@tangle-network/agent-eval 0.117.0 → 0.118.0

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 (143) hide show
  1. package/CHANGELOG.md +30 -0
  2. package/dist/analyst/index.d.ts +2772 -21
  3. package/dist/analyst/index.js +7 -6
  4. package/dist/analyst/index.js.map +1 -1
  5. package/dist/belief-state/index.d.ts +706 -10
  6. package/dist/belief-state/index.js +2 -1
  7. package/dist/belief-state/index.js.map +1 -1
  8. package/dist/benchmarks/index.d.ts +958 -14
  9. package/dist/benchmarks/index.js +12 -10
  10. package/dist/builder-eval/index.d.ts +449 -4
  11. package/dist/builder-eval/index.js +4 -3
  12. package/dist/builder-eval/index.js.map +1 -1
  13. package/dist/campaign/index.d.ts +4276 -74
  14. package/dist/campaign/index.js +12 -10
  15. package/dist/{chunk-VF3XSYTI.js → chunk-33JA4TFA.js} +6 -6
  16. package/dist/{chunk-4JLWXDYA.js → chunk-3EHHMC6E.js} +2 -2
  17. package/dist/{chunk-CCZIVI3F.js → chunk-BFW56GTT.js} +2 -2
  18. package/dist/{chunk-YZPO4UHR.js → chunk-DP3BBMYJ.js} +100 -143
  19. package/dist/chunk-DP3BBMYJ.js.map +1 -0
  20. package/dist/{chunk-E4BUPP7Z.js → chunk-HKUCJ437.js} +38 -66
  21. package/dist/chunk-HKUCJ437.js.map +1 -0
  22. package/dist/{chunk-HZHNRYHK.js → chunk-K6N6XJJX.js} +2 -2
  23. package/dist/chunk-KSDQVPLR.js +286 -0
  24. package/dist/chunk-KSDQVPLR.js.map +1 -0
  25. package/dist/chunk-MA6HLL3S.js +65 -0
  26. package/dist/chunk-MA6HLL3S.js.map +1 -0
  27. package/dist/{chunk-MGEHEHSN.js → chunk-OIIMMLRB.js} +11 -11
  28. package/dist/{chunk-DXZRATT5.js → chunk-OYZAPX5G.js} +3 -3
  29. package/dist/chunk-PXE2VKMX.js +140 -0
  30. package/dist/chunk-PXE2VKMX.js.map +1 -0
  31. package/dist/{chunk-ZUXV7UWZ.js → chunk-Q442S5AS.js} +73 -19
  32. package/dist/{chunk-ZUXV7UWZ.js.map → chunk-Q442S5AS.js.map} +1 -1
  33. package/dist/{chunk-ODVOOEWQ.js → chunk-QBRSJK47.js} +2 -2
  34. package/dist/{chunk-S2F4J57L.js → chunk-QKEGNI5B.js} +77 -32
  35. package/dist/chunk-QKEGNI5B.js.map +1 -0
  36. package/dist/{chunk-5UF54T55.js → chunk-S3UZOQ5Y.js} +34 -7
  37. package/dist/chunk-S3UZOQ5Y.js.map +1 -0
  38. package/dist/{chunk-FQNLDL4D.js → chunk-SVH2ANFD.js} +136 -4
  39. package/dist/chunk-SVH2ANFD.js.map +1 -0
  40. package/dist/{chunk-GQCZRZ7L.js → chunk-U5CHZ5M3.js} +9 -9
  41. package/dist/{chunk-TVVP3ZZQ.js → chunk-VQMK5FMP.js} +2 -1
  42. package/dist/{chunk-TVVP3ZZQ.js.map → chunk-VQMK5FMP.js.map} +1 -1
  43. package/dist/chunk-WDHBCA3M.js +31 -0
  44. package/dist/chunk-WDHBCA3M.js.map +1 -0
  45. package/dist/{chunk-HQPHZGL6.js → chunk-YLMUS4MM.js} +9 -9
  46. package/dist/{chunk-LQUTGLOZ.js → chunk-ZET2UAYW.js} +15 -65
  47. package/dist/chunk-ZET2UAYW.js.map +1 -0
  48. package/dist/contract/index.d.ts +4012 -38
  49. package/dist/contract/index.js +56 -29
  50. package/dist/contract/index.js.map +1 -1
  51. package/dist/control.d.ts +1013 -9
  52. package/dist/control.js +4 -3
  53. package/dist/fuzz.d.ts +194 -4
  54. package/dist/fuzz.js +3 -3
  55. package/dist/hosted/index.d.ts +498 -17
  56. package/dist/index.d.ts +10982 -1286
  57. package/dist/index.js +97 -80
  58. package/dist/index.js.map +1 -1
  59. package/dist/matrix/index.d.ts +139 -4
  60. package/dist/meta-eval/index.d.ts +862 -15
  61. package/dist/meta-eval/index.js +2 -1
  62. package/dist/meta-eval/index.js.map +1 -1
  63. package/dist/multishot/index.d.ts +214 -14
  64. package/dist/openapi.json +1 -1
  65. package/dist/pipelines/index.d.ts +392 -7
  66. package/dist/pipelines/index.js +5 -3
  67. package/dist/pipelines/index.js.map +1 -1
  68. package/dist/reporting.d.ts +1277 -17
  69. package/dist/rl.d.ts +2359 -28
  70. package/dist/rl.js +6 -5
  71. package/dist/rl.js.map +1 -1
  72. package/dist/storyboard/index.d.ts +86 -1
  73. package/dist/trace-attributes.d.ts +16 -0
  74. package/dist/trace-attributes.js +32 -0
  75. package/dist/trace-attributes.js.map +1 -0
  76. package/dist/traces.d.ts +1970 -697
  77. package/dist/traces.js +49 -32
  78. package/dist/wire/index.d.ts +655 -9
  79. package/docs/insight-report.md +44 -0
  80. package/package.json +9 -3
  81. package/dist/adversarial-B7loGVVX.d.ts +0 -19
  82. package/dist/analyst-C8HHvfJp.d.ts +0 -88
  83. package/dist/analyze-runs--2x39HZ7.d.ts +0 -81
  84. package/dist/baseline-DKq3gJpP.d.ts +0 -141
  85. package/dist/calibration-C8MTS7cw.d.ts +0 -101
  86. package/dist/chunk-5UF54T55.js.map +0 -1
  87. package/dist/chunk-E4BUPP7Z.js.map +0 -1
  88. package/dist/chunk-FQNLDL4D.js.map +0 -1
  89. package/dist/chunk-LQUTGLOZ.js.map +0 -1
  90. package/dist/chunk-S2F4J57L.js.map +0 -1
  91. package/dist/chunk-YZPO4UHR.js.map +0 -1
  92. package/dist/code-agent-session-CjZsVd19.d.ts +0 -87
  93. package/dist/control-6vuGfmDH.d.ts +0 -258
  94. package/dist/cost-ledger-DWy3XdJc.d.ts +0 -183
  95. package/dist/dataset-NENEzRgk.d.ts +0 -115
  96. package/dist/default-registry-DaK8b3fv.d.ts +0 -155
  97. package/dist/emitter-CjD7vUwv.d.ts +0 -122
  98. package/dist/errors-oeQrLqXC.d.ts +0 -74
  99. package/dist/failure-cluster-DOAcSJ87.d.ts +0 -76
  100. package/dist/feedback-trajectory-BUnM58xL.d.ts +0 -348
  101. package/dist/gepa-eESocoDi.d.ts +0 -642
  102. package/dist/index-PdX4VnPA.d.ts +0 -423
  103. package/dist/insight-report-DY4nDW9Q.d.ts +0 -310
  104. package/dist/integrity-DqlBiLyK.d.ts +0 -81
  105. package/dist/judge-calibration-7C-IDmKr.d.ts +0 -145
  106. package/dist/kind-factory-ClZmO25A.d.ts +0 -171
  107. package/dist/llm-client-qoDd18Qz.d.ts +0 -289
  108. package/dist/multi-layer-verifier-BsqKuLyN.d.ts +0 -150
  109. package/dist/off-policy-DiwuKKg7.d.ts +0 -132
  110. package/dist/outcome-store-rnXLEqSn.d.ts +0 -63
  111. package/dist/policy-edit-wG9uFEFm.d.ts +0 -455
  112. package/dist/pre-registration-BWQhJ3vz.d.ts +0 -761
  113. package/dist/provenance-DpjwyseI.d.ts +0 -541
  114. package/dist/query-CF7PG61p.d.ts +0 -35
  115. package/dist/raw-provider-sink-C46HDghv.d.ts +0 -132
  116. package/dist/release-report-C8G2i5Xi.d.ts +0 -236
  117. package/dist/researcher-C8XyxQsu.d.ts +0 -387
  118. package/dist/rubric-predictive-validity-p49lLVrE.d.ts +0 -105
  119. package/dist/run-record-BDH49H2E.d.ts +0 -360
  120. package/dist/runtime-trajectory-DGBIUt4B.d.ts +0 -49
  121. package/dist/schema-B3Q3l9Z_.d.ts +0 -201
  122. package/dist/semantic-concept-judge-CXnPEJbf.d.ts +0 -723
  123. package/dist/sequential-5iSVfzl2.d.ts +0 -139
  124. package/dist/series-convergence-D5OWMBg6.d.ts +0 -33
  125. package/dist/statistics-KUnG73jH.d.ts +0 -494
  126. package/dist/storage-DrX3v_5B.d.ts +0 -50
  127. package/dist/store-C1YxJDEK.d.ts +0 -248
  128. package/dist/store-DGqD0Pyo.d.ts +0 -116
  129. package/dist/summary-report-C5bKFfm-.d.ts +0 -445
  130. package/dist/test-graded-scenario-B0ybnPY7.d.ts +0 -166
  131. package/dist/types-BSw1rOUB.d.ts +0 -634
  132. package/dist/types-BUxNaJ8c.d.ts +0 -108
  133. package/dist/types-BkfcQnxV.d.ts +0 -313
  134. package/dist/verdict-C9MlYujm.d.ts +0 -35
  135. /package/dist/{chunk-VF3XSYTI.js.map → chunk-33JA4TFA.js.map} +0 -0
  136. /package/dist/{chunk-4JLWXDYA.js.map → chunk-3EHHMC6E.js.map} +0 -0
  137. /package/dist/{chunk-CCZIVI3F.js.map → chunk-BFW56GTT.js.map} +0 -0
  138. /package/dist/{chunk-HZHNRYHK.js.map → chunk-K6N6XJJX.js.map} +0 -0
  139. /package/dist/{chunk-MGEHEHSN.js.map → chunk-OIIMMLRB.js.map} +0 -0
  140. /package/dist/{chunk-DXZRATT5.js.map → chunk-OYZAPX5G.js.map} +0 -0
  141. /package/dist/{chunk-ODVOOEWQ.js.map → chunk-QBRSJK47.js.map} +0 -0
  142. /package/dist/{chunk-GQCZRZ7L.js.map → chunk-U5CHZ5M3.js.map} +0 -0
  143. /package/dist/{chunk-HQPHZGL6.js.map → chunk-YLMUS4MM.js.map} +0 -0
package/CHANGELOG.md CHANGED
@@ -4,6 +4,36 @@ All notable changes to `@tangle-network/agent-eval` and its sibling `agent-eval-
4
4
 
5
5
  ---
6
6
 
7
+ ## [0.118.0] — 2026-07-13 — complete execution accounting
8
+
9
+ ### Added
10
+
11
+ - `InsightReport.execution` reports duration, optional queue time, direct input, output, reasoning, cache-read, and cache-write tokens, model cohorts, model-call coverage, explicit failure counts, and separately labeled orchestration aggregates from the same `RunRecord[]` passed to `analyzeRuns()`.
12
+ - `summarizeExecution()` returns those execution facts and cost provenance without interpreting task quality.
13
+ - `RunTokenUsage.reasoning` preserves the reasoning subset of normalized output, and `RunTokenUsage.cacheWrite` preserves provider cache creation separately from cache reads.
14
+
15
+ ### Changed
16
+
17
+ - Trace capture, every OTLP exporter, OTLP intake, and code-agent session intake preserve reasoning, cache reads, and cache writes separately.
18
+ - Both OTLP intake paths use one field-by-field reconciliation rule for nested model-call wrappers, preserving complementary parent data without double-counting complete child data.
19
+ - Code-agent intake uses the shared provider-usage parser, including OpenCode's nested `cache.read` and `cache.write` fields and OpenAI-compatible token-detail objects.
20
+ - OTLP-derived run records explicitly label complete USD as observed, model-priced USD as estimated, and missing or partial USD as uncaptured instead of relying on a zero-value inference.
21
+ - Usage parsing reuses `@tangle-network/agent-core` token vocabulary and SSE framing, preserves agent-eval-specific reasoning and cache-write details, reconciles cumulative streams by default, and accepts explicit delta mode through `captureFetchToRawSink({ sseUsageMode })`.
22
+ - Run-record validation rejects negative execution measurements and unknown failure classes, and OTLP intake rejects duplicate span identities instead of corrupting totals.
23
+ - Declaration bundles build sequentially and package verification compiles a strict Node consumer, preserving the public subpath types without concurrent declaration workers exhausting memory.
24
+ - `@tangle-network/agent-interface` is updated to `0.26.x`.
25
+
26
+ ### Breaking
27
+
28
+ - `InsightReport.execution`, `CodeAgentSessionMetrics.reasoningTokens`, and `CodeAgentSessionMetrics.cacheWriteTokens` are required fields on newly constructed objects.
29
+
30
+ ## [0.117.1] — 2026-07-13 — retry-safe code-candidate cleanup
31
+
32
+ ### Fixed
33
+
34
+ - `gitWorktreeAdapter().discard()` now reconciles worktree and branch removal independently.
35
+ Repeated cleanup is safe, partial cleanup can be retried, and a Git command that reports an error after completing its mutation no longer strands candidate branches or worktrees.
36
+
7
37
  ## [0.117.0] — 2026-07-13 — durable cost and bounded behavioral evidence
8
38
 
9
39
  ### Added