@tangle-network/agent-eval 0.140.1 → 0.141.0

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 (59) hide show
  1. package/CHANGELOG.md +15 -0
  2. package/dist/analyst/index.d.ts +74 -6
  3. package/dist/analyst/index.d.ts.map +1 -1
  4. package/dist/analyst/index.js +6 -6
  5. package/dist/{benchmark-command-Bo_hkhQf.js → benchmark-command-D8xXNVVh.js} +2239 -1820
  6. package/dist/benchmark-command-D8xXNVVh.js.map +1 -0
  7. package/dist/benchmarks/index.js +1 -1
  8. package/dist/{benchmarks-HwoBE32G.js → benchmarks-Ce8GdJbc.js} +4 -4
  9. package/dist/{benchmarks-HwoBE32G.js.map → benchmarks-Ce8GdJbc.js.map} +1 -1
  10. package/dist/campaign/index.js +3 -3
  11. package/dist/{campaign-BzjYNYVZ.js → campaign-COFTVy3Q.js} +3 -3
  12. package/dist/{campaign-BzjYNYVZ.js.map → campaign-COFTVy3Q.js.map} +1 -1
  13. package/dist/cli.js +1 -1
  14. package/dist/contract/index.d.ts.map +1 -1
  15. package/dist/contract/index.js +56 -13
  16. package/dist/contract/index.js.map +1 -1
  17. package/dist/{default-registry-BOiBXNFh.js → default-registry-BupFSXSN.js} +108 -20
  18. package/dist/default-registry-BupFSXSN.js.map +1 -0
  19. package/dist/default-registry-Ci7wAAR8.d.ts.map +1 -1
  20. package/dist/{dspy-rlm-engine-Bkak4nzo.js → dspy-rlm-engine-Bw8Zz9CX.js} +7 -4
  21. package/dist/{dspy-rlm-engine-Bkak4nzo.js.map → dspy-rlm-engine-Bw8Zz9CX.js.map} +1 -1
  22. package/dist/{extract-usage-C5vMw-0R.js → extract-usage-7l1Xq5ti.js} +2 -2
  23. package/dist/{extract-usage-C5vMw-0R.js.map → extract-usage-7l1Xq5ti.js.map} +1 -1
  24. package/dist/index.d.ts +68 -31
  25. package/dist/index.d.ts.map +1 -1
  26. package/dist/index.js +122 -46
  27. package/dist/index.js.map +1 -1
  28. package/dist/{kind-factory-DB7nIs35.js → kind-factory-Bvwe3pup.js} +11 -1
  29. package/dist/kind-factory-Bvwe3pup.js.map +1 -0
  30. package/dist/multishot/index.d.ts +29 -6
  31. package/dist/multishot/index.d.ts.map +1 -1
  32. package/dist/multishot/index.js +53 -3
  33. package/dist/multishot/index.js.map +1 -1
  34. package/dist/openapi.json +1 -1
  35. package/dist/replay-BqTgoioO.d.ts.map +1 -1
  36. package/dist/{replay-k2MsOmv5.js → replay-DjUfTrHD.js} +40 -48
  37. package/dist/replay-DjUfTrHD.js.map +1 -0
  38. package/dist/{single-run-lock-CiQThJxB.js → single-run-lock-BcvPQY57.js} +65 -15
  39. package/dist/single-run-lock-BcvPQY57.js.map +1 -0
  40. package/dist/{skill-usage-CVVnoIx-.d.ts → skill-usage-CNZTk_LT.d.ts} +8 -4
  41. package/dist/{skill-usage-CVVnoIx-.d.ts.map → skill-usage-CNZTk_LT.d.ts.map} +1 -1
  42. package/dist/{skillopt-optimization-method-CSBQ8Qma.js → skillopt-optimization-method-DV6c4ooy.js} +2 -2
  43. package/dist/{skillopt-optimization-method-CSBQ8Qma.js.map → skillopt-optimization-method-DV6c4ooy.js.map} +1 -1
  44. package/dist/{store-otlp-vRByAR6h.js → store-otlp-D4I90_vR.js} +2 -2
  45. package/dist/{store-otlp-vRByAR6h.js.map → store-otlp-D4I90_vR.js.map} +1 -1
  46. package/dist/traces.js +4 -4
  47. package/docs/adapters-observability.md +35 -21
  48. package/docs/concepts.md +1 -0
  49. package/docs/distributed-driver.md +7 -3
  50. package/docs/feature-guide.md +1 -0
  51. package/docs/insight-report.md +1 -1
  52. package/docs/trace-analysis.md +11 -0
  53. package/docs/wire-protocol.md +70 -2
  54. package/package.json +5 -3
  55. package/dist/benchmark-command-Bo_hkhQf.js.map +0 -1
  56. package/dist/default-registry-BOiBXNFh.js.map +0 -1
  57. package/dist/kind-factory-DB7nIs35.js.map +0 -1
  58. package/dist/replay-k2MsOmv5.js.map +0 -1
  59. package/dist/single-run-lock-CiQThJxB.js.map +0 -1
package/dist/cli.js CHANGED
@@ -1,6 +1,6 @@
1
1
  #!/usr/bin/env node
2
2
  import { o as runRolloutReleaseCli } from "./hf-dataset-XggBupCr.js";
3
- import { n as runAnalystBenchmarkCommand } from "./benchmark-command-Bo_hkhQf.js";
3
+ import { n as runAnalystBenchmarkCommand } from "./benchmark-command-D8xXNVVh.js";
4
4
  import { a as runRpcBatch, o as runRpcOnce, p as handleVersion, r as startServerAsync, s as buildOpenApi } from "./server-Cu4M3NSO.js";
5
5
  import { writeFileSync } from "node:fs";
6
6
  //#region src/cli-config.ts
@@ -1 +1 @@
1
- {"version":3,"file":"index.d.ts","names":[],"sources":["../../src/contract/self-improve.ts","../../src/contract/define-agent-eval.ts","../../src/contract/measured-comparison.ts","../../src/contract/profile-measured-comparison.ts","../../src/contract/intake/run-record-dir.ts","../../src/contract/eval-reporting-suite.ts","../../src/contract/diff.ts","../../src/contract/intake/agent-trace.ts","../../src/contract/intake/code-agent-observation.ts","../../src/contract/intake/code-agent-session.ts","../../src/contract/intake/feedback-table.ts","../../src/contract/intake/otel-spans.ts"],"mappings":";;;;;;;;;;;;;UA+DiB;;;EAGf;;;;EAIA;;EAEA;;EAEA;;;EAGA;;;EAGA;;;;EAIA;;EAEA,mBAAmB;;;;;;;;;EASnB;;EAEA;;;;;EAKA;;KAGU;EACN;EAA0B;;EAC1B;EAA4B;EAAuB;;EACnD;EAA4B;EAAe;;EAC3C;EAA8B;EAAe;EAAuB;;EAGpE;EAAsB;EAAkB;;EACxC;EAAyB;EAAW;EAAY;EAAa;;UAElD,mBAAmB,kBAAkB,UAAU;;;;;;;;;;;;;;EAc9D,QAAQ,SAAS,gBAAgB,UAAU,WAAW,KAAK,oBAAoB,QAAQ;;;;;;;EAQvF;;;EAIA,WAAW;;;EAIX,OAAO,YAAY,WAAW;;;EAI9B,iBAAiB;;EAGjB,SAAS;;;;;;;;;;;;EAaT,sBAAsB,gCAAgC,WAAW;;;;;EAMjE,WAAW,gBAAgB;;;;;;EAO3B,SAAS,mBAAmB,WAAW;;;EAIvC,qBAAqB;;;EAIrB,OAAO,KAAK,WAAW;;;;;;;;EASvB,cAAc,eAAe,gBAAgB,iBAAiB,mBAAmB;;;;EAKjF,UAAU;;;;EAKV;;;EAIA,gBAAgB,QAAQ;;;;;EAMxB,iBAAiB;IACf,UAAU;IACV;IACA;;;;EAKF;;;EAIA,cAAc,OAAO;;;;EAKrB;EACA;EACA;;;;;;;;;;;;;;EAeA,eAAe;;;EAIf,eAAe;;;;EAKf,eAAe;;EAGf;;;;;;;;EASA;;;;;;;;;EAUA,oBAAoB,uBAAuB,WAAW;;;EAItD,WAAW;;;;;;;EAQX,mBAAmB,uBAAuB,WAAW;;UAGtC,kBAAkB,kBAAkB,UAAU;;;;EAI7D;IACE;IACA,aAAa;;;;;EAKf;IACE;IACA,aAAa;IACb,SAAS;;;IAGT;;;;IAIA;;;;;;EAMF;;;EAGA;;;;EAIA,YAAY;;EAEZ;;;EAGA;;EAEA;;EAEA;;EAEA,MAAM;;;EAGN,UAAU;;EAEV;IACE;IACA,MAAM;IACN;IACA,aAAa;;;;;;;;EAQf,SAAS;;;;;EAKT,QAAQ;;;;;;EAMR,KAAK,yBAAyB,WAAW;;;cAI9B,4BAA4B;WAC9B,MAAM;WACN,UAAU;EAEnB,YAAY,gBAAgB,QAAQ;;;;;;;;;;;;;;;;;;;;;;;;;;;;iBA+LhB,YAAY,kBAAkB,UAAU,WAC5D,MAAM,mBAAmB,WAAW,aACnC,QAAQ,kBAAkB,WAAW;;;KCviB5B,eAAe,kBAAkB,UAAU,cACrD,SAAS,gBACT,UAAU,WACV,KAAK,oBACF,QAAQ;KAED,uBAAuB,kBAAkB,UAAU,aAAa,mBAC1E,WACA;UAGe,yBAAyB,kBAAkB,UAAU,mBAC5D,KACN,eAAe,WAAW;;EAI5B,YAAY;;EAEZ,UAAU;;EAEV,QAAQ,eAAe,WAAW;;EAElC,QAAQ,YAAY,WAAW;;EAE/B,SAAS,YAAY,WAAW;;EAEhC;;KAGU,wBAAwB,kBAAkB,UAAU,aAAa,KAC3E,QAAQ,mBAAmB,WAAW;EAGtC,SAAS,QAAQ;EACjB,eAAe,QAAQ;;UAGR,iBAAiB,kBAAkB,UAAU;;WAEnD,oBAAoB;;WAEpB,iBAAiB;;;;;EAK1B,SACE,OAAO,yBAAyB,WAAW,aAC1C,QAAQ,eAAe,WAAW;;;;;;;EAOrC,QACE,OAAO,wBAAwB,WAAW,aACzC,QAAQ,kBAAkB,WAAW;;;;;;;;;iBAU1B,gBAAgB,kBAAkB,UAAU,WAC1D,UAAU,uBAAuB,WAAW,aAC3C,iBAAiB,WAAW;;;UCvDd;EACf,QAAQ,gCAAgC;EACxC;EACA;;UAGe;EACf,YAAY;EACZ;EACA,QAAQ;EACR,MAAM;EACN,eAAe;EACf;EACA,SAAS;;UAGM;EACf,YAAY;EACZ,QAAQ,OAAO,oCAAoC,QAAQ;;EAE3D;;EAEA,aAAa;EACb,SAAS;;UAGM;EACf,cAAc;EACd;IACE;IACA,MAAM;;;UAIO;EACf,YAAY;EACZ,cAAc;EACd;IACE;IACA,MAAM;;EAER,aAAa;EACb;EACA,YAAY;EACZ;EACA,WAAW;;;UAII,kBAAkB;EACjC;EACA,UAAU;EACV,WAAW;;;UAII,yBAAyB;EACxC,MAAM,KAAK;EACX,WAAW,KAAK;IAAkB;IAAc;;EAChD,QAAQ,KAAK;EACb,eAAe,KAAK,OAAO;EAC3B,UAAU,KAAK;EACf,UAAU,KAAK;EACf,OAAO,KAAK;;UAGG,kCAAkC;EACjD,uBAAuB,kBAAkB;EACzC,QAAQ;EACR,SAAS,yBAAyB;;EAElC;;EAEA,kBAAkB;;EAElB,kBAAkB;;;KAIR,8BAA8B,KACxC;EAGA,iBAAiB;EACjB,WAAW;EACX;;;iBAIc,2BACd,UAAU,sCACT;;iBAQa,4BACd,SAAS,qCACR;;iBAiBa,wBACd,UAAU,mCACT;iBAQa,0BAA0B,iBAAiB;;iBAarC,uBACpB,SAAS,gCACR,QAAQ;;;;;;;;iBAkEK,2BAA2B,MACzC,SAAS,kCAAkC,QAC1C;;iBAwRa,0CACd,SAAS,oCACR;;iBAuEa,oCACd,iBACC;iBAkHa,6BAA6B,iBAAiB;iBAM9C,oCACd,iBACC;iBAkBa,8BAA8B;;;;;;;;;;UCvsB7B;EACf,aAAa;EACb,QAAQ,gCAAgC;EACxC;EACA;;UAGe;EACf,YAAY;EACZ;EACA,aAAa;EACb,MAAM;EACN,SAAS;EACT;EACA,SAAS;;UAGM;EACf,YAAY;EACZ,QACE,OAAO,kDACN,QAAQ;;EAEX;;EAEA,aAAa;EACb,SAAS;;UAGM;EACf,cAAc;EACd;IACE;IACA,MAAM;;;UAIO;EACf,YAAY;EACZ,cAAc;EACd;IACE;IACA,MAAM;;EAER,aAAa;EACb;EACA,YAAY;EACZ;EACA,WAAW;;;iBAIG,gCACd,UAAU,sCACT;;iBAQa,iCACd,SAAS,0CACR;;iBAqBa,sCACd,UAAU,4CACT;iBAOa,kCAAkC,iBAAiB;iBAInD,yCACd,iBACC;iBAIa,wCACd,iBACC;;;;;;iBASmB,qCACpB,SAAS,8CACR,QAAQ;;iBA0CK,wDACd,SAAS,kDACR;;iBAuEa,kDACd,iBACC;;;;UCnPc;;EAEf;;EAEA;;EAEA;;UAGe;;;;;;EAMf;;;;;;;EAOA,WAAW;;;;;;EAMX;;UAGe;;EAEf,MAAM;;EAEN,UAAU;;EAEV;;;;;;;;;;iBAkBoB,iBACpB,cACA,UAAS,0BACR,QAAQ;;;;;KC7CC,0BAA0B;UAErB;;;;;EAKf,UAAU,KAAK;;EAEf,OAAO;;;;;;;;;;EAUP;;;;UAKe;;;EAGf,QAAQ;;EAER;;IAEE;;IAEA;;;IAGA;;IAEA;;;IAGA,UAAU;;;EAGZ;;;;;;;iBAUoB,mBACpB,OAAO,yBACP,UAAS,4BACR,QAAQ;;;;;;UC5DM;EACf;EACA;EACA;;;UAIe;EACf;EACA;EACA;EACA;EACA;;;EAGA,YAAY,eAAe,eAAe;;;;UAK3B;EACf;EACA;EACA;EACA;EACA;;EAEA,SAAS;;EAET,SAAS;;EAET,OAAO;;EAEP;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;;;;;UAMe;EACf;EACA;EACA;EACA;EACA,oBAAoB;EACpB,mBAAmB;EACnB;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;;EAEA,cAAc;;;;EAId,aAAa;;;;;;;;;iBAoDC,gBACd,QAAQ,2BACR,OAAO,4BACN;;;;;;iBAqEa,SAAS,QAAQ,cAAc,OAAO,eAAe;;;;;;;iBAsCrD,wBAAwB,KAAK,eAAe;;;KC1OhD;UAEK;EACf,MAAM;;EAEN;;UAGe;EACf;EACA;EACA;;;EAGA,cAAc;;UAGC;EACf;EACA,cAAc;EACd,QAAQ;;UAGO;EACf;EACA,eAAe;;UAGA;EACf;EACA;EACA;EACA;IAAQ;IAAc;;EACtB;IAAS;IAAe;;EACxB,OAAO;;;;UAOQ;EACf;;EAEA;;EAEA;EACA;EACA;;EAEA;;EAEA;;KAGU,kBAAkB,YAAY;;;;;;iBAW1B,gBAAgB,SAAS,qBAAqB;UAmE7C;;;;EAIf,SAAS,YAAY;;;EAGrB,cAAc;;;;;;;;iBASA,8BACd,MAAM,aACN,OAAO,kBACN;;;KCjLS;KAEA;KAEA;KAEA;KASA;UAEK;EACf;EACA;EACA;;UAGe;EACf;EACA;EACA,MAAM;EACN,SAAS;EACT;EACA,QAAQ;EACR;EACA;EACA,UAAU;;UAGK;EACf,QAAQ;EACR;EACA;EACA;IACE,QAAQ;IACR;;EAEF,SAAS;;UAGM;EACf,QAAQ;EACR;EACA;EACA,YAAY;;;;;;;iBAeE,wBACd,SAAS,iCACR;;;UCrCc;EACf;EACA;;UAGe;EACf;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;;UAGe;EACf,QAAQ;EACR;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA,WAAW;EACX;;UAGe;EACf,MAAM;EACN,aAAa;EACb,SAAS;EACT,cAAc;;UAGC;EACf;EACA;EACA;EACA;EACA;EACA;EACA,WAAW;EACX;EACA;EACA;EACA;EACA;EACA;;;EAGA,iBAAiB;;;EAGjB,YAAY;;iBAGE,oBAAoB,gBAAgB;iBAepC,iBACd,SAAS,gCACR;iBAIa,sBACd,SAAS,gCACR;iBAIa,oBACd,SAAS,gCACR;iBAIa,oBACd,SAAS,gCACR;iBAIa,cACd,SAAS,gCACR;cAIU,2BAAkB;;;UCpJd;;;EAGf;;EAEA;;;EAGA;;;EAGA,WAAW;;UAGI;EACf;;;EAGA;;EAEA;;EAEA;;EAEA;;EAEA;;EAEA;;EAEA;;EAEA;;EAEA;;;EAGA,WAAW;;;EAGX,SAAS;;UAGM;;EAEf,SAAS;;;EAGT,OAAO;;;;EAIP;IAAU;IAAa;;;;;EAIvB;;UAGe;EACf,MAAM;;;EAGN,aAAa;IAAQ;IAAe;IAAe;;;iBAGrC,kBAAkB,MAAM,2BAA2B;;;UCvBlD;EACf,OAAO;;EAEP,eAAe;;EAEf;;;;;;EAMA,eAAe,eAAe,gBAAgB;;iBAGhC,cAAc,MAAM,uBAAuB"}
1
+ {"version":3,"file":"index.d.ts","names":[],"sources":["../../src/contract/self-improve.ts","../../src/contract/define-agent-eval.ts","../../src/contract/measured-comparison.ts","../../src/contract/profile-measured-comparison.ts","../../src/contract/intake/run-record-dir.ts","../../src/contract/eval-reporting-suite.ts","../../src/contract/diff.ts","../../src/contract/intake/agent-trace.ts","../../src/contract/intake/code-agent-observation.ts","../../src/contract/intake/code-agent-session.ts","../../src/contract/intake/feedback-table.ts","../../src/contract/intake/otel-spans.ts"],"mappings":";;;;;;;;;;;;;UA+DiB;;;EAGf;;;;EAIA;;EAEA;;EAEA;;;EAGA;;;EAGA;;;;EAIA;;EAEA,mBAAmB;;;;;;;;;EASnB;;EAEA;;;;;EAKA;;KAGU;EACN;EAA0B;;EAC1B;EAA4B;EAAuB;;EACnD;EAA4B;EAAe;;EAC3C;EAA8B;EAAe;EAAuB;;EAGpE;EAAsB;EAAkB;;EACxC;EAAyB;EAAW;EAAY;EAAa;;UAElD,mBAAmB,kBAAkB,UAAU;;;;;;;;;;;;;;EAc9D,QAAQ,SAAS,gBAAgB,UAAU,WAAW,KAAK,oBAAoB,QAAQ;;;;;;;EAQvF;;;EAIA,WAAW;;;EAIX,OAAO,YAAY,WAAW;;;EAI9B,iBAAiB;;EAGjB,SAAS;;;;;;;;;;;;EAaT,sBAAsB,gCAAgC,WAAW;;;;;EAMjE,WAAW,gBAAgB;;;;;;EAO3B,SAAS,mBAAmB,WAAW;;;EAIvC,qBAAqB;;;EAIrB,OAAO,KAAK,WAAW;;;;;;;;EASvB,cAAc,eAAe,gBAAgB,iBAAiB,mBAAmB;;;;EAKjF,UAAU;;;;EAKV;;;EAIA,gBAAgB,QAAQ;;;;;EAMxB,iBAAiB;IACf,UAAU;IACV;IACA;;;;EAKF;;;EAIA,cAAc,OAAO;;;;EAKrB;EACA;EACA;;;;;;;;;;;;;;EAeA,eAAe;;;EAIf,eAAe;;;;EAKf,eAAe;;EAGf;;;;;;;;EASA;;;;;;;;;EAUA,oBAAoB,uBAAuB,WAAW;;;EAItD,WAAW;;;;;;;EAQX,mBAAmB,uBAAuB,WAAW;;UAGtC,kBAAkB,kBAAkB,UAAU;;;;EAI7D;IACE;IACA,aAAa;;;;;EAKf;IACE;IACA,aAAa;IACb,SAAS;;;IAGT;;;;IAIA;;;;;;EAMF;;;EAGA;;;;EAIA,YAAY;;EAEZ;;;EAGA;;EAEA;;EAEA;;EAEA,MAAM;;;EAGN,UAAU;;EAEV;IACE;IACA,MAAM;IACN;IACA,aAAa;;;;;;;;EAQf,SAAS;;;;;EAKT,QAAQ;;;;;;EAMR,KAAK,yBAAyB,WAAW;;;cAI9B,4BAA4B;WAC9B,MAAM;WACN,UAAU;EAEnB,YAAY,gBAAgB,QAAQ;;;;;;;;;;;;;;;;;;;;;;;;;;;;iBA+LhB,YAAY,kBAAkB,UAAU,WAC5D,MAAM,mBAAmB,WAAW,aACnC,QAAQ,kBAAkB,WAAW;;;KCviB5B,eAAe,kBAAkB,UAAU,cACrD,SAAS,gBACT,UAAU,WACV,KAAK,oBACF,QAAQ;KAED,uBAAuB,kBAAkB,UAAU,aAAa,mBAC1E,WACA;UAGe,yBAAyB,kBAAkB,UAAU,mBAC5D,KACN,eAAe,WAAW;;EAI5B,YAAY;;EAEZ,UAAU;;EAEV,QAAQ,eAAe,WAAW;;EAElC,QAAQ,YAAY,WAAW;;EAE/B,SAAS,YAAY,WAAW;;EAEhC;;KAGU,wBAAwB,kBAAkB,UAAU,aAAa,KAC3E,QAAQ,mBAAmB,WAAW;EAGtC,SAAS,QAAQ;EACjB,eAAe,QAAQ;;UAGR,iBAAiB,kBAAkB,UAAU;;WAEnD,oBAAoB;;WAEpB,iBAAiB;;;;;EAK1B,SACE,OAAO,yBAAyB,WAAW,aAC1C,QAAQ,eAAe,WAAW;;;;;;;EAOrC,QACE,OAAO,wBAAwB,WAAW,aACzC,QAAQ,kBAAkB,WAAW;;;;;;;;;iBAU1B,gBAAgB,kBAAkB,UAAU,WAC1D,UAAU,uBAAuB,WAAW,aAC3C,iBAAiB,WAAW;;;UCvDd;EACf,QAAQ,gCAAgC;EACxC;EACA;;UAGe;EACf,YAAY;EACZ;EACA,QAAQ;EACR,MAAM;EACN,eAAe;EACf;EACA,SAAS;;UAGM;EACf,YAAY;EACZ,QAAQ,OAAO,oCAAoC,QAAQ;;EAE3D;;EAEA,aAAa;EACb,SAAS;;UAGM;EACf,cAAc;EACd;IACE;IACA,MAAM;;;UAIO;EACf,YAAY;EACZ,cAAc;EACd;IACE;IACA,MAAM;;EAER,aAAa;EACb;EACA,YAAY;EACZ;EACA,WAAW;;;UAII,kBAAkB;EACjC;EACA,UAAU;EACV,WAAW;;;UAII,yBAAyB;EACxC,MAAM,KAAK;EACX,WAAW,KAAK;IAAkB;IAAc;;EAChD,QAAQ,KAAK;EACb,eAAe,KAAK,OAAO;EAC3B,UAAU,KAAK;EACf,UAAU,KAAK;EACf,OAAO,KAAK;;UAGG,kCAAkC;EACjD,uBAAuB,kBAAkB;EACzC,QAAQ;EACR,SAAS,yBAAyB;;EAElC;;EAEA,kBAAkB;;EAElB,kBAAkB;;;KAIR,8BAA8B,KACxC;EAGA,iBAAiB;EACjB,WAAW;EACX;;;iBAIc,2BACd,UAAU,sCACT;;iBAQa,4BACd,SAAS,qCACR;;iBAiBa,wBACd,UAAU,mCACT;iBAQa,0BAA0B,iBAAiB;;iBAarC,uBACpB,SAAS,gCACR,QAAQ;;;;;;;;iBAkEK,2BAA2B,MACzC,SAAS,kCAAkC,QAC1C;;iBA8Ya,0CACd,SAAS,oCACR;;iBAuEa,oCACd,iBACC;iBAkHa,6BAA6B,iBAAiB;iBAM9C,oCACd,iBACC;iBAkBa,8BAA8B;;;;;;;;;;UC7zB7B;EACf,aAAa;EACb,QAAQ,gCAAgC;EACxC;EACA;;UAGe;EACf,YAAY;EACZ;EACA,aAAa;EACb,MAAM;EACN,SAAS;EACT;EACA,SAAS;;UAGM;EACf,YAAY;EACZ,QACE,OAAO,kDACN,QAAQ;;EAEX;;EAEA,aAAa;EACb,SAAS;;UAGM;EACf,cAAc;EACd;IACE;IACA,MAAM;;;UAIO;EACf,YAAY;EACZ,cAAc;EACd;IACE;IACA,MAAM;;EAER,aAAa;EACb;EACA,YAAY;EACZ;EACA,WAAW;;;iBAIG,gCACd,UAAU,sCACT;;iBAQa,iCACd,SAAS,0CACR;;iBAqBa,sCACd,UAAU,4CACT;iBAOa,kCAAkC,iBAAiB;iBAInD,yCACd,iBACC;iBAIa,wCACd,iBACC;;;;;;iBASmB,qCACpB,SAAS,8CACR,QAAQ;;iBA0CK,wDACd,SAAS,kDACR;;iBAuEa,kDACd,iBACC;;;;UCnPc;;EAEf;;EAEA;;EAEA;;UAGe;;;;;;EAMf;;;;;;;EAOA,WAAW;;;;;;EAMX;;UAGe;;EAEf,MAAM;;EAEN,UAAU;;EAEV;;;;;;;;;;iBAkBoB,iBACpB,cACA,UAAS,0BACR,QAAQ;;;;;KC7CC,0BAA0B;UAErB;;;;;EAKf,UAAU,KAAK;;EAEf,OAAO;;;;;;;;;;EAUP;;;;UAKe;;;EAGf,QAAQ;;EAER;;IAEE;;IAEA;;;IAGA;;IAEA;;;IAGA,UAAU;;;EAGZ;;;;;;;iBAUoB,mBACpB,OAAO,yBACP,UAAS,4BACR,QAAQ;;;;;;UC5DM;EACf;EACA;EACA;;;UAIe;EACf;EACA;EACA;EACA;EACA;;;EAGA,YAAY,eAAe,eAAe;;;;UAK3B;EACf;EACA;EACA;EACA;EACA;;EAEA,SAAS;;EAET,SAAS;;EAET,OAAO;;EAEP;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;;;;;UAMe;EACf;EACA;EACA;EACA;EACA,oBAAoB;EACpB,mBAAmB;EACnB;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;;EAEA,cAAc;;;;EAId,aAAa;;;;;;;;;iBAoDC,gBACd,QAAQ,2BACR,OAAO,4BACN;;;;;;iBAqEa,SAAS,QAAQ,cAAc,OAAO,eAAe;;;;;;;iBAsCrD,wBAAwB,KAAK,eAAe;;;KC1OhD;UAEK;EACf,MAAM;;EAEN;;UAGe;EACf;EACA;EACA;;;EAGA,cAAc;;UAGC;EACf;EACA,cAAc;EACd,QAAQ;;UAGO;EACf;EACA,eAAe;;UAGA;EACf;EACA;EACA;EACA;IAAQ;IAAc;;EACtB;IAAS;IAAe;;EACxB,OAAO;;;;UAOQ;EACf;;EAEA;;EAEA;EACA;EACA;;EAEA;;EAEA;;KAGU,kBAAkB,YAAY;;;;;;iBAW1B,gBAAgB,SAAS,qBAAqB;UAmE7C;;;;EAIf,SAAS,YAAY;;;EAGrB,cAAc;;;;;;;;iBASA,8BACd,MAAM,aACN,OAAO,kBACN;;;KCjLS;KAEA;KAEA;KAEA;KASA;UAEK;EACf;EACA;EACA;;UAGe;EACf;EACA;EACA,MAAM;EACN,SAAS;EACT;EACA,QAAQ;EACR;EACA;EACA,UAAU;;UAGK;EACf,QAAQ;EACR;EACA;EACA;IACE,QAAQ;IACR;;EAEF,SAAS;;UAGM;EACf,QAAQ;EACR;EACA;EACA,YAAY;;;;;;;iBAeE,wBACd,SAAS,iCACR;;;UCrCc;EACf;EACA;;UAGe;EACf;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;;UAGe;EACf,QAAQ;EACR;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA,WAAW;EACX;;UAGe;EACf,MAAM;EACN,aAAa;EACb,SAAS;EACT,cAAc;;UAGC;EACf;EACA;EACA;EACA;EACA;EACA;EACA,WAAW;EACX;EACA;EACA;EACA;EACA;EACA;;;EAGA,iBAAiB;;;EAGjB,YAAY;;iBAGE,oBAAoB,gBAAgB;iBAepC,iBACd,SAAS,gCACR;iBAIa,sBACd,SAAS,gCACR;iBAIa,oBACd,SAAS,gCACR;iBAIa,oBACd,SAAS,gCACR;iBAIa,cACd,SAAS,gCACR;cAIU,2BAAkB;;;UCpJd;;;EAGf;;EAEA;;;EAGA;;;EAGA,WAAW;;UAGI;EACf;;;EAGA;;EAEA;;EAEA;;EAEA;;EAEA;;EAEA;;EAEA;;EAEA;;EAEA;;;EAGA,WAAW;;;EAGX,SAAS;;UAGM;;EAEf,SAAS;;;EAGT,OAAO;;;;EAIP;IAAU;IAAa;;;;;EAIvB;;UAGe;EACf,MAAM;;;EAGN,aAAa;IAAQ;IAAe;IAAe;;;iBAGrC,kBAAkB,MAAM,2BAA2B;;;UCvBlD;EACf,OAAO;;EAEP,eAAe;;EAEf;;;;;;EAMA,eAAe,eAAe,gBAAgB;;iBAGhC,cAAc,MAAM,uBAAuB"}
@@ -1,16 +1,16 @@
1
1
  import { c as ValidationError } from "../errors-D-LKuDhb.js";
2
2
  import { i as estimateCost, o as isModelPriced } from "../metrics-C9YY1OcL.js";
3
3
  import { i as CostLedger } from "../cost-ledger-DMFxsLKr.js";
4
- import { C as createChatClient, t as buildDefaultAnalystRegistry } from "../default-registry-BOiBXNFh.js";
4
+ import { C as createChatClient, t as buildDefaultAnalystRegistry } from "../default-registry-BupFSXSN.js";
5
5
  import { LLM_MODEL_ATTR_KEYS, SPAN_KIND_ATTR_KEYS } from "../trace-attributes.js";
6
- import { nt as classifyOtlpSpanRole, rt as isOtlpModelCall } from "../kind-factory-DB7nIs35.js";
6
+ import { nt as classifyOtlpSpanRole, rt as isOtlpModelCall } from "../kind-factory-Bvwe3pup.js";
7
7
  import { s as makeProposalFinding } from "../usage-receipt-CgxMEBZq.js";
8
- import { i as inMemoryCampaignStorage, n as createRunCostLedger, r as fsCampaignStorage } from "../single-run-lock-CiQThJxB.js";
8
+ import { i as inMemoryCampaignStorage, n as createRunCostLedger, r as fsCampaignStorage } from "../single-run-lock-BcvPQY57.js";
9
9
  import { m as mapConcurrentRange } from "../ledger-core-Dxz0Rkwa.js";
10
- import { A as assertOptimizationResult, B as surfaceHash, Dt as decidePairedPromotion, J as runCampaign, O as composeGate, S as defaultProductionGate, T as heldoutSignificance, Y as resolveRunDir, _ as buildEvidenceVector, a as emitLoopProvenance, at as createReferenceEquivalenceJudge, b as powerPreflight, d as runImprovementLoop, et as campaignSplitDigest, g as gepaOptimizationMethod, ht as llmJudge, it as REFERENCE_EQUIVALENCE_JUDGE_VERSION, j as compareOptimizationMethods, k as externalTextOptimizationMethod, o as loopProvenanceArgsFromResult, ot as runReferenceEquivalenceJudge, p as runEval, rt as REFERENCE_EQUIVALENCE_INPUT_LIMITS, t as skillOptOptimizationMethod, v as paretoPolicy, x as heldOutGate, y as paretoSignificanceGate, z as surfaceContentHash } from "../skillopt-optimization-method-CSBQ8Qma.js";
10
+ import { A as assertOptimizationResult, B as surfaceHash, Dt as decidePairedPromotion, J as runCampaign, O as composeGate, S as defaultProductionGate, T as heldoutSignificance, Y as resolveRunDir, _ as buildEvidenceVector, a as emitLoopProvenance, at as createReferenceEquivalenceJudge, b as powerPreflight, d as runImprovementLoop, et as campaignSplitDigest, g as gepaOptimizationMethod, ht as llmJudge, it as REFERENCE_EQUIVALENCE_JUDGE_VERSION, j as compareOptimizationMethods, k as externalTextOptimizationMethod, o as loopProvenanceArgsFromResult, ot as runReferenceEquivalenceJudge, p as runEval, rt as REFERENCE_EQUIVALENCE_INPUT_LIMITS, t as skillOptOptimizationMethod, v as paretoPolicy, x as heldOutGate, y as paretoSignificanceGate, z as surfaceContentHash } from "../skillopt-optimization-method-DV6c4ooy.js";
11
11
  import { E as pairedBootstrap } from "../statistics-ByxzSiOM.js";
12
12
  import { i as parseRunRecordSafe, r as modelHasSnapshot } from "../run-record-vRgqWmJw.js";
13
- import { a as recordAggregateMeasurements, i as readTaskFailureLabels, o as summarizeExecutionMeasurements, s as summarizeTraceErrors, t as extractUsage } from "../extract-usage-C5vMw-0R.js";
13
+ import { a as recordAggregateMeasurements, i as readTaskFailureLabels, o as summarizeExecutionMeasurements, s as summarizeTraceErrors, t as extractUsage } from "../extract-usage-7l1Xq5ti.js";
14
14
  import { n as summarizeExecution, t as analyzeRuns } from "../analyze-runs-BScZvqMV.js";
15
15
  import { a as campaignCellExecutionEvidence, c as campaignCellToRunRecord, o as campaignCellJudgeDimensions, s as campaignCellTaskScore } from "../reward-hacking-GyN0kMd8.js";
16
16
  import { n as InMemoryOutcomeStore, t as FileSystemOutcomeStore } from "../outcome-store-ChBKlTd_.js";
@@ -845,7 +845,9 @@ function evaluatePairedMeasurements(options) {
845
845
  const measurementCostUsd = measurements.reduce((sum, measurement) => sum + measurement.baseline.costUsd + measurement.candidate.costUsd, 0);
846
846
  const measurementWorkDurationMs = measurements.reduce((sum, measurement) => sum + measurement.baseline.latencyMs + measurement.candidate.latencyMs, 0);
847
847
  const incompleteRuns = measurements.flatMap((measurement) => [measurement.baseline, measurement.candidate]).filter((run) => !run.completed);
848
- const failedCandidateResults = measurements.filter((measurement) => !measurement.candidate.passed);
848
+ const regressedTasks = measurements.filter((measurement) => !measurement.candidate.passed && measurement.baseline.passed);
849
+ const sharedTaskFailures = measurements.filter((measurement) => !measurement.candidate.passed && !measurement.baseline.passed);
850
+ const repairedTasks = measurements.filter((measurement) => measurement.candidate.passed && !measurement.baseline.passed);
849
851
  const derivedMeasurementCost = {
850
852
  usd: measurementCostUsd,
851
853
  provenance: measurements.every((measurement) => measurement.baseline.costProvenance === "observed" && measurement.candidate.costProvenance === "observed") ? "observed" : "estimated"
@@ -872,8 +874,8 @@ function evaluatePairedMeasurements(options) {
872
874
  passed: incompleteRuns.length === 0
873
875
  },
874
876
  {
875
- name: "candidate-task-pass",
876
- passed: failedCandidateResults.length === 0
877
+ name: "no-task-regression",
878
+ passed: regressedTasks.length === 0
877
879
  },
878
880
  {
879
881
  name: "critical-dimensions",
@@ -885,15 +887,16 @@ function evaluatePairedMeasurements(options) {
885
887
  }
886
888
  ];
887
889
  const shipped = checks.every((check) => check.passed);
888
- const hardFailure = incompleteRuns.length > 0 || failedCandidateResults.length > 0 || regressions.length > 0 || missingCriticalDimensions.length > 0 || !budgetPassed;
890
+ const hardFailure = incompleteRuns.length > 0 || regressedTasks.length > 0 || regressions.length > 0 || missingCriticalDimensions.length > 0 || !budgetPassed;
889
891
  const reasons = [
890
- ...significance.significant ? [] : [significance.fewRuns ? `only ${significance.n} paired runs; ${significance.minimumRequired} required` : `paired interval lower bound ${significance.bootstrap.low} did not clear ${deltaThreshold}`],
892
+ ...significance.significant ? [] : [significanceReason(significance, deltaThreshold)],
891
893
  ...powerSufficient || significance.fewRuns ? [] : [power.reason],
892
894
  ...regressions.length === 0 ? [] : [`critical dimensions regressed: ${regressions.map((entry) => entry.name).join(", ")}`],
893
895
  ...missingCriticalDimensions.length === 0 ? [] : [`critical dimensions missing: ${missingCriticalDimensions.join(", ")}`],
894
896
  ...incompleteRuns.length === 0 ? [] : [`${incompleteRuns.length} benchmark executions did not exit successfully`],
895
- ...failedCandidateResults.length === 0 ? [] : [`candidate failed ${failedCandidateResults.length} benchmark tasks`],
896
- ...budgetPassed ? [] : [`total cost ${totalCost.usd} exceeded budget ${budgetUsd}`]
897
+ ...regressedTasks.length === 0 ? [] : [`candidate regressed ${regressedTasks.length} benchmark ${regressedTasks.length === 1 ? "task" : "tasks"} the baseline passed: ` + regressedTasks.map((measurement) => measurement.cellId).join(", ")],
898
+ ...budgetPassed ? [] : [`total cost ${totalCost.usd} exceeded budget ${budgetUsd}`],
899
+ ...shipped || sharedTaskFailures.length === 0 ? [] : [`benchmark tasks: ${repairedTasks.length} repaired, ${regressedTasks.length} regressed, ${sharedTaskFailures.length} still failing that the baseline also failed`]
897
900
  ];
898
901
  return {
899
902
  overall: {
@@ -904,7 +907,7 @@ function evaluatePairedMeasurements(options) {
904
907
  },
905
908
  objectives,
906
909
  decision: {
907
- outcome: shipped ? "ship" : hardFailure ? "hold" : significance.fewRuns || !powerSufficient ? "need_more_work" : "hold",
910
+ outcome: shipped ? "ship" : hardFailure ? "hold" : significance.fewRuns || !powerSufficient || significance.decision.indeterminate ? "need_more_work" : "hold",
908
911
  reasons: reasons.length > 0 ? reasons : ["all measured checks passed"],
909
912
  contributingChecks: checks
910
913
  },
@@ -922,6 +925,46 @@ function evaluatePairedMeasurements(options) {
922
925
  measurementWorkDurationMs
923
926
  };
924
927
  }
928
+ /**
929
+ * Say why the paired significance check refused, reading the interval that
930
+ * ACTUALLY decided.
931
+ *
932
+ * The old string always printed `significance.bootstrap.low`, which is a
933
+ * DIAGNOSTIC bootstrap that on several paths never enters the decision at all.
934
+ * That produced self-contradicting holds — observed live: "paired interval lower
935
+ * bound 0.6666666666666669 did not clear 0", where 0.667 plainly clears 0. Three
936
+ * paths where the old string was simply reporting the wrong number:
937
+ *
938
+ * - ZERO-VARIANCE sample. Every paired delta identical ⇒ the resample
939
+ * distribution is a point mass and the interval collapses to [g, g]. The
940
+ * refusal is correct and load-bearing, and the reason has to say so.
941
+ * Under the bounded asymmetric null this repo documents (2 % of pairs drop
942
+ * by 1.0, the rest gain 0.0204, true mean paired delta exactly 0), a
943
+ * constant-positive sample is exactly a sample in which no pair drew the
944
+ * drop, so its frequency is closed-form 0.98^n — 88.6 % at n = 6 and 66.8 %
945
+ * at n = 20. That reconciles with the false-promotion rates measured on the
946
+ * composable gate for the same null, 88.50 % at n = 6 and 65.65 % at n = 20
947
+ * (see {@link decidePairedPromotion} in `src/paired-promotion-decision.ts`);
948
+ * no number here is independently measured, and none needs to be. A
949
+ * constant positive delta is therefore NOT the strongest evidence — it is
950
+ * the signature of a held-out set too small or too uniform to have sampled
951
+ * the tail, and promoting on it would false-promote a true-zero candidate
952
+ * most of the time. No distribution-free bound rescues it either: a
953
+ * Hoeffding 95 % lower bound on a constant +0.667 with scores in [0, 1] is
954
+ * −0.33 at n = 6 and −0.11 at n = 10, first clearing 0 at n = 20. The
955
+ * estimator genuinely cannot express this sample, so the gate says that and
956
+ * routes to `need_more_work` — vary or extend the held-out cells.
957
+ * - TWO-POINT outcome, where the score interval decided and the bootstrap was
958
+ * never consulted.
959
+ * - McNEMAR VETO, where no interval failed at all.
960
+ */
961
+ function significanceReason(significance, deltaThreshold) {
962
+ if (significance.fewRuns) return `only ${significance.n} paired runs; ${significance.minimumRequired} required`;
963
+ const { decision } = significance;
964
+ if (decision.indeterminate) return `paired ${decision.statistic} interval is degenerate at [${decision.low}, ${decision.high}] (${decision.indeterminateCause}); a zero-width interval carries no information about its own error, so no threshold can be cleared on it — vary or extend the held-out cells`;
965
+ if (decision.exactTestVetoes && decision.mcnemar !== null) return `McNemar's exact test refuses at threshold ${deltaThreshold}: b=${decision.mcnemar.b}, c=${decision.mcnemar.c}, p=${decision.mcnemar.pValue}`;
966
+ return `paired ${decision.statistic} interval lower bound ${decision.low} did not clear ${deltaThreshold}.${decision.methodDetail}`;
967
+ }
925
968
  /** Build the only publishable comparison: paired statistics over Runtime receipts. */
926
969
  function measuredComparisonFromCandidateExperiment(options) {
927
970
  const experiment = verifyCandidateExperiment(options.experiment);