properties: {
148
+ source: {
149
+ type: 'string',
150
+ description: 'URL (http/https) or absolute local file path of the image to describe.',
151
+ },
152
+ prompt: {
153
+ type: 'string',
154
+ description: 'Optional instruction to guide the description ' +
155
+ '(default: describe all visible content with focus on factual details).',
156
+ },
157
+ },
158
+ required: ['source'],
159
+ },
160
+ };
161
+ async execute(input) {
162
+ const source = String(input['source'] ?? '').trim();
163
+ if (!source)
164
+ throw new Error('image_describe: `source` input is required and must be non-empty.');
165
+ const prompt = input['prompt'] != null ? String(input['prompt']).trim() : DEFAULT_PROMPT;
166
+ // Resolve API key — errors here are caught by the agent loop.
167
+ let apiKey;
168
+ try {
169
+ apiKey = resolveAnthropicApiKey(this.apiKey);
170
+ }
171
+ catch (e) {
172
+ return `[image_describe error] ${String(e)}`;
173
+ }
174
+ let imageBlock;
175
+ if (source.startsWith('http://') || source.startsWith('https://')) {
176
+ // URL image — Anthropic fetches it directly.
177
+ imageBlock = {
178
+ type: 'image',
179
+ source: { type: 'url', url: source },
180
+ };
181
+ }
182
+ else {
183
+ // Local file — validate, read, base64-encode.
184
+ if (!path.isAbsolute(source)) {
185
+ return (`[image_describe error] Local file paths must be absolute. ` +
186
+ `Got: "${source}".`);
187
+ }
188
+ let buf;
189
+ try {
190
+ buf = fs.readFileSync(source);
191
+ }
192
+ catch (e) {
193
+ const err = e;
194
+ if (err.code === 'ENOENT') {
195
+ return `[image_describe error] File not found: ${source}`;
196
+ }
197
+ return `[image_describe error] Cannot read file "${source}": ${String(e)}`;
198
+ }
199
+ const mime = detectMime(source, buf);
200
+ imageBlock = {
201
+ type: 'image',
202
+ source: {
203
+ type: 'base64',
204
+ media_type: mime,
205
+ data: buf.toString('base64'),
206
+ },
207
+ };
208
+ }
209
+ try {
210
+ const description = await callVisionApi(imageBlock, prompt, apiKey);
211
+ return `[image_describe: ${VISION_MODEL}]\n${description}`;
212
+ }
213
+ catch (e) {
214
+ // Return error as string — never throw so the agent loop stays alive.
215
+ return `[image_describe error] ${String(e)}`;
216
+ }
217
+ }
218
+ }
219
+ export function createImageDescribeTool(opts) {
220
+ return new ImageDescribeTool(opts?.apiKey);
221
+ }
222
+ //# sourceMappingURL=image_describe.js.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"image_describe.js","sourceRoot":"","sources":["../../../../src/benchmarks/gaia-tools/image_describe.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GA8BG;AAEH,OAAO,KAAK,EAAE,MAAM,SAAS,CAAC;AAC9B,OAAO,KAAK,IAAI,MAAM,WAAW,CAAC;AAClC,OAAO,EAAE,QAAQ,EAAE,MAAM,oBAAoB,CAAC;AAG9C,8EAA8E;AAC9E,YAAY;AACZ,8EAA8E;AAE9E,MAAM,YAAY,GAAG,kBAAkB,CAAC;AACxC,MAAM,qBAAqB,GAAG,YAAY,CAAC;AAC3C,MAAM,iBAAiB,GAAG,uCAAuC,CAAC;AAClE,MAAM,UAAU,GAAG,GAAG,CAAC;AACvB,MAAM,cAAc,GAClB,0EAA0E;IAC1E,wEAAwE;IACxE,iEAAiE,CAAC;AAQpE,MAAM,WAAW,GAAkC;IACjD,MAAM,EAAE,YAAY;IACpB,OAAO,EAAE,YAAY;IACrB,MAAM,EAAE,WAAW;IACnB,MAAM,EAAE,WAAW;IACnB,OAAO,EAAE,YAAY;CACtB,CAAC;AAEF,SAAS,UAAU,CAAC,QAAgB,EAAE,GAAW;IAC/C,oBAAoB;IACpB,IAAI,GAAG,CAAC,MAAM,IAAI,CAAC,IAAI,GAAG,CAAC,CAAC,CAAC,KAAK,IAAI,IAAI,GAAG,CAAC,CAAC,CAAC,KAAK,IAAI;QAAE,OAAO,YAAY,CAAC;IAC/E,IACE,GAAG,CAAC,MAAM,IAAI,CAAC;QACf,GAAG,CAAC,CAAC,CAAC,KAAK,IAAI;QACf,GAAG,CAAC,CAAC,CAAC,KAAK,IAAI;QACf,GAAG,CAAC,CAAC,CAAC,KAAK,IAAI;QACf,GAAG,CAAC,CAAC,CAAC,KAAK,IAAI;QAEf,OAAO,WAAW,CAAC;IACrB,IACE,GAAG,CAAC,MAAM,IAAI,CAAC;QACf,GAAG,CAAC,CAAC,CAAC,KAAK,IAAI;QACf,GAAG,CAAC,CAAC,CAAC,KAAK,IAAI;QACf,GAAG,CAAC,CAAC,CAAC,KAAK,IAAI;QACf,GAAG,CAAC,CAAC,CAAC,KAAK,IAAI;QAEf,OAAO,WAAW,CAAC;IACrB,IACE,GAAG,CAAC,MAAM,IAAI,CAAC;QACf,GAAG,CAAC,CAAC,CAAC,KAAK,IAAI;QACf,GAAG,CAAC,CAAC,CAAC,KAAK,IAAI;QACf,GAAG,CAAC,CAAC,CAAC,KAAK,IAAI;QACf,GAAG,CAAC,CAAC,CAAC,KAAK,IAAI;QAEf,OAAO,YAAY,CAAC;IACtB,yBAAyB;IACzB,MAAM,GAAG,GAAG,IAAI,CAAC,OAAO,CAAC,QAAQ,CAAC,CAAC,WAAW,EAAE,CAAC;IACjD,OAAO,WAAW,CAAC,GAAG,CAAC,IAAI,WAAW,CAAC;AACzC,CAAC;AAED,8EAA8E;AAC9E,6CAA6C;AAC7C,8EAA8E;AAE9E,SAAS,sBAAsB,CAAC,WAAoB;IAClD,IAAI,WAAW,IAAI,WAAW,CAAC,IAAI,EAAE;QAAE,OAAO,WAAW,CAAC,IAAI,EAAE,CAAC;IACjE,MAAM,MAAM,GAAG,OAAO,CAAC,GAAG,CAAC,iBAAiB,CAAC;IAC7C,IAAI,MAAM,IAAI,MAAM,CAAC,IAAI,EAAE;QAAE,OAAO,MAAM,CAAC,IAAI,EAAE,CAAC;IAClD,IAAI,CAAC;QACH,MAAM,GAAG,GAAG,QAAQ,CAClB,8EAA8E,EAC9E,EAAE,QAAQ,EAAE,OAAO,EAAE,OAAO,EAAE,MAAM,EAAE,CACvC,CAAC,IAAI,EAAE,CAAC;QACT,IAAI,GAAG;YAAE,OAAO,GAAG,CAAC;IACtB,CAAC;IAAC,MAAM,CAAC;QACP,kBAAkB;IACpB,CAAC;IACD,MAAM,IAAI,KAAK,CACb,kFAAkF;QAChF,4BAA4B,CAC/B,CAAC;AACJ,CAAC;AAuBD,KAAK,UAAU,aAAa,CAC1B,UAA+B,EAC/B,MAAc,EACd,MAAc;IAEd,MAAM,IAAI,GAAG,IAAI,CAAC,SAAS,CAAC;QAC1B,KAAK,EAAE,YAAY;QACnB,UAAU,EAAE,UAAU;QACtB,QAAQ,EAAE;YACR;gBACE,IAAI,EAAE,MAAM;gBACZ,OAAO,EAAE,CAAC,UAAU,EAAE,EAAE,IAAI,EAAE,MAAM,EAAE,IAAI,EAAE,MAAM,EAAwB,CAAC;aAC5E;SACF;KACF,CAAC,CAAC;IAEH,MAAM,IAAI,GAAG,MAAM,KAAK,CAAC,iBAAiB,EAAE;QAC1C,MAAM,EAAE,MAAM;QACd,OAAO,EAAE;YACP,WAAW,EAAE,MAAM;YACnB,mBAAmB,EAAE,qBAAqB;YAC1C,cAAc,EAAE,kBAAkB;SACnC;QACD,IAAI;KACL,CAAC,CAAC;IAEH,IAAI,CAAC,IAAI,CAAC,EAAE,EAAE,CAAC;QACb,MAAM,OAAO,GAAG,MAAM,IAAI,CAAC,IAAI,EAAE,CAAC,KAAK,CAAC,GAAG,EAAE,CAAC,WAAW,CAAC,CAAC;QAC3D,MAAM,IAAI,KAAK,CAAC,uBAAuB,IAAI,CAAC,MAAM,KAAK,OAAO,EAAE,CAAC,CAAC;IACpE,CAAC;IAED,MAAM,IAAI,GAAG,CAAC,MAAM,IAAI,CAAC,IAAI,EAAE,CAI9B,CAAC;IAEF,IAAI,IAAI,CAAC,KAAK;QAAE,MAAM,IAAI,KAAK,CAAC,wBAAwB,IAAI,CAAC,KAAK,CAAC,OAAO,EAAE,CAAC,CAAC;IAE9E,MAAM,SAAS,GAAG,IAAI,CAAC,OAAO,EAAE,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,KAAK,MAAM,CAAC,CAAC;IAC/D,OAAO,SAAS,EAAE,IAAI,IAAI,2BAA2B,CAAC;AACxD,CAAC;AAED,8EAA8E;AAC9E,0BAA0B;AAC1B,8EAA8E;AAE9E,MAAM,OAAO,iBAAiB;IACnB,IAAI,GAAG,gBAAgB,CAAC;IAChB,MAAM,CAAU;IAEjC,YAAY,MAAe;QACzB,IAAI,CAAC,MAAM,GAAG,MAAM,CAAC;IACvB,CAAC;IAEQ,UAAU,GAAmB;QACpC,IAAI,EAAE,gBAAgB;QACtB,WAAW,EACT,gEAAgE;YAChE,kEAAkE;YAClE,qEAAqE;YACrE,0CAA0C;YAC1C,QAAQ,YAAY,6CAA6C;YACjE,0CAA0C;QAC5C,YAAY,EAAE;YACZ,IAAI,EAAE,QAAQ;YACd,UAAU,EAAE;gBACV,MAAM,EAAE;oBACN,IAAI,EAAE,QAAQ;oBACd,WAAW,EACT,wEAAwE;iBAC3E;gBACD,MAAM,EAAE;oBACN,IAAI,EAAE,QAAQ;oBACd,WAAW,EACT,gDAAgD;wBAChD,wEAAwE;iBAC3E;aACF;YACD,QAAQ,EAAE,CAAC,QAAQ,CAAC;SACrB;KACF,CAAC;IAEF,KAAK,CAAC,OAAO,CAAC,KAA8B;QAC1C,MAAM,MAAM,GAAG,MAAM,CAAC,KAAK,CAAC,QAAQ,CAAC,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC;QACpD,IAAI,CAAC,MAAM;YAAE,MAAM,IAAI,KAAK,CAAC,mEAAmE,CAAC,CAAC;QAElG,MAAM,MAAM,GAAG,KAAK,CAAC,QAAQ,CAAC,IAAI,IAAI,CAAC,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,QAAQ,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,CAAC,cAAc,CAAC;QAEzF,8DAA8D;QAC9D,IAAI,MAAc,CAAC;QACnB,IAAI,CAAC;YACH,MAAM,GAAG,sBAAsB,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;QAC/C,CAAC;QAAC,OAAO,CAAU,EAAE,CAAC;YACpB,OAAO,0BAA0B,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC;QAC/C,CAAC;QAED,IAAI,UAA+B,CAAC;QAEpC,IAAI,MAAM,CAAC,UAAU,CAAC,SAAS,CAAC,IAAI,MAAM,CAAC,UAAU,CAAC,UAAU,CAAC,EAAE,CAAC;YAClE,6CAA6C;YAC7C,UAAU,GAAG;gBACX,IAAI,EAAE,OAAO;gBACb,MAAM,EAAE,EAAE,IAAI,EAAE,KAAK,EAAE,GAAG,EAAE,MAAM,EAAE;aACrC,CAAC;QACJ,CAAC;aAAM,CAAC;YACN,8CAA8C;YAC9C,IAAI,CAAC,IAAI,CAAC,UAAU,CAAC,MAAM,CAAC,EAAE,CAAC;gBAC7B,OAAO,CACL,4DAA4D;oBAC5D,SAAS,MAAM,IAAI,CACpB,CAAC;YACJ,CAAC;YAED,IAAI,GAAW,CAAC;YAChB,IAAI,CAAC;gBACH,GAAG,GAAG,EAAE,CAAC,YAAY,CAAC,MAAM,CAAC,CAAC;YAChC,CAAC;YAAC,OAAO,CAAU,EAAE,CAAC;gBACpB,MAAM,GAAG,GAAG,CAA0B,CAAC;gBACvC,IAAI,GAAG,CAAC,IAAI,KAAK,QAAQ,EAAE,CAAC;oBAC1B,OAAO,0CAA0C,MAAM,EAAE,CAAC;gBAC5D,CAAC;gBACD,OAAO,4CAA4C,MAAM,MAAM,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC;YAC7E,CAAC;YAED,MAAM,IAAI,GAAG,UAAU,CAAC,MAAM,EAAE,GAAG,CAAC,CAAC;YACrC,UAAU,GAAG;gBACX,IAAI,EAAE,OAAO;gBACb,MAAM,EAAE;oBACN,IAAI,EAAE,QAAQ;oBACd,UAAU,EAAE,IAAI;oBAChB,IAAI,EAAE,GAAG,CAAC,QAAQ,CAAC,QAAQ,CAAC;iBAC7B;aACF,CAAC;QACJ,CAAC;QAED,IAAI,CAAC;YACH,MAAM,WAAW,GAAG,MAAM,aAAa,CAAC,UAAU,EAAE,MAAM,EAAE,MAAM,CAAC,CAAC;YACpE,OAAO,oBAAoB,YAAY,MAAM,WAAW,EAAE,CAAC;QAC7D,CAAC;QAAC,OAAO,CAAU,EAAE,CAAC;YACpB,sEAAsE;YACtE,OAAO,0BAA0B,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC;QAC/C,CAAC;IACH,CAAC;CACF;AAWD,MAAM,UAAU,uBAAuB,CAAC,IAA+B;IACrE,OAAO,IAAI,iBAAiB,CAAC,IAAI,EAAE,MAAM,CAAC,CAAC;AAC7C,CAAC"}
@@ -0,0 +1,24 @@
1
+ /**
2
+ * Smoke tests for image_describe — ADR-133-PR5
3
+ *
4
+ * Run with:
5
+ * cd v3/@claude-flow/cli
6
+ * npx tsx src/benchmarks/gaia-tools/image_describe.smoke.ts
7
+ *
8
+ * Tests:
9
+ * 1. URL image (Wikipedia logo PNG) — description contains 'wiki' or 'logo' or 'W'
10
+ * 2. Missing API key — returns error string, does not crash
11
+ * 3. Non-existent local file — returns error string, does not crash
12
+ * 4. Invalid source (empty) — throws (caught by test harness)
13
+ *
14
+ * If ANTHROPIC_API_KEY is not set and gcloud is not available, Test 1 is
15
+ * marked SKIP (not FAIL) — the tool's graceful error path is validated in
16
+ * Test 2 instead.
17
+ *
18
+ * Cost estimate: ~$0.001 per live API call (Haiku vision, single small image).
19
+ * Total smoke cost: ≤$0.001 (only Test 1 makes a live call).
20
+ *
21
+ * Refs: ADR-133, #2156
22
+ */
23
+ export {};
24
+ //# sourceMappingURL=image_describe.smoke.d.ts.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"image_describe.smoke.d.ts","sourceRoot":"","sources":["../../../../src/benchmarks/gaia-tools/image_describe.smoke.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;GAqBG"}