@tangle-network/agent-eval 0.93.0 → 0.95.0
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.
- package/CHANGELOG.md +15 -0
- package/README.md +44 -30
- package/dist/action-policy.d.ts +24 -0
- package/dist/action-policy.d.ts.map +1 -0
- package/dist/action-policy.test.d.ts +2 -0
- package/dist/action-policy.test.d.ts.map +1 -0
- package/dist/active-learning.d.ts +41 -0
- package/dist/active-learning.d.ts.map +1 -0
- package/dist/adapters/http.d.ts +15 -21
- package/dist/adapters/http.d.ts.map +1 -0
- package/dist/adapters/http.js.map +1 -1
- package/dist/adapters/langchain.d.ts +7 -13
- package/dist/adapters/langchain.d.ts.map +1 -0
- package/dist/adapters/otel.d.ts +13 -24
- package/dist/adapters/otel.d.ts.map +1 -0
- package/dist/agent-profile-cell.d.ts +101 -0
- package/dist/agent-profile-cell.d.ts.map +1 -0
- package/dist/agent-profile.d.ts +27 -0
- package/dist/agent-profile.d.ts.map +1 -0
- package/dist/agent-profile.test.d.ts +2 -0
- package/dist/agent-profile.test.d.ts.map +1 -0
- package/dist/analyst/adapters.d.ts +62 -0
- package/dist/analyst/adapters.d.ts.map +1 -0
- package/dist/analyst/analyst.test.d.ts +2 -0
- package/dist/analyst/analyst.test.d.ts.map +1 -0
- package/dist/analyst/ax-service.d.ts +27 -0
- package/dist/analyst/ax-service.d.ts.map +1 -0
- package/dist/analyst/behavioral-analyst.d.ts +28 -0
- package/dist/analyst/behavioral-analyst.d.ts.map +1 -0
- package/dist/analyst/chat-client.d.ts +91 -0
- package/dist/analyst/chat-client.d.ts.map +1 -0
- package/dist/analyst/default-registry.d.ts +27 -0
- package/dist/analyst/default-registry.d.ts.map +1 -0
- package/dist/analyst/default-registry.test.d.ts +2 -0
- package/dist/analyst/default-registry.test.d.ts.map +1 -0
- package/dist/analyst/finding-signature.d.ts +48 -0
- package/dist/analyst/finding-signature.d.ts.map +1 -0
- package/dist/analyst/finding-subject.d.ts +146 -0
- package/dist/analyst/finding-subject.d.ts.map +1 -0
- package/dist/analyst/finding-subject.test.d.ts +2 -0
- package/dist/analyst/finding-subject.test.d.ts.map +1 -0
- package/dist/analyst/findings-store.d.ts +75 -0
- package/dist/analyst/findings-store.d.ts.map +1 -0
- package/dist/analyst/index.d.ts +28 -256
- package/dist/analyst/index.d.ts.map +1 -0
- package/dist/analyst/index.js +6 -66
- package/dist/analyst/index.js.map +1 -1
- package/dist/{kind-factory-5b7xXXOr.d.ts → analyst/kind-factory.d.ts} +11 -63
- package/dist/analyst/kind-factory.d.ts.map +1 -0
- package/dist/analyst/kinds/failure-mode.d.ts +19 -0
- package/dist/analyst/kinds/failure-mode.d.ts.map +1 -0
- package/dist/analyst/kinds/improvement.d.ts +23 -0
- package/dist/analyst/kinds/improvement.d.ts.map +1 -0
- package/dist/analyst/kinds/index.d.ts +22 -0
- package/dist/analyst/kinds/index.d.ts.map +1 -0
- package/dist/analyst/kinds/kinds.test.d.ts +2 -0
- package/dist/analyst/kinds/kinds.test.d.ts.map +1 -0
- package/dist/analyst/kinds/knowledge-gap.d.ts +28 -0
- package/dist/analyst/kinds/knowledge-gap.d.ts.map +1 -0
- package/dist/analyst/kinds/knowledge-poisoning.d.ts +22 -0
- package/dist/analyst/kinds/knowledge-poisoning.d.ts.map +1 -0
- package/dist/analyst/kinds/skill-usage.d.ts +84 -0
- package/dist/analyst/kinds/skill-usage.d.ts.map +1 -0
- package/dist/analyst/kinds/skill-usage.test.d.ts +2 -0
- package/dist/analyst/kinds/skill-usage.test.d.ts.map +1 -0
- package/dist/analyst/parse-tolerant.d.ts +26 -0
- package/dist/analyst/parse-tolerant.d.ts.map +1 -0
- package/dist/analyst/parse-tolerant.test.d.ts +2 -0
- package/dist/analyst/parse-tolerant.test.d.ts.map +1 -0
- package/dist/analyst/registry.budget.test.d.ts +2 -0
- package/dist/analyst/registry.budget.test.d.ts.map +1 -0
- package/dist/{default-registry-zoGHUQEH.d.ts → analyst/registry.d.ts} +8 -37
- package/dist/analyst/registry.d.ts.map +1 -0
- package/dist/analyst/steer-firewall.d.ts +35 -0
- package/dist/analyst/steer-firewall.d.ts.map +1 -0
- package/dist/analyst/steer-firewall.test.d.ts +2 -0
- package/dist/analyst/steer-firewall.test.d.ts.map +1 -0
- package/dist/analyst/structure-findings.d.ts +37 -0
- package/dist/analyst/structure-findings.d.ts.map +1 -0
- package/dist/analyst/structure-findings.test.d.ts +2 -0
- package/dist/analyst/structure-findings.test.d.ts.map +1 -0
- package/dist/analyst/tool-groups.d.ts +34 -0
- package/dist/analyst/tool-groups.d.ts.map +1 -0
- package/dist/{types-2VVIL04s.d.ts → analyst/types.d.ts} +19 -112
- package/dist/analyst/types.d.ts.map +1 -0
- package/dist/anti-slop.d.ts +59 -0
- package/dist/anti-slop.d.ts.map +1 -0
- package/dist/artifact-validator.d.ts +74 -0
- package/dist/artifact-validator.d.ts.map +1 -0
- package/dist/attestation.d.ts +63 -0
- package/dist/attestation.d.ts.map +1 -0
- package/dist/attestation.test.d.ts +2 -0
- package/dist/attestation.test.d.ts.map +1 -0
- package/dist/authenticity/index.d.ts +15 -16
- package/dist/authenticity/index.d.ts.map +1 -0
- package/dist/authenticity/index.test.d.ts +2 -0
- package/dist/authenticity/index.test.d.ts.map +1 -0
- package/dist/auto-pr.d.ts +120 -0
- package/dist/auto-pr.d.ts.map +1 -0
- package/dist/{baseline-DE36-Np7.d.ts → baseline.d.ts} +8 -44
- package/dist/baseline.d.ts.map +1 -0
- package/dist/behavior-dsl.d.ts +73 -0
- package/dist/behavior-dsl.d.ts.map +1 -0
- package/dist/belief-state/calibration.d.ts +10 -0
- package/dist/belief-state/calibration.d.ts.map +1 -0
- package/dist/belief-state/calibration.test.d.ts +2 -0
- package/dist/belief-state/calibration.test.d.ts.map +1 -0
- package/dist/belief-state/code-agent-corpus.d.ts +66 -0
- package/dist/belief-state/code-agent-corpus.d.ts.map +1 -0
- package/dist/belief-state/code-agent-corpus.test.d.ts +2 -0
- package/dist/belief-state/code-agent-corpus.test.d.ts.map +1 -0
- package/dist/belief-state/code-agent-evidence.d.ts +22 -0
- package/dist/belief-state/code-agent-evidence.d.ts.map +1 -0
- package/dist/belief-state/code-agent-evidence.test.d.ts +2 -0
- package/dist/belief-state/code-agent-evidence.test.d.ts.map +1 -0
- package/dist/belief-state/extract.d.ts +7 -0
- package/dist/belief-state/extract.d.ts.map +1 -0
- package/dist/belief-state/extract.test.d.ts +2 -0
- package/dist/belief-state/extract.test.d.ts.map +1 -0
- package/dist/belief-state/index.d.ts +14 -604
- package/dist/belief-state/index.d.ts.map +1 -0
- package/dist/belief-state/index.js +1 -1
- package/dist/belief-state/ope.d.ts +17 -0
- package/dist/belief-state/ope.d.ts.map +1 -0
- package/dist/belief-state/ope.test.d.ts +2 -0
- package/dist/belief-state/ope.test.d.ts.map +1 -0
- package/dist/belief-state/phase0-measurement.d.ts +55 -0
- package/dist/belief-state/phase0-measurement.d.ts.map +1 -0
- package/dist/belief-state/report.d.ts +17 -0
- package/dist/belief-state/report.d.ts.map +1 -0
- package/dist/belief-state/report.test.d.ts +2 -0
- package/dist/belief-state/report.test.d.ts.map +1 -0
- package/dist/belief-state/research-evidence.d.ts +23 -0
- package/dist/belief-state/research-evidence.d.ts.map +1 -0
- package/dist/belief-state/research-evidence.test.d.ts +2 -0
- package/dist/belief-state/research-evidence.test.d.ts.map +1 -0
- package/dist/belief-state/runtime-benchmark-corpus.d.ts +32 -0
- package/dist/belief-state/runtime-benchmark-corpus.d.ts.map +1 -0
- package/dist/belief-state/runtime-hooks.d.ts +87 -0
- package/dist/belief-state/runtime-hooks.d.ts.map +1 -0
- package/dist/belief-state/runtime-hooks.test.d.ts +2 -0
- package/dist/belief-state/runtime-hooks.test.d.ts.map +1 -0
- package/dist/belief-state/selective.d.ts +15 -0
- package/dist/belief-state/selective.d.ts.map +1 -0
- package/dist/belief-state/selective.test.d.ts +2 -0
- package/dist/belief-state/selective.test.d.ts.map +1 -0
- package/dist/belief-state/shadow-probe.d.ts +80 -0
- package/dist/belief-state/shadow-probe.d.ts.map +1 -0
- package/dist/belief-state/shadow-probe.test.d.ts +2 -0
- package/dist/belief-state/shadow-probe.test.d.ts.map +1 -0
- package/dist/belief-state/types.d.ts +195 -0
- package/dist/belief-state/types.d.ts.map +1 -0
- package/dist/belief-state/types.test.d.ts +2 -0
- package/dist/belief-state/types.test.d.ts.map +1 -0
- package/dist/benchmark.d.ts +14 -0
- package/dist/benchmark.d.ts.map +1 -0
- package/dist/benchmarks/index.d.ts +23 -4
- package/dist/benchmarks/index.d.ts.map +1 -0
- package/dist/benchmarks/routing/dataset.d.ts +34 -0
- package/dist/benchmarks/routing/dataset.d.ts.map +1 -0
- package/dist/benchmarks/routing/index.d.ts +34 -0
- package/dist/benchmarks/routing/index.d.ts.map +1 -0
- package/dist/benchmarks/types.d.ts +49 -0
- package/dist/benchmarks/types.d.ts.map +1 -0
- package/dist/bisector.d.ts +81 -0
- package/dist/bisector.d.ts.map +1 -0
- package/dist/budget-guard.d.ts +31 -0
- package/dist/budget-guard.d.ts.map +1 -0
- package/dist/builder-eval/builder-session.d.ts +111 -0
- package/dist/builder-eval/builder-session.d.ts.map +1 -0
- package/dist/builder-eval/correlation.d.ts +32 -0
- package/dist/builder-eval/correlation.d.ts.map +1 -0
- package/dist/builder-eval/index.d.ts +5 -250
- package/dist/builder-eval/index.d.ts.map +1 -0
- package/dist/builder-eval/index.js +2 -2
- package/dist/builder-eval/project-registry.d.ts +51 -0
- package/dist/builder-eval/project-registry.d.ts.map +1 -0
- package/dist/builder-eval/three-layer-eval.d.ts +55 -0
- package/dist/builder-eval/three-layer-eval.d.ts.map +1 -0
- package/dist/campaign/analyst-surface.d.ts +108 -0
- package/dist/campaign/analyst-surface.d.ts.map +1 -0
- package/dist/campaign/analyst-surface.test.d.ts +2 -0
- package/dist/campaign/analyst-surface.test.d.ts.map +1 -0
- package/dist/campaign/auto-pr.d.ts +46 -0
- package/dist/campaign/auto-pr.d.ts.map +1 -0
- package/dist/campaign/distillation/agreement-judge.d.ts +69 -0
- package/dist/campaign/distillation/agreement-judge.d.ts.map +1 -0
- package/dist/campaign/distillation/cli.d.ts +35 -0
- package/dist/campaign/distillation/cli.d.ts.map +1 -0
- package/dist/campaign/distillation/distillation.test.d.ts +2 -0
- package/dist/campaign/distillation/distillation.test.d.ts.map +1 -0
- package/dist/campaign/distillation/gold-scenarios.d.ts +54 -0
- package/dist/campaign/distillation/gold-scenarios.d.ts.map +1 -0
- package/dist/campaign/distillation/run-distillation.d.ts +119 -0
- package/dist/campaign/distillation/run-distillation.d.ts.map +1 -0
- package/dist/campaign/gates/compose.d.ts +12 -0
- package/dist/campaign/gates/compose.d.ts.map +1 -0
- package/dist/campaign/gates/default-production-gate.d.ts +58 -0
- package/dist/campaign/gates/default-production-gate.d.ts.map +1 -0
- package/dist/campaign/gates/heldout-gate.d.ts +12 -0
- package/dist/campaign/gates/heldout-gate.d.ts.map +1 -0
- package/dist/campaign/gates/promotion-policy.d.ts +125 -0
- package/dist/campaign/gates/promotion-policy.d.ts.map +1 -0
- package/dist/campaign/gates/promotion-policy.test.d.ts +2 -0
- package/dist/campaign/gates/promotion-policy.test.d.ts.map +1 -0
- package/dist/campaign/gates/sequential.d.ts +146 -0
- package/dist/campaign/gates/sequential.d.ts.map +1 -0
- package/dist/campaign/gates/sequential.test.d.ts +2 -0
- package/dist/campaign/gates/sequential.test.d.ts.map +1 -0
- package/dist/campaign/gates/statistical-heldout.d.ts +99 -0
- package/dist/campaign/gates/statistical-heldout.d.ts.map +1 -0
- package/dist/campaign/gates/statistical-heldout.test.d.ts +2 -0
- package/dist/campaign/gates/statistical-heldout.test.d.ts.map +1 -0
- package/dist/campaign/index.d.ts +38 -1341
- package/dist/campaign/index.d.ts.map +1 -0
- package/dist/campaign/index.js +1865 -1318
- package/dist/campaign/index.js.map +1 -1
- package/dist/campaign/labeled-store/fs-adapter.d.ts +59 -0
- package/dist/campaign/labeled-store/fs-adapter.d.ts.map +1 -0
- package/dist/campaign/presets/compare-proposers.d.ts +146 -0
- package/dist/campaign/presets/compare-proposers.d.ts.map +1 -0
- package/dist/campaign/presets/playback.d.ts +120 -0
- package/dist/campaign/presets/playback.d.ts.map +1 -0
- package/dist/campaign/presets/playback.test.d.ts +2 -0
- package/dist/campaign/presets/playback.test.d.ts.map +1 -0
- package/dist/campaign/presets/run-eval.d.ts +14 -0
- package/dist/campaign/presets/run-eval.d.ts.map +1 -0
- package/dist/campaign/presets/run-improvement-loop.d.ts +63 -0
- package/dist/campaign/presets/run-improvement-loop.d.ts.map +1 -0
- package/dist/campaign/presets/run-improvement-loop.test.d.ts +2 -0
- package/dist/campaign/presets/run-improvement-loop.test.d.ts.map +1 -0
- package/dist/campaign/presets/run-optimization.d.ts +92 -0
- package/dist/campaign/presets/run-optimization.d.ts.map +1 -0
- package/dist/campaign/presets/run-profile-matrix.d.ts +151 -0
- package/dist/campaign/presets/run-profile-matrix.d.ts.map +1 -0
- package/dist/campaign/presets/run-skill-opt.d.ts +96 -0
- package/dist/campaign/presets/run-skill-opt.d.ts.map +1 -0
- package/dist/campaign/proposers/_findings-text.d.ts +22 -0
- package/dist/campaign/proposers/_findings-text.d.ts.map +1 -0
- package/dist/campaign/proposers/ace.d.ts +33 -0
- package/dist/campaign/proposers/ace.d.ts.map +1 -0
- package/dist/campaign/proposers/ace.test.d.ts +2 -0
- package/dist/campaign/proposers/ace.test.d.ts.map +1 -0
- package/dist/campaign/proposers/analysis-edit.d.ts +32 -0
- package/dist/campaign/proposers/analysis-edit.d.ts.map +1 -0
- package/dist/campaign/proposers/evolutionary.d.ts +20 -0
- package/dist/campaign/proposers/evolutionary.d.ts.map +1 -0
- package/dist/campaign/proposers/fapo.d.ts +120 -0
- package/dist/campaign/proposers/fapo.d.ts.map +1 -0
- package/dist/campaign/proposers/gepa.d.ts +86 -0
- package/dist/campaign/proposers/gepa.d.ts.map +1 -0
- package/dist/campaign/proposers/halo.d.ts +44 -0
- package/dist/campaign/proposers/halo.d.ts.map +1 -0
- package/dist/campaign/proposers/halo.test.d.ts +2 -0
- package/dist/campaign/proposers/halo.test.d.ts.map +1 -0
- package/dist/campaign/proposers/memory.d.ts +47 -0
- package/dist/campaign/proposers/memory.d.ts.map +1 -0
- package/dist/campaign/proposers/memory.test.d.ts +2 -0
- package/dist/campaign/proposers/memory.test.d.ts.map +1 -0
- package/dist/campaign/proposers/skill-opt.d.ts +88 -0
- package/dist/campaign/proposers/skill-opt.d.ts.map +1 -0
- package/dist/campaign/proposers/trace-analyst.d.ts +48 -0
- package/dist/campaign/proposers/trace-analyst.d.ts.map +1 -0
- package/dist/campaign/proposers/trace-analyst.test.d.ts +2 -0
- package/dist/campaign/proposers/trace-analyst.test.d.ts.map +1 -0
- package/dist/campaign/provenance.d.ts +185 -0
- package/dist/campaign/provenance.d.ts.map +1 -0
- package/dist/campaign/run-campaign.d.ts +90 -0
- package/dist/campaign/run-campaign.d.ts.map +1 -0
- package/dist/campaign/score-utils.d.ts +26 -0
- package/dist/campaign/score-utils.d.ts.map +1 -0
- package/dist/campaign/skill-patch.d.ts +62 -0
- package/dist/campaign/skill-patch.d.ts.map +1 -0
- package/dist/campaign/storage.d.ts +38 -0
- package/dist/campaign/storage.d.ts.map +1 -0
- package/dist/{types-BU-7W85F.d.ts → campaign/types.d.ts} +98 -99
- package/dist/campaign/types.d.ts.map +1 -0
- package/dist/campaign/worktree/index.d.ts +53 -0
- package/dist/campaign/worktree/index.d.ts.map +1 -0
- package/dist/canary.d.ts +101 -0
- package/dist/canary.d.ts.map +1 -0
- package/dist/causal-attribution.d.ts +45 -0
- package/dist/causal-attribution.d.ts.map +1 -0
- package/dist/{chunk-TWS7AZEY.js → chunk-2T4EZACH.js} +2 -2
- package/dist/chunk-2T4EZACH.js.map +1 -0
- package/dist/{chunk-LMZQ2Z4U.js → chunk-56GC6VYO.js} +126 -1
- package/dist/chunk-56GC6VYO.js.map +1 -0
- package/dist/{chunk-QG2OVF2D.js → chunk-77T4STFI.js} +154 -112
- package/dist/chunk-77T4STFI.js.map +1 -0
- package/dist/{chunk-ZFIBGEOL.js → chunk-7QTQKIDD.js} +179 -178
- package/dist/chunk-7QTQKIDD.js.map +1 -0
- package/dist/{chunk-UMMZHCPB.js → chunk-AQ5WQAIV.js} +26 -9
- package/dist/chunk-AQ5WQAIV.js.map +1 -0
- package/dist/{chunk-CVVHBFGN.js → chunk-CWNP4DV4.js} +53 -5
- package/dist/chunk-CWNP4DV4.js.map +1 -0
- package/dist/{chunk-QS3RBQPI.js → chunk-D5KBVULY.js} +2 -2
- package/dist/chunk-DJWX3GVS.js +81 -0
- package/dist/chunk-DJWX3GVS.js.map +1 -0
- package/dist/{chunk-S6OZEZQK.js → chunk-HMA63UEO.js} +37 -9
- package/dist/{chunk-S6OZEZQK.js.map → chunk-HMA63UEO.js.map} +1 -1
- package/dist/{chunk-UHMJT4T7.js → chunk-IZCEK2HR.js} +2 -2
- package/dist/{chunk-6SOJM3VR.js → chunk-KKWJD5E6.js} +21 -21
- package/dist/chunk-KKWJD5E6.js.map +1 -0
- package/dist/{chunk-KWRRMR3J.js → chunk-LO6IOIJ2.js} +10 -10
- package/dist/chunk-LO6IOIJ2.js.map +1 -0
- package/dist/{chunk-L3JOU6XM.js → chunk-NACM5RS7.js} +3 -3
- package/dist/{chunk-E4GH6USR.js → chunk-NZEQVRH5.js} +2 -2
- package/dist/chunk-NZEQVRH5.js.map +1 -0
- package/dist/{chunk-XY4DDNEG.js → chunk-PSWWQXHF.js} +14 -89
- package/dist/chunk-PSWWQXHF.js.map +1 -0
- package/dist/{chunk-4FBZZIYD.js → chunk-QTMYW64F.js} +2 -2
- package/dist/{chunk-D3V5B42D.js → chunk-S4SYLDFX.js} +7 -4
- package/dist/chunk-S4SYLDFX.js.map +1 -0
- package/dist/{chunk-47X6LRCE.js → chunk-UFSG7ACU.js} +10 -7
- package/dist/chunk-UFSG7ACU.js.map +1 -0
- package/dist/{chunk-CY6U5S3X.js → chunk-URWVKRCS.js} +2 -2
- package/dist/{chunk-GSH6QNNS.js → chunk-X74V6ESX.js} +102 -79
- package/dist/chunk-X74V6ESX.js.map +1 -0
- package/dist/{chunk-QAY5UIJO.js → chunk-YBIGNSCZ.js} +178 -59
- package/dist/chunk-YBIGNSCZ.js.map +1 -0
- package/dist/{chunk-Y47J2LJ3.js → chunk-Z6L6YSU6.js} +5 -5
- package/dist/{chunk-T375SUOZ.js → chunk-ZOPLCJSY.js} +86 -31
- package/dist/chunk-ZOPLCJSY.js.map +1 -0
- package/dist/ci-gate.d.ts +44 -0
- package/dist/ci-gate.d.ts.map +1 -0
- package/dist/cli.d.ts +2 -0
- package/dist/cli.d.ts.map +1 -0
- package/dist/cli.js +2 -2
- package/dist/client.d.ts +77 -0
- package/dist/client.d.ts.map +1 -0
- package/dist/client.test.d.ts +2 -0
- package/dist/client.test.d.ts.map +1 -0
- package/dist/command-runner.d.ts +74 -0
- package/dist/command-runner.d.ts.map +1 -0
- package/dist/command-runner.test.d.ts +2 -0
- package/dist/command-runner.test.d.ts.map +1 -0
- package/dist/completion-verifier.d.ts +147 -0
- package/dist/completion-verifier.d.ts.map +1 -0
- package/dist/completion-verifier.test.d.ts +9 -0
- package/dist/completion-verifier.test.d.ts.map +1 -0
- package/dist/concurrency.d.ts +23 -0
- package/dist/concurrency.d.ts.map +1 -0
- package/dist/contamination-guard.d.ts +81 -0
- package/dist/contamination-guard.d.ts.map +1 -0
- package/dist/{analyze-runs-DwCEkpO_.d.ts → contract/analyze-runs.d.ts} +9 -10
- package/dist/contract/analyze-runs.d.ts.map +1 -0
- package/dist/contract/define-agent-eval.d.ts +52 -0
- package/dist/contract/define-agent-eval.d.ts.map +1 -0
- package/dist/contract/diff.d.ts +114 -0
- package/dist/contract/diff.d.ts.map +1 -0
- package/dist/contract/index.d.ts +106 -640
- package/dist/contract/index.d.ts.map +1 -0
- package/dist/contract/index.js +131 -21
- package/dist/contract/index.js.map +1 -1
- package/dist/{insight-report-BBwvOh6x.d.ts → contract/insight-report.d.ts} +16 -19
- package/dist/contract/insight-report.d.ts.map +1 -0
- package/dist/contract/insight-types-fwd.d.ts +7 -0
- package/dist/contract/insight-types-fwd.d.ts.map +1 -0
- package/dist/contract/intake/agent-trace.d.ts +97 -0
- package/dist/contract/intake/agent-trace.d.ts.map +1 -0
- package/dist/{code-agent-session-BO8nCnv3.d.ts → contract/intake/code-agent-session.d.ts} +15 -17
- package/dist/contract/intake/code-agent-session.d.ts.map +1 -0
- package/dist/contract/intake/feedback-table.d.ts +87 -0
- package/dist/contract/intake/feedback-table.d.ts.map +1 -0
- package/dist/contract/intake/index.d.ts +22 -0
- package/dist/contract/intake/index.d.ts.map +1 -0
- package/dist/contract/intake/otel-spans.d.ts +33 -0
- package/dist/contract/intake/otel-spans.d.ts.map +1 -0
- package/dist/contract/self-improve.d.ts +284 -0
- package/dist/contract/self-improve.d.ts.map +1 -0
- package/dist/{control-runtime-DuFBYg7A.d.ts → control-runtime.d.ts} +23 -26
- package/dist/control-runtime.d.ts.map +1 -0
- package/dist/control-runtime.test.d.ts +2 -0
- package/dist/control-runtime.test.d.ts.map +1 -0
- package/dist/control.d.ts +11 -9
- package/dist/control.d.ts.map +1 -0
- package/dist/control.js +2 -2
- package/dist/convergence.d.ts +29 -0
- package/dist/convergence.d.ts.map +1 -0
- package/dist/{cost-ledger-DuSqlw5B.d.ts → cost-ledger.d.ts} +10 -11
- package/dist/cost-ledger.d.ts.map +1 -0
- package/dist/cost-ledger.test.d.ts +2 -0
- package/dist/cost-ledger.test.d.ts.map +1 -0
- package/dist/cost-report.d.ts +43 -0
- package/dist/cost-report.d.ts.map +1 -0
- package/dist/cost-report.test.d.ts +2 -0
- package/dist/cost-report.test.d.ts.map +1 -0
- package/dist/cost-tracker.d.ts +76 -0
- package/dist/cost-tracker.d.ts.map +1 -0
- package/dist/{counterfactual-Dwibr5IW.d.ts → counterfactual.d.ts} +12 -13
- package/dist/counterfactual.d.ts.map +1 -0
- package/dist/cross-trace-diff.d.ts +56 -0
- package/dist/cross-trace-diff.d.ts.map +1 -0
- package/dist/{dataset-BbGkaN2I.d.ts → dataset.d.ts} +11 -14
- package/dist/dataset.d.ts.map +1 -0
- package/dist/deploy-gate-layer.d.ts +125 -0
- package/dist/deploy-gate-layer.d.ts.map +1 -0
- package/dist/deploy-gate-layer.test.d.ts +2 -0
- package/dist/deploy-gate-layer.test.d.ts.map +1 -0
- package/dist/description-length-gate.d.ts +119 -0
- package/dist/description-length-gate.d.ts.map +1 -0
- package/dist/detectors/edge.test.d.ts +2 -0
- package/dist/detectors/edge.test.d.ts.map +1 -0
- package/dist/detectors/index.d.ts +81 -0
- package/dist/detectors/index.d.ts.map +1 -0
- package/dist/detectors/index.test.d.ts +2 -0
- package/dist/detectors/index.test.d.ts.map +1 -0
- package/dist/diagnose/causal-sweep.d.ts +100 -0
- package/dist/diagnose/causal-sweep.d.ts.map +1 -0
- package/dist/diagnose/index.d.ts +36 -0
- package/dist/diagnose/index.d.ts.map +1 -0
- package/dist/diagnose/remediation.d.ts +68 -0
- package/dist/diagnose/remediation.d.ts.map +1 -0
- package/dist/diagnose/repair.d.ts +77 -0
- package/dist/diagnose/repair.d.ts.map +1 -0
- package/dist/diagnose.d.ts +1 -251
- package/dist/diagnose.js +2 -2
- package/dist/discover-personas.d.ts +35 -0
- package/dist/discover-personas.d.ts.map +1 -0
- package/dist/driver.d.ts +95 -0
- package/dist/driver.d.ts.map +1 -0
- package/dist/driver.test.d.ts +8 -0
- package/dist/driver.test.d.ts.map +1 -0
- package/dist/dual-agent-bench.d.ts +81 -0
- package/dist/dual-agent-bench.d.ts.map +1 -0
- package/dist/error-count-extractor.d.ts +47 -0
- package/dist/error-count-extractor.d.ts.map +1 -0
- package/dist/error-count-extractor.test.d.ts +2 -0
- package/dist/error-count-extractor.test.d.ts.map +1 -0
- package/dist/{errors-CzMUYo7b.d.ts → errors.d.ts} +10 -11
- package/dist/errors.d.ts.map +1 -0
- package/dist/{researcher-DE6Gpnb4.d.ts → eval-campaign.d.ts} +34 -156
- package/dist/eval-campaign.d.ts.map +1 -0
- package/dist/eval-campaign.test.d.ts +2 -0
- package/dist/eval-campaign.test.d.ts.map +1 -0
- package/dist/eval-tools.d.ts +55 -0
- package/dist/eval-tools.d.ts.map +1 -0
- package/dist/eval-trace-store.d.ts +107 -0
- package/dist/eval-trace-store.d.ts.map +1 -0
- package/dist/eval-trace-store.test.d.ts +2 -0
- package/dist/eval-trace-store.test.d.ts.map +1 -0
- package/dist/executor.d.ts +38 -0
- package/dist/executor.d.ts.map +1 -0
- package/dist/executor.test.d.ts +10 -0
- package/dist/executor.test.d.ts.map +1 -0
- package/dist/experiment-tracker.d.ts +178 -0
- package/dist/experiment-tracker.d.ts.map +1 -0
- package/dist/experiment-tracker.test.d.ts +2 -0
- package/dist/experiment-tracker.test.d.ts.map +1 -0
- package/dist/failure-taxonomy.d.ts +38 -0
- package/dist/failure-taxonomy.d.ts.map +1 -0
- package/dist/{feedback-trajectory-D9OVLrg9.d.ts → feedback-trajectory.d.ts} +36 -38
- package/dist/feedback-trajectory.d.ts.map +1 -0
- package/dist/feedback-trajectory.test.d.ts +2 -0
- package/dist/feedback-trajectory.test.d.ts.map +1 -0
- package/dist/flow-layer.d.ts +90 -0
- package/dist/flow-layer.d.ts.map +1 -0
- package/dist/flow-layer.test.d.ts +2 -0
- package/dist/flow-layer.test.d.ts.map +1 -0
- package/dist/fuzz/capsule.d.ts +46 -0
- package/dist/fuzz/capsule.d.ts.map +1 -0
- package/dist/fuzz/cube.d.ts +36 -0
- package/dist/fuzz/cube.d.ts.map +1 -0
- package/dist/fuzz/explorer-cost.test.d.ts +2 -0
- package/dist/fuzz/explorer-cost.test.d.ts.map +1 -0
- package/dist/fuzz/explorer.d.ts +64 -0
- package/dist/fuzz/explorer.d.ts.map +1 -0
- package/dist/fuzz/fuzz-agent.d.ts +16 -0
- package/dist/fuzz/fuzz-agent.d.ts.map +1 -0
- package/dist/fuzz/fuzz-agent.test.d.ts +2 -0
- package/dist/fuzz/fuzz-agent.test.d.ts.map +1 -0
- package/dist/fuzz/gates.d.ts +33 -0
- package/dist/fuzz/gates.d.ts.map +1 -0
- package/dist/fuzz/index.d.ts +26 -0
- package/dist/fuzz/index.d.ts.map +1 -0
- package/dist/fuzz/policies.d.ts +28 -0
- package/dist/fuzz/policies.d.ts.map +1 -0
- package/dist/fuzz/tools.d.ts +20 -0
- package/dist/fuzz/tools.d.ts.map +1 -0
- package/dist/fuzz/types.d.ts +307 -0
- package/dist/fuzz/types.d.ts.map +1 -0
- package/dist/fuzz.d.ts +1 -547
- package/dist/golden-matcher.d.ts +71 -0
- package/dist/golden-matcher.d.ts.map +1 -0
- package/dist/governance/eu-ai-act.d.ts +37 -0
- package/dist/governance/eu-ai-act.d.ts.map +1 -0
- package/dist/governance/index.d.ts +5 -135
- package/dist/governance/index.d.ts.map +1 -0
- package/dist/governance/nist-ai-rmf.d.ts +15 -0
- package/dist/governance/nist-ai-rmf.d.ts.map +1 -0
- package/dist/governance/soc2.d.ts +12 -0
- package/dist/governance/soc2.d.ts.map +1 -0
- package/dist/governance/types.d.ts +66 -0
- package/dist/governance/types.d.ts.map +1 -0
- package/dist/harness-optimizer.d.ts +82 -0
- package/dist/harness-optimizer.d.ts.map +1 -0
- package/dist/held-out-gate.d.ts +135 -0
- package/dist/held-out-gate.d.ts.map +1 -0
- package/dist/hosted/client.d.ts +73 -0
- package/dist/hosted/client.d.ts.map +1 -0
- package/dist/hosted/from-env.test.d.ts +8 -0
- package/dist/hosted/from-env.test.d.ts.map +1 -0
- package/dist/hosted/index.d.ts +10 -238
- package/dist/hosted/index.d.ts.map +1 -0
- package/dist/hosted/types.d.ts +159 -0
- package/dist/hosted/types.d.ts.map +1 -0
- package/dist/index.d.ts +277 -5643
- package/dist/index.d.ts.map +1 -0
- package/dist/index.js +317 -108
- package/dist/index.js.map +1 -1
- package/dist/integrity/backend-integrity.d.ts +71 -0
- package/dist/integrity/backend-integrity.d.ts.map +1 -0
- package/dist/integrity/preflight.d.ts +72 -0
- package/dist/integrity/preflight.d.ts.map +1 -0
- package/dist/integrity/preflight.test.d.ts +2 -0
- package/dist/integrity/preflight.test.d.ts.map +1 -0
- package/dist/integrity/single-backend.d.ts +67 -0
- package/dist/integrity/single-backend.d.ts.map +1 -0
- package/dist/intent-match-judge.d.ts +69 -0
- package/dist/intent-match-judge.d.ts.map +1 -0
- package/dist/intent-match-judge.test.d.ts +2 -0
- package/dist/intent-match-judge.test.d.ts.map +1 -0
- package/dist/{judge-calibration-0p2QcWNE.d.ts → judge-calibration.d.ts} +16 -17
- package/dist/judge-calibration.d.ts.map +1 -0
- package/dist/judge-ensemble.d.ts +66 -0
- package/dist/judge-ensemble.d.ts.map +1 -0
- package/dist/judge-ensemble.test.d.ts +8 -0
- package/dist/judge-ensemble.test.d.ts.map +1 -0
- package/dist/judge-families.d.ts +38 -0
- package/dist/judge-families.d.ts.map +1 -0
- package/dist/judge-panel.d.ts +65 -0
- package/dist/judge-panel.d.ts.map +1 -0
- package/dist/judge-retry.d.ts +70 -0
- package/dist/judge-retry.d.ts.map +1 -0
- package/dist/judge-runner.d.ts +36 -0
- package/dist/judge-runner.d.ts.map +1 -0
- package/dist/judge-runner.test.d.ts +2 -0
- package/dist/judge-runner.test.d.ts.map +1 -0
- package/dist/judges.d.ts +74 -0
- package/dist/judges.d.ts.map +1 -0
- package/dist/keyword-coverage-judge.d.ts +89 -0
- package/dist/keyword-coverage-judge.d.ts.map +1 -0
- package/dist/keyword-coverage-judge.test.d.ts +2 -0
- package/dist/keyword-coverage-judge.test.d.ts.map +1 -0
- package/dist/knowledge/index.d.ts +3 -103
- package/dist/knowledge/index.d.ts.map +1 -0
- package/dist/knowledge/readiness.d.ts +26 -0
- package/dist/knowledge/readiness.d.ts.map +1 -0
- package/dist/knowledge/types.d.ts +75 -0
- package/dist/knowledge/types.d.ts.map +1 -0
- package/dist/live-proof.d.ts +62 -0
- package/dist/live-proof.d.ts.map +1 -0
- package/dist/{llm-client-BeEcAokY.d.ts → llm-client.d.ts} +52 -23
- package/dist/llm-client.d.ts.map +1 -0
- package/dist/llm-client.test.d.ts +2 -0
- package/dist/llm-client.test.d.ts.map +1 -0
- package/dist/locked-jsonl-appender.d.ts +19 -0
- package/dist/locked-jsonl-appender.d.ts.map +1 -0
- package/dist/matrix/aggregation.d.ts +16 -0
- package/dist/matrix/aggregation.d.ts.map +1 -0
- package/dist/matrix/index.d.ts +12 -30
- package/dist/matrix/index.d.ts.map +1 -0
- package/dist/matrix/runner.d.ts +15 -0
- package/dist/matrix/runner.d.ts.map +1 -0
- package/dist/{types-mn5Aqk7x.d.ts → matrix/types.d.ts} +11 -15
- package/dist/matrix/types.d.ts.map +1 -0
- package/dist/meta-eval/calibration.d.ts +47 -0
- package/dist/meta-eval/calibration.d.ts.map +1 -0
- package/dist/meta-eval/correlation-study.d.ts +53 -0
- package/dist/meta-eval/correlation-study.d.ts.map +1 -0
- package/dist/meta-eval/index.d.ts +6 -181
- package/dist/meta-eval/index.d.ts.map +1 -0
- package/dist/meta-eval/index.js +1 -1
- package/dist/{outcome-store-rnXLEqSn.d.ts → meta-eval/outcome-store.d.ts} +7 -8
- package/dist/meta-eval/outcome-store.d.ts.map +1 -0
- package/dist/{rubric-predictive-validity-Cy_W-hWZ.d.ts → meta-eval/rubric-predictive-validity.d.ts} +8 -11
- package/dist/meta-eval/rubric-predictive-validity.d.ts.map +1 -0
- package/dist/meta-eval/sentinel.d.ts +169 -0
- package/dist/meta-eval/sentinel.d.ts.map +1 -0
- package/dist/metrics.d.ts +63 -0
- package/dist/metrics.d.ts.map +1 -0
- package/dist/model-seats.d.ts +71 -0
- package/dist/model-seats.d.ts.map +1 -0
- package/dist/model-seats.test.d.ts +2 -0
- package/dist/model-seats.test.d.ts.map +1 -0
- package/dist/muffled-gate-scanner.d.ts +102 -0
- package/dist/muffled-gate-scanner.d.ts.map +1 -0
- package/dist/{multi-layer-verifier-DUZXrPDA.d.ts → multi-layer-verifier.d.ts} +12 -15
- package/dist/multi-layer-verifier.d.ts.map +1 -0
- package/dist/multi-layer-verifier.test.d.ts +2 -0
- package/dist/multi-layer-verifier.test.d.ts.map +1 -0
- package/dist/multi-toolchain-layer.d.ts +80 -0
- package/dist/multi-toolchain-layer.d.ts.map +1 -0
- package/dist/multi-toolchain-layer.test.d.ts +2 -0
- package/dist/multi-toolchain-layer.test.d.ts.map +1 -0
- package/dist/multishot/default-tools.d.ts +34 -0
- package/dist/multishot/default-tools.d.ts.map +1 -0
- package/dist/multishot/index.d.ts +7 -290
- package/dist/multishot/index.d.ts.map +1 -0
- package/dist/multishot/index.js.map +1 -1
- package/dist/multishot/judges.d.ts +32 -0
- package/dist/multishot/judges.d.ts.map +1 -0
- package/dist/multishot/matrix.d.ts +107 -0
- package/dist/multishot/matrix.d.ts.map +1 -0
- package/dist/multishot/multishot.d.ts +23 -0
- package/dist/multishot/multishot.d.ts.map +1 -0
- package/dist/multishot/router.d.ts +37 -0
- package/dist/multishot/router.d.ts.map +1 -0
- package/dist/multishot/types.d.ts +60 -0
- package/dist/multishot/types.d.ts.map +1 -0
- package/dist/observability.d.ts +71 -0
- package/dist/observability.d.ts.map +1 -0
- package/dist/openapi.json +1 -1
- package/dist/oracle.d.ts +55 -0
- package/dist/oracle.d.ts.map +1 -0
- package/dist/orthogonality.d.ts +35 -0
- package/dist/orthogonality.d.ts.map +1 -0
- package/dist/otel-pipeline.d.ts +31 -0
- package/dist/otel-pipeline.d.ts.map +1 -0
- package/dist/paraphrase.d.ts +107 -0
- package/dist/paraphrase.d.ts.map +1 -0
- package/dist/{pareto-E-pembql.d.ts → pareto.d.ts} +9 -10
- package/dist/pareto.d.ts.map +1 -0
- package/dist/partition-held-out.d.ts +70 -0
- package/dist/partition-held-out.d.ts.map +1 -0
- package/dist/partition-held-out.test.d.ts +2 -0
- package/dist/partition-held-out.test.d.ts.map +1 -0
- package/dist/perf/index.d.ts +13 -119
- package/dist/perf/index.d.ts.map +1 -0
- package/dist/perf/integrity.d.ts +30 -0
- package/dist/perf/integrity.d.ts.map +1 -0
- package/dist/perf/journey.d.ts +45 -0
- package/dist/perf/journey.d.ts.map +1 -0
- package/dist/perf/ratchet.d.ts +47 -0
- package/dist/perf/ratchet.d.ts.map +1 -0
- package/dist/pipelines/budget-breach.d.ts +31 -0
- package/dist/pipelines/budget-breach.d.ts.map +1 -0
- package/dist/pipelines/budget-breach.test.d.ts +2 -0
- package/dist/pipelines/budget-breach.test.d.ts.map +1 -0
- package/dist/pipelines/failure-cluster.d.ts +38 -0
- package/dist/pipelines/failure-cluster.d.ts.map +1 -0
- package/dist/pipelines/failure-cluster.test.d.ts +2 -0
- package/dist/pipelines/failure-cluster.test.d.ts.map +1 -0
- package/dist/pipelines/first-divergence.d.ts +26 -0
- package/dist/pipelines/first-divergence.d.ts.map +1 -0
- package/dist/pipelines/first-divergence.test.d.ts +2 -0
- package/dist/pipelines/first-divergence.test.d.ts.map +1 -0
- package/dist/pipelines/index.d.ts +8 -173
- package/dist/pipelines/index.d.ts.map +1 -0
- package/dist/pipelines/index.js +40 -15
- package/dist/pipelines/index.js.map +1 -1
- package/dist/pipelines/judge-agreement.d.ts +26 -0
- package/dist/pipelines/judge-agreement.d.ts.map +1 -0
- package/dist/pipelines/judge-agreement.test.d.ts +2 -0
- package/dist/pipelines/judge-agreement.test.d.ts.map +1 -0
- package/dist/pipelines/regression.d.ts +23 -0
- package/dist/pipelines/regression.d.ts.map +1 -0
- package/dist/pipelines/regression.test.d.ts +2 -0
- package/dist/pipelines/regression.test.d.ts.map +1 -0
- package/dist/pipelines/stuck-loop.d.ts +32 -0
- package/dist/pipelines/stuck-loop.d.ts.map +1 -0
- package/dist/pipelines/stuck-loop.test.d.ts +2 -0
- package/dist/pipelines/stuck-loop.test.d.ts.map +1 -0
- package/dist/pipelines/tool-waste.d.ts +34 -0
- package/dist/pipelines/tool-waste.d.ts.map +1 -0
- package/dist/pipelines/tool-waste.test.d.ts +2 -0
- package/dist/pipelines/tool-waste.test.d.ts.map +1 -0
- package/dist/playbook.d.ts +16 -0
- package/dist/playbook.d.ts.map +1 -0
- package/dist/pr-review-benchmark.d.ts +88 -0
- package/dist/pr-review-benchmark.d.ts.map +1 -0
- package/dist/pr-review-benchmark.test.d.ts +2 -0
- package/dist/pr-review-benchmark.test.d.ts.map +1 -0
- package/dist/pre-registration.d.ts +125 -0
- package/dist/pre-registration.d.ts.map +1 -0
- package/dist/prm/builtin-rubrics.d.ts +33 -0
- package/dist/prm/builtin-rubrics.d.ts.map +1 -0
- package/dist/prm/index.d.ts +5 -100
- package/dist/prm/index.d.ts.map +1 -0
- package/dist/prm/index.js +55 -12
- package/dist/prm/index.js.map +1 -1
- package/dist/prm/inference.d.ts +29 -0
- package/dist/prm/inference.d.ts.map +1 -0
- package/dist/prm/inference.test.d.ts +2 -0
- package/dist/prm/inference.test.d.ts.map +1 -0
- package/dist/{rubric-BOfxn4ja.d.ts → prm/rubric.d.ts} +10 -13
- package/dist/prm/rubric.d.ts.map +1 -0
- package/dist/prm/training-export.d.ts +38 -0
- package/dist/prm/training-export.d.ts.map +1 -0
- package/dist/produced-state.d.ts +63 -0
- package/dist/produced-state.d.ts.map +1 -0
- package/dist/produced-state.test.d.ts +8 -0
- package/dist/produced-state.test.d.ts.map +1 -0
- package/dist/profile/baselines.d.ts +37 -0
- package/dist/profile/baselines.d.ts.map +1 -0
- package/dist/profile/index.d.ts +105 -0
- package/dist/profile/index.d.ts.map +1 -0
- package/dist/promotion-gate.d.ts +93 -0
- package/dist/promotion-gate.d.ts.map +1 -0
- package/dist/prompt-registry.d.ts +41 -0
- package/dist/prompt-registry.d.ts.map +1 -0
- package/dist/propose-review-control.d.ts +49 -0
- package/dist/propose-review-control.d.ts.map +1 -0
- package/dist/propose-review-control.test.d.ts +2 -0
- package/dist/propose-review-control.test.d.ts.map +1 -0
- package/dist/propose-review.d.ts +155 -0
- package/dist/propose-review.d.ts.map +1 -0
- package/dist/{red-team-BXHil6c8.d.ts → red-team.d.ts} +14 -16
- package/dist/red-team.d.ts.map +1 -0
- package/dist/reference-replay-steering.d.ts +11 -0
- package/dist/reference-replay-steering.d.ts.map +1 -0
- package/dist/reference-replay.d.ts +176 -0
- package/dist/reference-replay.d.ts.map +1 -0
- package/dist/reflective-mutation.d.ts +79 -0
- package/dist/reflective-mutation.d.ts.map +1 -0
- package/dist/registry.d.ts +31 -0
- package/dist/registry.d.ts.map +1 -0
- package/dist/release-confidence.d.ts +128 -0
- package/dist/release-confidence.d.ts.map +1 -0
- package/dist/release-report.d.ts +11 -0
- package/dist/release-report.d.ts.map +1 -0
- package/dist/replay.d.ts +120 -0
- package/dist/replay.d.ts.map +1 -0
- package/dist/reporter.d.ts +14 -0
- package/dist/reporter.d.ts.map +1 -0
- package/dist/reporting.d.ts +15 -15
- package/dist/reporting.d.ts.map +1 -0
- package/dist/reporting.js +3 -3
- package/dist/researcher.d.ts +140 -0
- package/dist/researcher.d.ts.map +1 -0
- package/dist/reviewer.d.ts +118 -0
- package/dist/reviewer.d.ts.map +1 -0
- package/dist/reviewer.test.d.ts +2 -0
- package/dist/reviewer.test.d.ts.map +1 -0
- package/dist/reward-model-export.d.ts +60 -0
- package/dist/reward-model-export.d.ts.map +1 -0
- package/dist/rl/active-curriculum.d.ts +110 -0
- package/dist/rl/active-curriculum.d.ts.map +1 -0
- package/dist/rl/adaptation-eval.d.ts +109 -0
- package/dist/rl/adaptation-eval.d.ts.map +1 -0
- package/dist/{adversarial-DIVcDoI_.d.ts → rl/adversarial.d.ts} +6 -7
- package/dist/rl/adversarial.d.ts.map +1 -0
- package/dist/rl/compute-curves.d.ts +127 -0
- package/dist/rl/compute-curves.d.ts.map +1 -0
- package/dist/rl/contamination.d.ts +117 -0
- package/dist/rl/contamination.d.ts.map +1 -0
- package/dist/rl/corpus.d.ts +55 -0
- package/dist/rl/corpus.d.ts.map +1 -0
- package/dist/rl/corpus.test.d.ts +2 -0
- package/dist/rl/corpus.test.d.ts.map +1 -0
- package/dist/rl/dataset.d.ts +102 -0
- package/dist/rl/dataset.d.ts.map +1 -0
- package/dist/rl/dataset.test.d.ts +2 -0
- package/dist/rl/dataset.test.d.ts.map +1 -0
- package/dist/rl/exporters.d.ts +141 -0
- package/dist/rl/exporters.d.ts.map +1 -0
- package/dist/rl/index.d.ts +49 -0
- package/dist/rl/index.d.ts.map +1 -0
- package/dist/{off-policy-DiwuKKg7.d.ts → rl/off-policy.d.ts} +8 -9
- package/dist/rl/off-policy.d.ts.map +1 -0
- package/dist/rl/predictive-validity-researcher.d.ts +69 -0
- package/dist/rl/predictive-validity-researcher.d.ts.map +1 -0
- package/dist/rl/preferences.d.ts +141 -0
- package/dist/rl/preferences.d.ts.map +1 -0
- package/dist/rl/process-reward.d.ts +122 -0
- package/dist/rl/process-reward.d.ts.map +1 -0
- package/dist/rl/reward-hacking.d.ts +104 -0
- package/dist/rl/reward-hacking.d.ts.map +1 -0
- package/dist/rl/rl-campaign.d.ts +85 -0
- package/dist/rl/rl-campaign.d.ts.map +1 -0
- package/dist/rl/run-record-adapters.d.ts +56 -0
- package/dist/rl/run-record-adapters.d.ts.map +1 -0
- package/dist/rl/sim-fidelity.d.ts +166 -0
- package/dist/rl/sim-fidelity.d.ts.map +1 -0
- package/dist/rl/sim-fidelity.test.d.ts +2 -0
- package/dist/rl/sim-fidelity.test.d.ts.map +1 -0
- package/dist/rl/tournament.d.ts +115 -0
- package/dist/rl/tournament.d.ts.map +1 -0
- package/dist/rl/verifiable-reward.d.ts +124 -0
- package/dist/rl/verifiable-reward.d.ts.map +1 -0
- package/dist/rl.d.ts +1 -1192
- package/dist/rl.js +615 -615
- package/dist/rl.js.map +1 -1
- package/dist/{run-campaign-RDGAM5KJ.js → run-campaign-WXY7KI67.js} +3 -3
- package/dist/run-critic.d.ts +23 -0
- package/dist/run-critic.d.ts.map +1 -0
- package/dist/run-evidence.d.ts +32 -0
- package/dist/run-evidence.d.ts.map +1 -0
- package/dist/{run-record-e7vj1uZQ.d.ts → run-record.d.ts} +16 -122
- package/dist/run-record.d.ts.map +1 -0
- package/dist/run-record.test.d.ts +2 -0
- package/dist/run-record.test.d.ts.map +1 -0
- package/dist/run-score.d.ts +31 -0
- package/dist/run-score.d.ts.map +1 -0
- package/dist/runtime-trajectory.d.ts +47 -0
- package/dist/runtime-trajectory.d.ts.map +1 -0
- package/dist/{test-graded-scenario-BdVaPyHT.d.ts → sandbox-harness.d.ts} +34 -60
- package/dist/sandbox-harness.d.ts.map +1 -0
- package/dist/sandbox-harness.test.d.ts +2 -0
- package/dist/sandbox-harness.test.d.ts.map +1 -0
- package/dist/sandbox-pool.d.ts +74 -0
- package/dist/sandbox-pool.d.ts.map +1 -0
- package/dist/sandbox-pool.test.d.ts +2 -0
- package/dist/sandbox-pool.test.d.ts.map +1 -0
- package/dist/scorecard.d.ts +133 -0
- package/dist/scorecard.d.ts.map +1 -0
- package/dist/scorecard.test.d.ts +2 -0
- package/dist/scorecard.test.d.ts.map +1 -0
- package/dist/self-play.d.ts +69 -0
- package/dist/self-play.d.ts.map +1 -0
- package/dist/semantic-concept-judge.d.ts +135 -0
- package/dist/semantic-concept-judge.d.ts.map +1 -0
- package/dist/semantic-concept-judge.test.d.ts +2 -0
- package/dist/semantic-concept-judge.test.d.ts.map +1 -0
- package/dist/{sequential-5iSVfzl2.d.ts → sequential.d.ts} +9 -10
- package/dist/sequential.d.ts.map +1 -0
- package/dist/{series-convergence-D5OWMBg6.d.ts → series-convergence.d.ts} +4 -5
- package/dist/series-convergence.d.ts.map +1 -0
- package/dist/slo.d.ts +48 -0
- package/dist/slo.d.ts.map +1 -0
- package/dist/state-continuity.d.ts +47 -0
- package/dist/state-continuity.d.ts.map +1 -0
- package/dist/{statistics-C7PozGrZ.d.ts → statistics.d.ts} +152 -39
- package/dist/statistics.d.ts.map +1 -0
- package/dist/statistics.test.d.ts +2 -0
- package/dist/statistics.test.d.ts.map +1 -0
- package/dist/steering-optimizer.d.ts +58 -0
- package/dist/steering-optimizer.d.ts.map +1 -0
- package/dist/steering.d.ts +24 -0
- package/dist/steering.d.ts.map +1 -0
- package/dist/storyboard/code-edit.d.ts +64 -0
- package/dist/storyboard/code-edit.d.ts.map +1 -0
- package/dist/storyboard/code-edit.test.d.ts +2 -0
- package/dist/storyboard/code-edit.test.d.ts.map +1 -0
- package/dist/storyboard/index.d.ts +16 -81
- package/dist/storyboard/index.d.ts.map +1 -0
- package/dist/storyboard/index.test.d.ts +2 -0
- package/dist/storyboard/index.test.d.ts.map +1 -0
- package/dist/{summary-report-DGmUucwQ.d.ts → summary-report.d.ts} +23 -160
- package/dist/summary-report.d.ts.map +1 -0
- package/dist/telemetry/client.d.ts +35 -0
- package/dist/telemetry/client.d.ts.map +1 -0
- package/dist/telemetry/index.d.ts +17 -35
- package/dist/telemetry/index.d.ts.map +1 -0
- package/dist/telemetry/schema.d.ts +61 -0
- package/dist/telemetry/schema.d.ts.map +1 -0
- package/dist/telemetry/sink-fetch.d.ts +39 -0
- package/dist/telemetry/sink-fetch.d.ts.map +1 -0
- package/dist/telemetry/{file.d.ts → sink-file.d.ts} +5 -7
- package/dist/telemetry/sink-file.d.ts.map +1 -0
- package/dist/test-graded-scenario.d.ts +42 -0
- package/dist/test-graded-scenario.d.ts.map +1 -0
- package/dist/testing.d.ts +5 -0
- package/dist/testing.d.ts.map +1 -0
- package/dist/testing.js +8 -0
- package/dist/testing.js.map +1 -0
- package/dist/tool-use-metrics.d.ts +35 -0
- package/dist/tool-use-metrics.d.ts.map +1 -0
- package/dist/trace/capture-fetch.d.ts +48 -0
- package/dist/trace/capture-fetch.d.ts.map +1 -0
- package/dist/trace/capture-fetch.test.d.ts +2 -0
- package/dist/trace/capture-fetch.test.d.ts.map +1 -0
- package/dist/{emitter-DEZwY14K.d.ts → trace/emitter.d.ts} +9 -12
- package/dist/trace/emitter.d.ts.map +1 -0
- package/dist/trace/extract-usage.d.ts +46 -0
- package/dist/trace/extract-usage.d.ts.map +1 -0
- package/dist/trace/extract-usage.test.d.ts +2 -0
- package/dist/trace/extract-usage.test.d.ts.map +1 -0
- package/dist/trace/index.d.ts +15 -0
- package/dist/trace/index.d.ts.map +1 -0
- package/dist/{integrity-VJ9A7aST.d.ts → trace/integrity.d.ts} +11 -14
- package/dist/trace/integrity.d.ts.map +1 -0
- package/dist/trace/otel-bridge.d.ts +29 -0
- package/dist/trace/otel-bridge.d.ts.map +1 -0
- package/dist/trace/otel-export.d.ts +52 -0
- package/dist/trace/otel-export.d.ts.map +1 -0
- package/dist/trace/otel.d.ts +57 -0
- package/dist/trace/otel.d.ts.map +1 -0
- package/dist/trace/otlp-attributes.d.ts +17 -0
- package/dist/trace/otlp-attributes.d.ts.map +1 -0
- package/dist/trace/query.d.ts +29 -0
- package/dist/trace/query.d.ts.map +1 -0
- package/dist/trace/query.test.d.ts +2 -0
- package/dist/trace/query.test.d.ts.map +1 -0
- package/dist/{raw-provider-sink-C46HDghv.d.ts → trace/raw-provider-sink.d.ts} +13 -14
- package/dist/trace/raw-provider-sink.d.ts.map +1 -0
- package/dist/{redact-B40YG2M_.d.ts → trace/redact.d.ts} +7 -8
- package/dist/trace/redact.d.ts.map +1 -0
- package/dist/{schema-m0gsnbt3.d.ts → trace/schema.d.ts} +29 -30
- package/dist/trace/schema.d.ts.map +1 -0
- package/dist/trace/store-to-otlp.d.ts +72 -0
- package/dist/trace/store-to-otlp.d.ts.map +1 -0
- package/dist/trace/store-to-otlp.test.d.ts +2 -0
- package/dist/trace/store-to-otlp.test.d.ts.map +1 -0
- package/dist/{store-CKUAgsJz.d.ts → trace/store.d.ts} +37 -13
- package/dist/trace/store.d.ts.map +1 -0
- package/dist/trace/store.test.d.ts +2 -0
- package/dist/trace/store.test.d.ts.map +1 -0
- package/dist/{analyst-C8HHvfJp.d.ts → trace-analyst/analyst.d.ts} +12 -14
- package/dist/trace-analyst/analyst.d.ts.map +1 -0
- package/dist/trace-analyst/analyst.test.d.ts +2 -0
- package/dist/trace-analyst/analyst.test.d.ts.map +1 -0
- package/dist/trace-analyst/behavioral-metrics.d.ts +40 -0
- package/dist/trace-analyst/behavioral-metrics.d.ts.map +1 -0
- package/dist/trace-analyst/behavioral-metrics.test.d.ts +2 -0
- package/dist/trace-analyst/behavioral-metrics.test.d.ts.map +1 -0
- package/dist/trace-analyst/hook.d.ts +55 -0
- package/dist/trace-analyst/hook.d.ts.map +1 -0
- package/dist/trace-analyst/index.d.ts +18 -0
- package/dist/trace-analyst/index.d.ts.map +1 -0
- package/dist/trace-analyst/insights.d.ts +71 -0
- package/dist/trace-analyst/insights.d.ts.map +1 -0
- package/dist/trace-analyst/insights.test.d.ts +2 -0
- package/dist/trace-analyst/insights.test.d.ts.map +1 -0
- package/dist/trace-analyst/otlp-flatten.d.ts +42 -0
- package/dist/trace-analyst/otlp-flatten.d.ts.map +1 -0
- package/dist/trace-analyst/otlp-span.d.ts +85 -0
- package/dist/trace-analyst/otlp-span.d.ts.map +1 -0
- package/dist/trace-analyst/otlp-span.test.d.ts +2 -0
- package/dist/trace-analyst/otlp-span.test.d.ts.map +1 -0
- package/dist/trace-analyst/otlp-to-run-records.d.ts +115 -0
- package/dist/trace-analyst/otlp-to-run-records.d.ts.map +1 -0
- package/dist/trace-analyst/otlp-to-run-records.test.d.ts +2 -0
- package/dist/trace-analyst/otlp-to-run-records.test.d.ts.map +1 -0
- package/dist/trace-analyst/otlp-to-run-records.timestamps.test.d.ts +2 -0
- package/dist/trace-analyst/otlp-to-run-records.timestamps.test.d.ts.map +1 -0
- package/dist/trace-analyst/prompts.d.ts +6 -0
- package/dist/trace-analyst/prompts.d.ts.map +1 -0
- package/dist/trace-analyst/store-otlp.d.ts +126 -0
- package/dist/trace-analyst/store-otlp.d.ts.map +1 -0
- package/dist/trace-analyst/store-otlp.test.d.ts +8 -0
- package/dist/trace-analyst/store-otlp.test.d.ts.map +1 -0
- package/dist/trace-analyst/store-otlp.timestamps.test.d.ts +2 -0
- package/dist/trace-analyst/store-otlp.timestamps.test.d.ts.map +1 -0
- package/dist/trace-analyst/store.d.ts +63 -0
- package/dist/trace-analyst/store.d.ts.map +1 -0
- package/dist/trace-analyst/tools.d.ts +44 -0
- package/dist/trace-analyst/tools.d.ts.map +1 -0
- package/dist/trace-analyst/tools.test.d.ts +10 -0
- package/dist/trace-analyst/tools.test.d.ts.map +1 -0
- package/dist/{store-C1YxJDEK.d.ts → trace-analyst/types.d.ts} +18 -74
- package/dist/trace-analyst/types.d.ts.map +1 -0
- package/dist/trace-contracts.d.ts +180 -0
- package/dist/trace-contracts.d.ts.map +1 -0
- package/dist/traced-analyst.d.ts +26 -0
- package/dist/traced-analyst.d.ts.map +1 -0
- package/dist/traced-judges.d.ts +27 -0
- package/dist/traced-judges.d.ts.map +1 -0
- package/dist/traces.d.ts +4 -947
- package/dist/traces.d.ts.map +1 -0
- package/dist/traces.js +42 -12
- package/dist/{trajectory-GEdXJCL5.d.ts → trajectory.d.ts} +8 -9
- package/dist/trajectory.d.ts.map +1 -0
- package/dist/{types-Croy5h7V.d.ts → types.d.ts} +33 -33
- package/dist/types.d.ts.map +1 -0
- package/dist/ui-finding.d.ts +104 -0
- package/dist/ui-finding.d.ts.map +1 -0
- package/dist/verdict-cache.d.ts +78 -0
- package/dist/verdict-cache.d.ts.map +1 -0
- package/dist/verdict-cache.test.d.ts +2 -0
- package/dist/verdict-cache.test.d.ts.map +1 -0
- package/dist/{verdict-C9MlYujm.d.ts → verdict.d.ts} +2 -3
- package/dist/verdict.d.ts.map +1 -0
- package/dist/visual-diff.d.ts +32 -0
- package/dist/visual-diff.d.ts.map +1 -0
- package/dist/wire/handlers.d.ts +54 -0
- package/dist/wire/handlers.d.ts.map +1 -0
- package/dist/wire/index.d.ts +13 -570
- package/dist/wire/index.d.ts.map +1 -0
- package/dist/wire/index.js +2 -2
- package/dist/wire/openapi.d.ts +3 -0
- package/dist/wire/openapi.d.ts.map +1 -0
- package/dist/wire/rpc.d.ts +21 -0
- package/dist/wire/rpc.d.ts.map +1 -0
- package/dist/wire/rubrics.d.ts +34 -0
- package/dist/wire/rubrics.d.ts.map +1 -0
- package/dist/wire/schemas.d.ts +410 -0
- package/dist/wire/schemas.d.ts.map +1 -0
- package/dist/wire/server.d.ts +60 -0
- package/dist/wire/server.d.ts.map +1 -0
- package/dist/workflow/event-schema.d.ts +5 -0
- package/dist/workflow/event-schema.d.ts.map +1 -0
- package/dist/workflow/feedback-pack.d.ts +99 -0
- package/dist/workflow/feedback-pack.d.ts.map +1 -0
- package/dist/workflow/index.d.ts +22 -495
- package/dist/workflow/index.d.ts.map +1 -0
- package/dist/workflow/index.js +2 -2
- package/dist/workflow/intelligence-export.d.ts +62 -0
- package/dist/workflow/intelligence-export.d.ts.map +1 -0
- package/dist/workflow/partner-report.d.ts +49 -0
- package/dist/workflow/partner-report.d.ts.map +1 -0
- package/dist/workflow/phase-graph.d.ts +43 -0
- package/dist/workflow/phase-graph.d.ts.map +1 -0
- package/dist/workflow/promotion-gate.d.ts +61 -0
- package/dist/workflow/promotion-gate.d.ts.map +1 -0
- package/dist/workflow/run-record.d.ts +12 -0
- package/dist/workflow/run-record.d.ts.map +1 -0
- package/dist/workflow/runtime-adapter.d.ts +20 -0
- package/dist/workflow/runtime-adapter.d.ts.map +1 -0
- package/dist/workflow/sanitize.d.ts +21 -0
- package/dist/workflow/sanitize.d.ts.map +1 -0
- package/dist/workflow/schema.d.ts +5 -0
- package/dist/workflow/schema.d.ts.map +1 -0
- package/dist/workflow/summary.d.ts +43 -0
- package/dist/workflow/summary.d.ts.map +1 -0
- package/dist/workflow/trace-event-fields.d.ts +6 -0
- package/dist/workflow/trace-event-fields.d.ts.map +1 -0
- package/dist/workflow/trajectory.d.ts +15 -0
- package/dist/workflow/trajectory.d.ts.map +1 -0
- package/dist/workflow/types.d.ts +68 -0
- package/dist/workflow/types.d.ts.map +1 -0
- package/dist/workspace-inspector.d.ts +67 -0
- package/dist/workspace-inspector.d.ts.map +1 -0
- package/dist/wrangler-deploy-runner.test.d.ts +2 -0
- package/dist/wrangler-deploy-runner.test.d.ts.map +1 -0
- package/docs/campaign-proposers.md +170 -0
- package/docs/concepts.md +8 -4
- package/docs/customer-journeys.md +15 -13
- package/docs/design/loop-taxonomy.md +34 -66
- package/docs/distributed-driver.md +14 -14
- package/docs/feature-guide.md +1 -1
- package/docs/hosted-ingest-spec.md +2 -3
- package/docs/multi-shot-optimization.md +8 -8
- package/docs/product-eval-adoption.md +1 -1
- package/docs/self-improvement-map.md +33 -29
- package/package.json +9 -15
- package/dist/calibration-Cpr3WaX3.d.ts +0 -101
- package/dist/chunk-47X6LRCE.js.map +0 -1
- package/dist/chunk-6SOJM3VR.js.map +0 -1
- package/dist/chunk-CVVHBFGN.js.map +0 -1
- package/dist/chunk-D3V5B42D.js.map +0 -1
- package/dist/chunk-E4GH6USR.js.map +0 -1
- package/dist/chunk-GSH6QNNS.js.map +0 -1
- package/dist/chunk-KWRRMR3J.js.map +0 -1
- package/dist/chunk-LMZQ2Z4U.js.map +0 -1
- package/dist/chunk-QAY5UIJO.js.map +0 -1
- package/dist/chunk-QG2OVF2D.js.map +0 -1
- package/dist/chunk-T375SUOZ.js.map +0 -1
- package/dist/chunk-TWS7AZEY.js.map +0 -1
- package/dist/chunk-UMMZHCPB.js.map +0 -1
- package/dist/chunk-XY4DDNEG.js.map +0 -1
- package/dist/chunk-ZFIBGEOL.js.map +0 -1
- package/dist/control-_Qb7skHX.d.ts +0 -259
- package/dist/corpus-BoR-041R.d.ts +0 -560
- package/dist/failure-cluster-CL7IVgkJ.d.ts +0 -76
- package/dist/harness-optimizer-EnEnQPsr.d.ts +0 -106
- package/dist/index-Bx3gZ8xl.d.ts +0 -159
- package/dist/pre-registration-mAnCugl9.d.ts +0 -523
- package/dist/provenance-LnqRT0sS.d.ts +0 -441
- package/dist/query-CqTxMwDw.d.ts +0 -31
- package/dist/release-report-euXIV_Sk.d.ts +0 -233
- package/dist/run-critic-BAIjX99r.d.ts +0 -56
- package/dist/run-improvement-loop-5z_l5zDz.d.ts +0 -427
- package/dist/runtime-trajectory-BDgfGZSr.d.ts +0 -49
- package/dist/semantic-concept-judge-Dn8Z6KEG.d.ts +0 -624
- package/dist/sink-fetch-B1Yg4Til.d.ts +0 -101
- package/docs/design/external-agent-wedge.md +0 -89
- package/docs/design/phase-d-rfc.md +0 -125
- package/docs/design/phase4-consumer-migration.md +0 -70
- package/docs/design/primitives-integration-spec.md +0 -393
- package/docs/design/product-self-improvement-loop.md +0 -146
- package/docs/design/self-improvement-engine.md +0 -140
- package/docs/design/self-improvement-protocol.md +0 -223
- package/docs/design/self-improvement-roadmap.md +0 -106
- package/docs/design/substrate-gaps.md +0 -118
- package/docs/phase-b-pairing-kit.md +0 -188
- package/docs/phase-b-runbook.md +0 -176
- package/docs/pilot/README.md +0 -62
- package/docs/pilot/customer-checklist.md +0 -90
- package/docs/pilot/integration-foreign-stack.md +0 -296
- package/docs/pilot/integration-tangle-stack.md +0 -248
- package/docs/pilot/one-pager.md +0 -161
- package/docs/pilot/sample-insight-report.json +0 -172
- package/docs/quickstart-external.md +0 -229
- package/docs/research/belief-state-agent-eval-roadmap.md +0 -593
- package/docs/research/research-roadmap.md +0 -205
- package/docs/specs/driver-honest-spec.md +0 -251
- package/docs/specs/hermes-self-improvement-audit.md +0 -93
- package/docs/specs/profile-versioning.md +0 -291
- package/docs/three-package-architecture.md +0 -168
- /package/dist/{chunk-QS3RBQPI.js.map → chunk-D5KBVULY.js.map} +0 -0
- /package/dist/{chunk-UHMJT4T7.js.map → chunk-IZCEK2HR.js.map} +0 -0
- /package/dist/{chunk-L3JOU6XM.js.map → chunk-NACM5RS7.js.map} +0 -0
- /package/dist/{chunk-4FBZZIYD.js.map → chunk-QTMYW64F.js.map} +0 -0
- /package/dist/{chunk-CY6U5S3X.js.map → chunk-URWVKRCS.js.map} +0 -0
- /package/dist/{chunk-Y47J2LJ3.js.map → chunk-Z6L6YSU6.js.map} +0 -0
- /package/dist/{run-campaign-RDGAM5KJ.js.map → run-campaign-WXY7KI67.js.map} +0 -0
|
@@ -0,0 +1,178 @@
|
|
|
1
|
+
/**
|
|
2
|
+
* Experiment tracker — git-provenanced experiment log with N-rep stats and a
|
|
3
|
+
* KEEP / REGRESSION / NOISE verdict against a parent.
|
|
4
|
+
*
|
|
5
|
+
* Every loop the fleet runs reduces to the same question: "I ran the candidate
|
|
6
|
+
* N times — is the median measurably better than the parent, or is the delta
|
|
7
|
+
* inside the noise band?" The hand-rolled copies bake a fixed score scale
|
|
8
|
+
* (percentage points), a fixed store path (`.evolve/experiments-v2.json`), and
|
|
9
|
+
* `execSync('git …')` straight into the module. This is the canonical version:
|
|
10
|
+
* provenance and persistence are injected, thresholds are configurable, and the
|
|
11
|
+
* stats + verdict are pure functions you can unit-test without a git repo or a
|
|
12
|
+
* filesystem.
|
|
13
|
+
*
|
|
14
|
+
* Stats per experiment: median / mean / min / max / iqr / stddev / passRate /
|
|
15
|
+
* n, plus a `stable` flag (`iqr < iqrUnstableAbove && stddev < stddevUnstableAbove`).
|
|
16
|
+
*
|
|
17
|
+
* Verdict against a parent (both must have `n >= minRepsForVerdict`):
|
|
18
|
+
* - NOISE — the candidate is too unstable to judge (`!stable`)
|
|
19
|
+
* - KEEP — `medianDelta > keepThreshold`
|
|
20
|
+
* - REGRESSION — `medianDelta < -regressionThreshold`
|
|
21
|
+
* - NOISE — otherwise (delta inside the band)
|
|
22
|
+
* With no parent (or insufficient reps) the verdict is the neutral ITERATE.
|
|
23
|
+
*/
|
|
24
|
+
/** Verdict for one experiment relative to its parent. ITERATE is the neutral
|
|
25
|
+
* "keep collecting reps / no parent to compare against" state. */
|
|
26
|
+
export type ExperimentVerdict = 'KEEP' | 'ITERATE' | 'NOISE' | 'REGRESSION';
|
|
27
|
+
/** Git provenance for the working tree an experiment was run from. */
|
|
28
|
+
export interface ExperimentProvenance {
|
|
29
|
+
/** Commit sha (short or full — the tracker does not interpret it). */
|
|
30
|
+
commit: string;
|
|
31
|
+
/** First line of the commit message. */
|
|
32
|
+
message: string;
|
|
33
|
+
/** Files changed vs the parent commit, or a marker like 'uncommitted'. */
|
|
34
|
+
changedFiles: string[];
|
|
35
|
+
}
|
|
36
|
+
/** A single repetition of an experiment, carrying the score the verdict is
|
|
37
|
+
* computed on plus any free-form per-rep metrics the consumer wants kept. */
|
|
38
|
+
export interface ExperimentRep {
|
|
39
|
+
/** 0-indexed repetition number within the experiment. */
|
|
40
|
+
rep: number;
|
|
41
|
+
/** The score this rep is judged on (same scale as the thresholds). */
|
|
42
|
+
score: number;
|
|
43
|
+
/** ISO timestamp the rep completed. */
|
|
44
|
+
timestamp: string;
|
|
45
|
+
/** Whether this rep passed the consumer's own gate — folded into `passRate`. */
|
|
46
|
+
passed?: boolean;
|
|
47
|
+
/** Free-form numeric metrics retained for later analysis. */
|
|
48
|
+
metrics?: Record<string, number>;
|
|
49
|
+
}
|
|
50
|
+
export interface ExperimentStats {
|
|
51
|
+
median: number;
|
|
52
|
+
mean: number;
|
|
53
|
+
min: number;
|
|
54
|
+
max: number;
|
|
55
|
+
/** Inter-quartile range of the rep scores. */
|
|
56
|
+
iqr: number;
|
|
57
|
+
/** Population standard deviation of the rep scores. */
|
|
58
|
+
stddev: number;
|
|
59
|
+
/** Fraction of reps with `passed === true`, over reps that set `passed`.
|
|
60
|
+
* null when no rep declared a pass/fail outcome. */
|
|
61
|
+
passRate: number | null;
|
|
62
|
+
/** Number of reps. */
|
|
63
|
+
n: number;
|
|
64
|
+
/** True when the sample is tight enough to trust for a verdict. */
|
|
65
|
+
stable: boolean;
|
|
66
|
+
}
|
|
67
|
+
export interface Experiment {
|
|
68
|
+
/** Stable id for the experiment. */
|
|
69
|
+
id: string;
|
|
70
|
+
/** Free-form label / config descriptor. */
|
|
71
|
+
label: string;
|
|
72
|
+
/** Git provenance captured when the experiment was created. */
|
|
73
|
+
provenance: ExperimentProvenance;
|
|
74
|
+
/** Parent experiment id this candidate is compared against, if any. */
|
|
75
|
+
parentId?: string;
|
|
76
|
+
/** One-line summary of what changed from the parent. */
|
|
77
|
+
changeSummary: string;
|
|
78
|
+
reps: ExperimentRep[];
|
|
79
|
+
stats: ExperimentStats;
|
|
80
|
+
verdict: ExperimentVerdict;
|
|
81
|
+
/** ISO timestamp the experiment was created. */
|
|
82
|
+
createdAt: string;
|
|
83
|
+
}
|
|
84
|
+
export interface ImprovementThresholds {
|
|
85
|
+
/** medianDelta strictly above this ⇒ KEEP. Default 5. */
|
|
86
|
+
keepThreshold?: number;
|
|
87
|
+
/** medianDelta strictly below the negative of this ⇒ REGRESSION. Default 5. */
|
|
88
|
+
regressionThreshold?: number;
|
|
89
|
+
/** iqr at or above this ⇒ unstable. Default 10. */
|
|
90
|
+
iqrUnstableAbove?: number;
|
|
91
|
+
/** stddev at or above this ⇒ unstable. Default Infinity (iqr-only stability). */
|
|
92
|
+
stddevUnstableAbove?: number;
|
|
93
|
+
/** Reps required on BOTH candidate and parent before a verdict is rendered.
|
|
94
|
+
* Default 3. */
|
|
95
|
+
minRepsForVerdict?: number;
|
|
96
|
+
}
|
|
97
|
+
export interface ImprovementVerdictResult {
|
|
98
|
+
verdict: ExperimentVerdict;
|
|
99
|
+
/** candidate.median − parent.median; null when no parent or insufficient reps. */
|
|
100
|
+
medianDelta: number | null;
|
|
101
|
+
/** Human-readable reason for the verdict — for dashboards and logs. */
|
|
102
|
+
reason: string;
|
|
103
|
+
}
|
|
104
|
+
/**
|
|
105
|
+
* Compute the N-rep statistics for a set of reps. Pure — no I/O. The `stable`
|
|
106
|
+
* flag is the trust gate the verdict depends on: a sample whose spread exceeds
|
|
107
|
+
* the configured bounds can't distinguish a real delta from run-to-run noise.
|
|
108
|
+
*/
|
|
109
|
+
export declare function computeExperimentStats(reps: ExperimentRep[], thresholds?: ImprovementThresholds): ExperimentStats;
|
|
110
|
+
/**
|
|
111
|
+
* Verdict for a candidate against its parent. Pure — operates on already-computed
|
|
112
|
+
* stats. KEEP/REGRESSION require both sides to have `>= minRepsForVerdict` reps
|
|
113
|
+
* AND the candidate to be `stable`; otherwise the result is NOISE (unstable) or
|
|
114
|
+
* ITERATE (not enough reps / no parent).
|
|
115
|
+
*/
|
|
116
|
+
export declare function improvementVerdict(candidate: ExperimentStats, parent: ExperimentStats | null, thresholds?: ImprovementThresholds): ImprovementVerdictResult;
|
|
117
|
+
/** Reads git provenance for the working tree. Inject a fake in tests; the
|
|
118
|
+
* default implementation shells out to `git`. */
|
|
119
|
+
export type ProvenanceReader = () => ExperimentProvenance | Promise<ExperimentProvenance>;
|
|
120
|
+
/** Persistence seam for the experiment log. Inject in-memory in tests; the
|
|
121
|
+
* filesystem implementation is `fileExperimentStore`. */
|
|
122
|
+
export interface ExperimentStore {
|
|
123
|
+
load(): Promise<Experiment[]>;
|
|
124
|
+
save(experiments: Experiment[]): Promise<void>;
|
|
125
|
+
}
|
|
126
|
+
/**
|
|
127
|
+
* Default provenance reader: `git rev-parse HEAD`, the subject line, and the
|
|
128
|
+
* files changed vs `HEAD~1`. Fail-loud — a tracker that silently logs
|
|
129
|
+
* `commit: 'unknown'` corrupts the provenance the whole point of the log is to
|
|
130
|
+
* carry. When the working tree genuinely has no parent commit, pass an override.
|
|
131
|
+
*/
|
|
132
|
+
export declare const gitProvenanceReader: ProvenanceReader;
|
|
133
|
+
/** In-memory store — the default when no persistence is wanted (tests, ephemeral
|
|
134
|
+
* runs). State lives on the instance. */
|
|
135
|
+
export declare function inMemoryExperimentStore(initial?: Experiment[]): ExperimentStore;
|
|
136
|
+
/** Filesystem store — a single JSON array at `path`, created on first save. */
|
|
137
|
+
export declare function fileExperimentStore(path: string): ExperimentStore;
|
|
138
|
+
export interface ExperimentTrackerOptions {
|
|
139
|
+
store?: ExperimentStore;
|
|
140
|
+
provenanceReader?: ProvenanceReader;
|
|
141
|
+
thresholds?: ImprovementThresholds;
|
|
142
|
+
/** Clock seam for deterministic timestamps in tests. Default `Date.now`. */
|
|
143
|
+
now?: () => number;
|
|
144
|
+
}
|
|
145
|
+
export interface CreateExperimentInput {
|
|
146
|
+
id: string;
|
|
147
|
+
label: string;
|
|
148
|
+
changeSummary: string;
|
|
149
|
+
parentId?: string;
|
|
150
|
+
/** Override provenance instead of reading from git (e.g. CI metadata). */
|
|
151
|
+
provenance?: ExperimentProvenance;
|
|
152
|
+
}
|
|
153
|
+
/**
|
|
154
|
+
* Stateful tracker over an `ExperimentStore`. Create an experiment (provenance
|
|
155
|
+
* is captured once), append reps as they complete (stats + verdict recompute on
|
|
156
|
+
* every append), and read the log back for a dashboard. All persistence and git
|
|
157
|
+
* access flow through the injected seams, so the tracker is fully testable
|
|
158
|
+
* without a repo or disk.
|
|
159
|
+
*/
|
|
160
|
+
export declare class ExperimentTracker {
|
|
161
|
+
private readonly store;
|
|
162
|
+
private readonly provenanceReader;
|
|
163
|
+
private readonly thresholds;
|
|
164
|
+
private readonly now;
|
|
165
|
+
constructor(options?: ExperimentTrackerOptions);
|
|
166
|
+
create(input: CreateExperimentInput): Promise<Experiment>;
|
|
167
|
+
/** Append a rep (its `rep` index defaults to the current rep count) and
|
|
168
|
+
* recompute stats + verdict. Returns the updated experiment. */
|
|
169
|
+
addRep(experimentId: string, rep: Omit<ExperimentRep, 'rep' | 'timestamp'> & {
|
|
170
|
+
rep?: number;
|
|
171
|
+
timestamp?: string;
|
|
172
|
+
}): Promise<Experiment>;
|
|
173
|
+
get(experimentId: string): Promise<Experiment | undefined>;
|
|
174
|
+
list(): Promise<Experiment[]>;
|
|
175
|
+
/** Full verdict (not just the enum) for an experiment vs its parent. */
|
|
176
|
+
verdictFor(experimentId: string): Promise<ImprovementVerdictResult>;
|
|
177
|
+
}
|
|
178
|
+
//# sourceMappingURL=experiment-tracker.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"experiment-tracker.d.ts","sourceRoot":"","sources":["../src/experiment-tracker.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;GAsBG;AAMH;mEACmE;AACnE,MAAM,MAAM,iBAAiB,GAAG,MAAM,GAAG,SAAS,GAAG,OAAO,GAAG,YAAY,CAAA;AAE3E,sEAAsE;AACtE,MAAM,WAAW,oBAAoB;IACnC,sEAAsE;IACtE,MAAM,EAAE,MAAM,CAAA;IACd,wCAAwC;IACxC,OAAO,EAAE,MAAM,CAAA;IACf,0EAA0E;IAC1E,YAAY,EAAE,MAAM,EAAE,CAAA;CACvB;AAED;8EAC8E;AAC9E,MAAM,WAAW,aAAa;IAC5B,yDAAyD;IACzD,GAAG,EAAE,MAAM,CAAA;IACX,sEAAsE;IACtE,KAAK,EAAE,MAAM,CAAA;IACb,uCAAuC;IACvC,SAAS,EAAE,MAAM,CAAA;IACjB,gFAAgF;IAChF,MAAM,CAAC,EAAE,OAAO,CAAA;IAChB,6DAA6D;IAC7D,OAAO,CAAC,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,CAAA;CACjC;AAED,MAAM,WAAW,eAAe;IAC9B,MAAM,EAAE,MAAM,CAAA;IACd,IAAI,EAAE,MAAM,CAAA;IACZ,GAAG,EAAE,MAAM,CAAA;IACX,GAAG,EAAE,MAAM,CAAA;IACX,8CAA8C;IAC9C,GAAG,EAAE,MAAM,CAAA;IACX,uDAAuD;IACvD,MAAM,EAAE,MAAM,CAAA;IACd;yDACqD;IACrD,QAAQ,EAAE,MAAM,GAAG,IAAI,CAAA;IACvB,sBAAsB;IACtB,CAAC,EAAE,MAAM,CAAA;IACT,mEAAmE;IACnE,MAAM,EAAE,OAAO,CAAA;CAChB;AAED,MAAM,WAAW,UAAU;IACzB,oCAAoC;IACpC,EAAE,EAAE,MAAM,CAAA;IACV,2CAA2C;IAC3C,KAAK,EAAE,MAAM,CAAA;IACb,+DAA+D;IAC/D,UAAU,EAAE,oBAAoB,CAAA;IAChC,uEAAuE;IACvE,QAAQ,CAAC,EAAE,MAAM,CAAA;IACjB,wDAAwD;IACxD,aAAa,EAAE,MAAM,CAAA;IACrB,IAAI,EAAE,aAAa,EAAE,CAAA;IACrB,KAAK,EAAE,eAAe,CAAA;IACtB,OAAO,EAAE,iBAAiB,CAAA;IAC1B,gDAAgD;IAChD,SAAS,EAAE,MAAM,CAAA;CAClB;AAED,MAAM,WAAW,qBAAqB;IACpC,yDAAyD;IACzD,aAAa,CAAC,EAAE,MAAM,CAAA;IACtB,+EAA+E;IAC/E,mBAAmB,CAAC,EAAE,MAAM,CAAA;IAC5B,mDAAmD;IACnD,gBAAgB,CAAC,EAAE,MAAM,CAAA;IACzB,iFAAiF;IACjF,mBAAmB,CAAC,EAAE,MAAM,CAAA;IAC5B;qBACiB;IACjB,iBAAiB,CAAC,EAAE,MAAM,CAAA;CAC3B;AAED,MAAM,WAAW,wBAAwB;IACvC,OAAO,EAAE,iBAAiB,CAAA;IAC1B,kFAAkF;IAClF,WAAW,EAAE,MAAM,GAAG,IAAI,CAAA;IAC1B,uEAAuE;IACvE,MAAM,EAAE,MAAM,CAAA;CACf;AA4CD;;;;GAIG;AACH,wBAAgB,sBAAsB,CACpC,IAAI,EAAE,aAAa,EAAE,EACrB,UAAU,CAAC,EAAE,qBAAqB,GACjC,eAAe,CAwCjB;AAED;;;;;GAKG;AACH,wBAAgB,kBAAkB,CAChC,SAAS,EAAE,eAAe,EAC1B,MAAM,EAAE,eAAe,GAAG,IAAI,EAC9B,UAAU,CAAC,EAAE,qBAAqB,GACjC,wBAAwB,CAuC1B;AAID;kDACkD;AAClD,MAAM,MAAM,gBAAgB,GAAG,MAAM,oBAAoB,GAAG,OAAO,CAAC,oBAAoB,CAAC,CAAA;AAEzF;0DAC0D;AAC1D,MAAM,WAAW,eAAe;IAC9B,IAAI,IAAI,OAAO,CAAC,UAAU,EAAE,CAAC,CAAA;IAC7B,IAAI,CAAC,WAAW,EAAE,UAAU,EAAE,GAAG,OAAO,CAAC,IAAI,CAAC,CAAA;CAC/C;AAED;;;;;GAKG;AACH,eAAO,MAAM,mBAAmB,EAAE,gBAOjC,CAAA;AAED;0CAC0C;AAC1C,wBAAgB,uBAAuB,CAAC,OAAO,GAAE,UAAU,EAAO,GAAG,eAAe,CAUnF;AAED,+EAA+E;AAC/E,wBAAgB,mBAAmB,CAAC,IAAI,EAAE,MAAM,GAAG,eAAe,CAuBjE;AAED,MAAM,WAAW,wBAAwB;IACvC,KAAK,CAAC,EAAE,eAAe,CAAA;IACvB,gBAAgB,CAAC,EAAE,gBAAgB,CAAA;IACnC,UAAU,CAAC,EAAE,qBAAqB,CAAA;IAClC,4EAA4E;IAC5E,GAAG,CAAC,EAAE,MAAM,MAAM,CAAA;CACnB;AAED,MAAM,WAAW,qBAAqB;IACpC,EAAE,EAAE,MAAM,CAAA;IACV,KAAK,EAAE,MAAM,CAAA;IACb,aAAa,EAAE,MAAM,CAAA;IACrB,QAAQ,CAAC,EAAE,MAAM,CAAA;IACjB,0EAA0E;IAC1E,UAAU,CAAC,EAAE,oBAAoB,CAAA;CAClC;AAED;;;;;;GAMG;AACH,qBAAa,iBAAiB;IAC5B,OAAO,CAAC,QAAQ,CAAC,KAAK,CAAiB;IACvC,OAAO,CAAC,QAAQ,CAAC,gBAAgB,CAAkB;IACnD,OAAO,CAAC,QAAQ,CAAC,UAAU,CAAiC;IAC5D,OAAO,CAAC,QAAQ,CAAC,GAAG,CAAc;gBAEtB,OAAO,GAAE,wBAA6B;IAO5C,MAAM,CAAC,KAAK,EAAE,qBAAqB,GAAG,OAAO,CAAC,UAAU,CAAC;IA2B/D;qEACiE;IAC3D,MAAM,CACV,YAAY,EAAE,MAAM,EACpB,GAAG,EAAE,IAAI,CAAC,aAAa,EAAE,KAAK,GAAG,WAAW,CAAC,GAAG;QAAE,GAAG,CAAC,EAAE,MAAM,CAAC;QAAC,SAAS,CAAC,EAAE,MAAM,CAAA;KAAE,GACnF,OAAO,CAAC,UAAU,CAAC;IAoBhB,GAAG,CAAC,YAAY,EAAE,MAAM,GAAG,OAAO,CAAC,UAAU,GAAG,SAAS,CAAC;IAM1D,IAAI,IAAI,OAAO,CAAC,UAAU,EAAE,CAAC;IAInC,wEAAwE;IAClE,UAAU,CAAC,YAAY,EAAE,MAAM,GAAG,OAAO,CAAC,wBAAwB,CAAC;CAQ1E"}
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"experiment-tracker.test.d.ts","sourceRoot":"","sources":["../src/experiment-tracker.test.ts"],"names":[],"mappings":""}
|
|
@@ -0,0 +1,38 @@
|
|
|
1
|
+
/**
|
|
2
|
+
* Failure taxonomy — canonical classes + a default classifier.
|
|
3
|
+
*
|
|
4
|
+
* Every failed run should end up in a named class. The classifier here
|
|
5
|
+
* is rule-based (fast, deterministic); an LLM fallback can be added by
|
|
6
|
+
* the consumer for novel cases and trained into the rule base over time.
|
|
7
|
+
*
|
|
8
|
+
* Consumers call `classifyFailure(run, spans, events)` and persist the
|
|
9
|
+
* returned class as `Run.outcome.failureClass`.
|
|
10
|
+
*/
|
|
11
|
+
import type { FailureClass, Run, Span, TraceEvent } from './trace/schema';
|
|
12
|
+
import { FAILURE_CLASSES } from './trace/schema';
|
|
13
|
+
export { FAILURE_CLASSES, type FailureClass };
|
|
14
|
+
export interface FailureContext {
|
|
15
|
+
run: Run;
|
|
16
|
+
spans: Span[];
|
|
17
|
+
events: TraceEvent[];
|
|
18
|
+
}
|
|
19
|
+
export interface FailureClassification {
|
|
20
|
+
failureClass: FailureClass;
|
|
21
|
+
reason: string;
|
|
22
|
+
triggerSpanId?: string;
|
|
23
|
+
triggerEventId?: string;
|
|
24
|
+
}
|
|
25
|
+
/** Ordered rules — first match wins. */
|
|
26
|
+
export interface FailureRule {
|
|
27
|
+
id: string;
|
|
28
|
+
match: (ctx: FailureContext) => {
|
|
29
|
+
failureClass: FailureClass;
|
|
30
|
+
reason: string;
|
|
31
|
+
triggerSpanId?: string;
|
|
32
|
+
triggerEventId?: string;
|
|
33
|
+
} | null;
|
|
34
|
+
}
|
|
35
|
+
export declare const DEFAULT_RULES: FailureRule[];
|
|
36
|
+
/** Classify the failure mode of a run using an ordered rule list. */
|
|
37
|
+
export declare function classifyFailure(ctx: FailureContext, rules?: FailureRule[]): FailureClassification;
|
|
38
|
+
//# sourceMappingURL=failure-taxonomy.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"failure-taxonomy.d.ts","sourceRoot":"","sources":["../src/failure-taxonomy.ts"],"names":[],"mappings":"AAAA;;;;;;;;;GASG;AAEH,OAAO,KAAK,EAAE,YAAY,EAAE,GAAG,EAAE,IAAI,EAAE,UAAU,EAAE,MAAM,gBAAgB,CAAA;AACzE,OAAO,EAAE,eAAe,EAAE,MAAM,gBAAgB,CAAA;AAEhD,OAAO,EAAE,eAAe,EAAE,KAAK,YAAY,EAAE,CAAA;AAE7C,MAAM,WAAW,cAAc;IAC7B,GAAG,EAAE,GAAG,CAAA;IACR,KAAK,EAAE,IAAI,EAAE,CAAA;IACb,MAAM,EAAE,UAAU,EAAE,CAAA;CACrB;AAED,MAAM,WAAW,qBAAqB;IACpC,YAAY,EAAE,YAAY,CAAA;IAC1B,MAAM,EAAE,MAAM,CAAA;IACd,aAAa,CAAC,EAAE,MAAM,CAAA;IACtB,cAAc,CAAC,EAAE,MAAM,CAAA;CACxB;AAED,wCAAwC;AACxC,MAAM,WAAW,WAAW;IAC1B,EAAE,EAAE,MAAM,CAAA;IACV,KAAK,EAAE,CAAC,GAAG,EAAE,cAAc,KAAK;QAC9B,YAAY,EAAE,YAAY,CAAA;QAC1B,MAAM,EAAE,MAAM,CAAA;QACd,aAAa,CAAC,EAAE,MAAM,CAAA;QACtB,cAAc,CAAC,EAAE,MAAM,CAAA;KACxB,GAAG,IAAI,CAAA;CACT;AAED,eAAO,MAAM,aAAa,EAAE,WAAW,EAmXtC,CAAA;AAqCD,qEAAqE;AACrE,wBAAgB,eAAe,CAC7B,GAAG,EAAE,cAAc,EACnB,KAAK,GAAE,WAAW,EAAkB,GACnC,qBAAqB,CASvB"}
|
|
@@ -1,15 +1,14 @@
|
|
|
1
|
-
import {
|
|
2
|
-
import {
|
|
3
|
-
|
|
4
|
-
type
|
|
5
|
-
type
|
|
6
|
-
type
|
|
7
|
-
|
|
8
|
-
interface FeedbackTask {
|
|
1
|
+
import type { ControlEvalResult, ControlRunResult, ControlStep } from './control-runtime';
|
|
2
|
+
import type { DatasetScenario, DatasetSplit } from './dataset';
|
|
3
|
+
export type FeedbackArtifactType = 'text' | 'code' | 'plan' | 'research' | 'action' | 'ui' | 'decision' | 'data' | 'other';
|
|
4
|
+
export type FeedbackLabelSource = 'user' | 'judge' | 'environment' | 'metric' | 'policy' | 'system';
|
|
5
|
+
export type FeedbackLabelKind = 'approve' | 'reject' | 'select' | 'edit' | 'rank' | 'rate' | 'comment' | 'metric_outcome' | 'policy_block' | 'revision_request';
|
|
6
|
+
export type FeedbackSeverity = 'info' | 'warning' | 'error' | 'critical';
|
|
7
|
+
export interface FeedbackTask {
|
|
9
8
|
intent: string;
|
|
10
9
|
context?: unknown;
|
|
11
10
|
}
|
|
12
|
-
interface ProposedSideEffect {
|
|
11
|
+
export interface ProposedSideEffect {
|
|
13
12
|
type: string;
|
|
14
13
|
risk?: 'low' | 'medium' | 'high';
|
|
15
14
|
costUsd?: number;
|
|
@@ -17,7 +16,7 @@ interface ProposedSideEffect {
|
|
|
17
16
|
requiresApproval?: boolean;
|
|
18
17
|
metadata?: Record<string, unknown>;
|
|
19
18
|
}
|
|
20
|
-
interface FeedbackLabel {
|
|
19
|
+
export interface FeedbackLabel {
|
|
21
20
|
id?: string;
|
|
22
21
|
source: FeedbackLabelSource;
|
|
23
22
|
kind: FeedbackLabelKind;
|
|
@@ -27,7 +26,7 @@ interface FeedbackLabel {
|
|
|
27
26
|
createdAt: string;
|
|
28
27
|
metadata?: Record<string, unknown>;
|
|
29
28
|
}
|
|
30
|
-
interface FeedbackAttempt {
|
|
29
|
+
export interface FeedbackAttempt {
|
|
31
30
|
id: string;
|
|
32
31
|
stepIndex: number;
|
|
33
32
|
artifactType: FeedbackArtifactType;
|
|
@@ -39,7 +38,7 @@ interface FeedbackAttempt {
|
|
|
39
38
|
createdAt: string;
|
|
40
39
|
metadata?: Record<string, unknown>;
|
|
41
40
|
}
|
|
42
|
-
interface FeedbackOutcome {
|
|
41
|
+
export interface FeedbackOutcome {
|
|
43
42
|
success?: boolean;
|
|
44
43
|
score?: number;
|
|
45
44
|
metrics?: Record<string, number>;
|
|
@@ -48,7 +47,7 @@ interface FeedbackOutcome {
|
|
|
48
47
|
observedAt?: string;
|
|
49
48
|
metadata?: Record<string, unknown>;
|
|
50
49
|
}
|
|
51
|
-
interface FeedbackTrajectory {
|
|
50
|
+
export interface FeedbackTrajectory {
|
|
52
51
|
id: string;
|
|
53
52
|
projectId?: string;
|
|
54
53
|
scenarioId?: string;
|
|
@@ -62,26 +61,26 @@ interface FeedbackTrajectory {
|
|
|
62
61
|
updatedAt?: string;
|
|
63
62
|
metadata?: Record<string, unknown>;
|
|
64
63
|
}
|
|
65
|
-
interface FeedbackTrajectoryStore {
|
|
64
|
+
export interface FeedbackTrajectoryStore {
|
|
66
65
|
save(trajectory: FeedbackTrajectory): Promise<void>;
|
|
67
66
|
get(id: string): Promise<FeedbackTrajectory | null>;
|
|
68
67
|
list(filter?: FeedbackTrajectoryFilter): Promise<FeedbackTrajectory[]>;
|
|
69
68
|
appendAttempt(id: string, attempt: FeedbackAttempt): Promise<FeedbackTrajectory>;
|
|
70
69
|
appendLabel(id: string, label: FeedbackLabel, attemptId?: string): Promise<FeedbackTrajectory>;
|
|
71
70
|
}
|
|
72
|
-
interface FeedbackTrajectoryFilter {
|
|
71
|
+
export interface FeedbackTrajectoryFilter {
|
|
73
72
|
projectId?: string;
|
|
74
73
|
scenarioId?: string;
|
|
75
74
|
split?: DatasetSplit;
|
|
76
75
|
tag?: [string, string];
|
|
77
76
|
}
|
|
78
|
-
interface FeedbackSplitPolicy {
|
|
77
|
+
export interface FeedbackSplitPolicy {
|
|
79
78
|
trainPct?: number;
|
|
80
79
|
devPct?: number;
|
|
81
80
|
testPct?: number;
|
|
82
81
|
holdoutPct?: number;
|
|
83
82
|
}
|
|
84
|
-
interface PreferenceMemoryEntry {
|
|
83
|
+
export interface PreferenceMemoryEntry {
|
|
85
84
|
instruction: string;
|
|
86
85
|
rationale: string;
|
|
87
86
|
weight: number;
|
|
@@ -89,14 +88,14 @@ interface PreferenceMemoryEntry {
|
|
|
89
88
|
sourceLabelId?: string;
|
|
90
89
|
category?: string;
|
|
91
90
|
}
|
|
92
|
-
interface FeedbackOptimizerRow {
|
|
91
|
+
export interface FeedbackOptimizerRow {
|
|
93
92
|
scenarioId: string;
|
|
94
93
|
trajectoryId: string;
|
|
95
94
|
labelKinds: FeedbackLabelKind[];
|
|
96
95
|
score?: number;
|
|
97
96
|
metadata?: Record<string, unknown>;
|
|
98
97
|
}
|
|
99
|
-
interface FeedbackReplayResult {
|
|
98
|
+
export interface FeedbackReplayResult {
|
|
100
99
|
trajectoryId: string;
|
|
101
100
|
pass: boolean;
|
|
102
101
|
score?: number;
|
|
@@ -104,10 +103,10 @@ interface FeedbackReplayResult {
|
|
|
104
103
|
outcome?: FeedbackOutcome;
|
|
105
104
|
metadata?: Record<string, unknown>;
|
|
106
105
|
}
|
|
107
|
-
interface FeedbackReplayAdapter {
|
|
106
|
+
export interface FeedbackReplayAdapter {
|
|
108
107
|
replay(trajectory: FeedbackTrajectory): Promise<Omit<FeedbackReplayResult, 'trajectoryId'>> | Omit<FeedbackReplayResult, 'trajectoryId'>;
|
|
109
108
|
}
|
|
110
|
-
declare class InMemoryFeedbackTrajectoryStore implements FeedbackTrajectoryStore {
|
|
109
|
+
export declare class InMemoryFeedbackTrajectoryStore implements FeedbackTrajectoryStore {
|
|
111
110
|
private readonly trajectories;
|
|
112
111
|
save(trajectory: FeedbackTrajectory): Promise<void>;
|
|
113
112
|
get(id: string): Promise<FeedbackTrajectory | null>;
|
|
@@ -115,7 +114,7 @@ declare class InMemoryFeedbackTrajectoryStore implements FeedbackTrajectoryStore
|
|
|
115
114
|
appendAttempt(id: string, attempt: FeedbackAttempt): Promise<FeedbackTrajectory>;
|
|
116
115
|
appendLabel(id: string, label: FeedbackLabel, attemptId?: string): Promise<FeedbackTrajectory>;
|
|
117
116
|
}
|
|
118
|
-
declare class FileSystemFeedbackTrajectoryStore implements FeedbackTrajectoryStore {
|
|
117
|
+
export declare class FileSystemFeedbackTrajectoryStore implements FeedbackTrajectoryStore {
|
|
119
118
|
private readonly dir;
|
|
120
119
|
private readonly memory;
|
|
121
120
|
private loaded;
|
|
@@ -130,7 +129,7 @@ declare class FileSystemFeedbackTrajectoryStore implements FeedbackTrajectorySto
|
|
|
130
129
|
private append;
|
|
131
130
|
private load;
|
|
132
131
|
}
|
|
133
|
-
declare function createFeedbackTrajectory(input: {
|
|
132
|
+
export declare function createFeedbackTrajectory(input: {
|
|
134
133
|
id?: string;
|
|
135
134
|
projectId?: string;
|
|
136
135
|
scenarioId?: string;
|
|
@@ -143,21 +142,21 @@ declare function createFeedbackTrajectory(input: {
|
|
|
143
142
|
createdAt?: string;
|
|
144
143
|
metadata?: Record<string, unknown>;
|
|
145
144
|
}): FeedbackTrajectory;
|
|
146
|
-
declare function assignFeedbackSplit(trajectory: Pick<FeedbackTrajectory, 'id' | 'projectId' | 'scenarioId' | 'task'>, policy?: FeedbackSplitPolicy): DatasetSplit;
|
|
147
|
-
declare function withAssignedFeedbackSplit(trajectory: FeedbackTrajectory, policy?: FeedbackSplitPolicy): FeedbackTrajectory;
|
|
148
|
-
declare function feedbackTrajectoryToDatasetScenario(trajectory: FeedbackTrajectory): DatasetScenario;
|
|
149
|
-
declare function feedbackTrajectoriesToDatasetScenarios(trajectories: FeedbackTrajectory[]): DatasetScenario[];
|
|
150
|
-
declare function feedbackTrajectoryToOptimizerRow(trajectory: FeedbackTrajectory): FeedbackOptimizerRow;
|
|
151
|
-
declare function feedbackTrajectoriesToOptimizerRows(trajectories: FeedbackTrajectory[]): FeedbackOptimizerRow[];
|
|
152
|
-
declare function replayFeedbackTrajectory(trajectory: FeedbackTrajectory, adapter: FeedbackReplayAdapter): Promise<FeedbackReplayResult>;
|
|
153
|
-
declare function replayFeedbackTrajectories(trajectories: FeedbackTrajectory[], adapter: FeedbackReplayAdapter): Promise<FeedbackReplayResult[]>;
|
|
154
|
-
declare function summarizePreferenceMemory(trajectories: FeedbackTrajectory[], options?: {
|
|
145
|
+
export declare function assignFeedbackSplit(trajectory: Pick<FeedbackTrajectory, 'id' | 'projectId' | 'scenarioId' | 'task'>, policy?: FeedbackSplitPolicy): DatasetSplit;
|
|
146
|
+
export declare function withAssignedFeedbackSplit(trajectory: FeedbackTrajectory, policy?: FeedbackSplitPolicy): FeedbackTrajectory;
|
|
147
|
+
export declare function feedbackTrajectoryToDatasetScenario(trajectory: FeedbackTrajectory): DatasetScenario;
|
|
148
|
+
export declare function feedbackTrajectoriesToDatasetScenarios(trajectories: FeedbackTrajectory[]): DatasetScenario[];
|
|
149
|
+
export declare function feedbackTrajectoryToOptimizerRow(trajectory: FeedbackTrajectory): FeedbackOptimizerRow;
|
|
150
|
+
export declare function feedbackTrajectoriesToOptimizerRows(trajectories: FeedbackTrajectory[]): FeedbackOptimizerRow[];
|
|
151
|
+
export declare function replayFeedbackTrajectory(trajectory: FeedbackTrajectory, adapter: FeedbackReplayAdapter): Promise<FeedbackReplayResult>;
|
|
152
|
+
export declare function replayFeedbackTrajectories(trajectories: FeedbackTrajectory[], adapter: FeedbackReplayAdapter): Promise<FeedbackReplayResult[]>;
|
|
153
|
+
export declare function summarizePreferenceMemory(trajectories: FeedbackTrajectory[], options?: {
|
|
155
154
|
maxEntries?: number;
|
|
156
155
|
}): PreferenceMemoryEntry[];
|
|
157
|
-
declare function renderPreferenceMemoryMarkdown(entries: PreferenceMemoryEntry[]): string;
|
|
158
|
-
declare function serializeFeedbackTrajectoriesJsonl(trajectories: FeedbackTrajectory[]): string;
|
|
159
|
-
declare function parseFeedbackTrajectoriesJsonl(jsonl: string): FeedbackTrajectory[];
|
|
160
|
-
declare function controlRunToFeedbackTrajectory<TState, TAction, TActionResult>(run: ControlRunResult<TState, TAction, TActionResult>, options?: {
|
|
156
|
+
export declare function renderPreferenceMemoryMarkdown(entries: PreferenceMemoryEntry[]): string;
|
|
157
|
+
export declare function serializeFeedbackTrajectoriesJsonl(trajectories: FeedbackTrajectory[]): string;
|
|
158
|
+
export declare function parseFeedbackTrajectoriesJsonl(jsonl: string): FeedbackTrajectory[];
|
|
159
|
+
export declare function controlRunToFeedbackTrajectory<TState, TAction, TActionResult>(run: ControlRunResult<TState, TAction, TActionResult>, options?: {
|
|
161
160
|
projectId?: string;
|
|
162
161
|
scenarioId?: string;
|
|
163
162
|
artifactType?: FeedbackArtifactType;
|
|
@@ -165,5 +164,4 @@ declare function controlRunToFeedbackTrajectory<TState, TAction, TActionResult>(
|
|
|
165
164
|
proposedActionFromStep?: (step: ControlStep<TState, TAction, TActionResult>) => ProposedSideEffect | undefined;
|
|
166
165
|
createdAt?: string;
|
|
167
166
|
}): FeedbackTrajectory;
|
|
168
|
-
|
|
169
|
-
export { replayFeedbackTrajectory as A, serializeFeedbackTrajectoriesJsonl as B, summarizePreferenceMemory as C, withAssignedFeedbackSplit as D, type FeedbackTrajectoryStore as F, InMemoryFeedbackTrajectoryStore as I, type PreferenceMemoryEntry as P, type FeedbackTrajectory as a, type FeedbackLabel as b, type FeedbackArtifactType as c, type FeedbackAttempt as d, type FeedbackLabelKind as e, type FeedbackLabelSource as f, type FeedbackOptimizerRow as g, type FeedbackOutcome as h, type FeedbackReplayAdapter as i, type FeedbackReplayResult as j, type FeedbackSeverity as k, type FeedbackSplitPolicy as l, type FeedbackTask as m, type FeedbackTrajectoryFilter as n, FileSystemFeedbackTrajectoryStore as o, type ProposedSideEffect as p, assignFeedbackSplit as q, controlRunToFeedbackTrajectory as r, createFeedbackTrajectory as s, feedbackTrajectoriesToDatasetScenarios as t, feedbackTrajectoriesToOptimizerRows as u, feedbackTrajectoryToDatasetScenario as v, feedbackTrajectoryToOptimizerRow as w, parseFeedbackTrajectoriesJsonl as x, renderPreferenceMemoryMarkdown as y, replayFeedbackTrajectories as z };
|
|
167
|
+
//# sourceMappingURL=feedback-trajectory.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"feedback-trajectory.d.ts","sourceRoot":"","sources":["../src/feedback-trajectory.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,iBAAiB,EAAE,gBAAgB,EAAE,WAAW,EAAE,MAAM,mBAAmB,CAAA;AACzF,OAAO,KAAK,EAAE,eAAe,EAAE,YAAY,EAAE,MAAM,WAAW,CAAA;AAE9D,MAAM,MAAM,oBAAoB,GAC5B,MAAM,GACN,MAAM,GACN,MAAM,GACN,UAAU,GACV,QAAQ,GACR,IAAI,GACJ,UAAU,GACV,MAAM,GACN,OAAO,CAAA;AAEX,MAAM,MAAM,mBAAmB,GAAG,MAAM,GAAG,OAAO,GAAG,aAAa,GAAG,QAAQ,GAAG,QAAQ,GAAG,QAAQ,CAAA;AAEnG,MAAM,MAAM,iBAAiB,GACzB,SAAS,GACT,QAAQ,GACR,QAAQ,GACR,MAAM,GACN,MAAM,GACN,MAAM,GACN,SAAS,GACT,gBAAgB,GAChB,cAAc,GACd,kBAAkB,CAAA;AAEtB,MAAM,MAAM,gBAAgB,GAAG,MAAM,GAAG,SAAS,GAAG,OAAO,GAAG,UAAU,CAAA;AAExE,MAAM,WAAW,YAAY;IAC3B,MAAM,EAAE,MAAM,CAAA;IACd,OAAO,CAAC,EAAE,OAAO,CAAA;CAClB;AAED,MAAM,WAAW,kBAAkB;IACjC,IAAI,EAAE,MAAM,CAAA;IACZ,IAAI,CAAC,EAAE,KAAK,GAAG,QAAQ,GAAG,MAAM,CAAA;IAChC,OAAO,CAAC,EAAE,MAAM,CAAA;IAChB,kBAAkB,CAAC,EAAE,OAAO,CAAA;IAC5B,gBAAgB,CAAC,EAAE,OAAO,CAAA;IAC1B,QAAQ,CAAC,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,CAAA;CACnC;AAED,MAAM,WAAW,aAAa;IAC5B,EAAE,CAAC,EAAE,MAAM,CAAA;IACX,MAAM,EAAE,mBAAmB,CAAA;IAC3B,IAAI,EAAE,iBAAiB,CAAA;IACvB,KAAK,EAAE,OAAO,CAAA;IACd,MAAM,CAAC,EAAE,MAAM,CAAA;IACf,QAAQ,CAAC,EAAE,gBAAgB,CAAA;IAC3B,SAAS,EAAE,MAAM,CAAA;IACjB,QAAQ,CAAC,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,CAAA;CACnC;AAED,MAAM,WAAW,eAAe;IAC9B,EAAE,EAAE,MAAM,CAAA;IACV,SAAS,EAAE,MAAM,CAAA;IACjB,YAAY,EAAE,oBAAoB,CAAA;IAClC,QAAQ,EAAE,OAAO,CAAA;IACjB,OAAO,CAAC,EAAE,OAAO,EAAE,CAAA;IACnB,cAAc,CAAC,EAAE,kBAAkB,CAAA;IACnC,KAAK,CAAC,EAAE,iBAAiB,EAAE,CAAA;IAC3B,QAAQ,CAAC,EAAE,aAAa,EAAE,CAAA;IAC1B,SAAS,EAAE,MAAM,CAAA;IACjB,QAAQ,CAAC,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,CAAA;CACnC;AAED,MAAM,WAAW,eAAe;IAC9B,OAAO,CAAC,EAAE,OAAO,CAAA;IACjB,KAAK,CAAC,EAAE,MAAM,CAAA;IACd,OAAO,CAAC,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,CAAA;IAChC,OAAO,CAAC,EAAE,MAAM,CAAA;IAChB,MAAM,CAAC,EAAE,MAAM,CAAA;IACf,UAAU,CAAC,EAAE,MAAM,CAAA;IACnB,QAAQ,CAAC,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,CAAA;CACnC;AAED,MAAM,WAAW,kBAAkB;IACjC,EAAE,EAAE,MAAM,CAAA;IACV,SAAS,CAAC,EAAE,MAAM,CAAA;IAClB,UAAU,CAAC,EAAE,MAAM,CAAA;IACnB,IAAI,EAAE,YAAY,CAAA;IAClB,QAAQ,EAAE,eAAe,EAAE,CAAA;IAC3B,MAAM,EAAE,aAAa,EAAE,CAAA;IACvB,OAAO,CAAC,EAAE,eAAe,CAAA;IACzB,KAAK,CAAC,EAAE,YAAY,CAAA;IACpB,IAAI,CAAC,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,CAAA;IAC7B,SAAS,EAAE,MAAM,CAAA;IACjB,SAAS,CAAC,EAAE,MAAM,CAAA;IAClB,QAAQ,CAAC,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,CAAA;CACnC;AAED,MAAM,WAAW,uBAAuB;IACtC,IAAI,CAAC,UAAU,EAAE,kBAAkB,GAAG,OAAO,CAAC,IAAI,CAAC,CAAA;IACnD,GAAG,CAAC,EAAE,EAAE,MAAM,GAAG,OAAO,CAAC,kBAAkB,GAAG,IAAI,CAAC,CAAA;IACnD,IAAI,CAAC,MAAM,CAAC,EAAE,wBAAwB,GAAG,OAAO,CAAC,kBAAkB,EAAE,CAAC,CAAA;IACtE,aAAa,CAAC,EAAE,EAAE,MAAM,EAAE,OAAO,EAAE,eAAe,GAAG,OAAO,CAAC,kBAAkB,CAAC,CAAA;IAChF,WAAW,CAAC,EAAE,EAAE,MAAM,EAAE,KAAK,EAAE,aAAa,EAAE,SAAS,CAAC,EAAE,MAAM,GAAG,OAAO,CAAC,kBAAkB,CAAC,CAAA;CAC/F;AAED,MAAM,WAAW,wBAAwB;IACvC,SAAS,CAAC,EAAE,MAAM,CAAA;IAClB,UAAU,CAAC,EAAE,MAAM,CAAA;IACnB,KAAK,CAAC,EAAE,YAAY,CAAA;IACpB,GAAG,CAAC,EAAE,CAAC,MAAM,EAAE,MAAM,CAAC,CAAA;CACvB;AAED,MAAM,WAAW,mBAAmB;IAClC,QAAQ,CAAC,EAAE,MAAM,CAAA;IACjB,MAAM,CAAC,EAAE,MAAM,CAAA;IACf,OAAO,CAAC,EAAE,MAAM,CAAA;IAChB,UAAU,CAAC,EAAE,MAAM,CAAA;CACpB;AAED,MAAM,WAAW,qBAAqB;IACpC,WAAW,EAAE,MAAM,CAAA;IACnB,SAAS,EAAE,MAAM,CAAA;IACjB,MAAM,EAAE,MAAM,CAAA;IACd,kBAAkB,EAAE,MAAM,CAAA;IAC1B,aAAa,CAAC,EAAE,MAAM,CAAA;IACtB,QAAQ,CAAC,EAAE,MAAM,CAAA;CAClB;AAED,MAAM,WAAW,oBAAoB;IACnC,UAAU,EAAE,MAAM,CAAA;IAClB,YAAY,EAAE,MAAM,CAAA;IACpB,UAAU,EAAE,iBAAiB,EAAE,CAAA;IAC/B,KAAK,CAAC,EAAE,MAAM,CAAA;IACd,QAAQ,CAAC,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,CAAA;CACnC;AAED,MAAM,WAAW,oBAAoB;IACnC,YAAY,EAAE,MAAM,CAAA;IACpB,IAAI,EAAE,OAAO,CAAA;IACb,KAAK,CAAC,EAAE,MAAM,CAAA;IACd,MAAM,EAAE,aAAa,EAAE,CAAA;IACvB,OAAO,CAAC,EAAE,eAAe,CAAA;IACzB,QAAQ,CAAC,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,CAAA;CACnC;AAED,MAAM,WAAW,qBAAqB;IACpC,MAAM,CACJ,UAAU,EAAE,kBAAkB,GAE5B,OAAO,CAAC,IAAI,CAAC,oBAAoB,EAAE,cAAc,CAAC,CAAC,GACnD,IAAI,CAAC,oBAAoB,EAAE,cAAc,CAAC,CAAA;CAC/C;AASD,qBAAa,+BAAgC,YAAW,uBAAuB;IAC7E,OAAO,CAAC,QAAQ,CAAC,YAAY,CAAwC;IAE/D,IAAI,CAAC,UAAU,EAAE,kBAAkB,GAAG,OAAO,CAAC,IAAI,CAAC;IAInD,GAAG,CAAC,EAAE,EAAE,MAAM,GAAG,OAAO,CAAC,kBAAkB,GAAG,IAAI,CAAC;IAKnD,IAAI,CAAC,MAAM,GAAE,wBAA6B,GAAG,OAAO,CAAC,kBAAkB,EAAE,CAAC;IAM1E,aAAa,CAAC,EAAE,EAAE,MAAM,EAAE,OAAO,EAAE,eAAe,GAAG,OAAO,CAAC,kBAAkB,CAAC;IAahF,WAAW,CACf,EAAE,EAAE,MAAM,EACV,KAAK,EAAE,aAAa,EACpB,SAAS,CAAC,EAAE,MAAM,GACjB,OAAO,CAAC,kBAAkB,CAAC;CAoB/B;AAED,qBAAa,iCAAkC,YAAW,uBAAuB;IAC/E,OAAO,CAAC,QAAQ,CAAC,GAAG,CAAQ;IAC5B,OAAO,CAAC,QAAQ,CAAC,MAAM,CAAwC;IAC/D,OAAO,CAAC,MAAM,CAAQ;gBAEV,OAAO,EAAE;QAAE,GAAG,EAAE,MAAM,CAAA;KAAE;IAI9B,IAAI,CAAC,UAAU,EAAE,kBAAkB,GAAG,OAAO,CAAC,IAAI,CAAC;IAMnD,GAAG,CAAC,EAAE,EAAE,MAAM,GAAG,OAAO,CAAC,kBAAkB,GAAG,IAAI,CAAC;IAKnD,IAAI,CAAC,MAAM,GAAE,wBAA6B,GAAG,OAAO,CAAC,kBAAkB,EAAE,CAAC;IAK1E,aAAa,CAAC,EAAE,EAAE,MAAM,EAAE,OAAO,EAAE,eAAe,GAAG,OAAO,CAAC,kBAAkB,CAAC;IAOhF,WAAW,CACf,EAAE,EAAE,MAAM,EACV,KAAK,EAAE,aAAa,EACpB,SAAS,CAAC,EAAE,MAAM,GACjB,OAAO,CAAC,kBAAkB,CAAC;YAOhB,MAAM;YAWN,IAAI;CA4BnB;AAED,wBAAgB,wBAAwB,CAAC,KAAK,EAAE;IAC9C,EAAE,CAAC,EAAE,MAAM,CAAA;IACX,SAAS,CAAC,EAAE,MAAM,CAAA;IAClB,UAAU,CAAC,EAAE,MAAM,CAAA;IACnB,IAAI,EAAE,YAAY,CAAA;IAClB,QAAQ,CAAC,EAAE,eAAe,EAAE,CAAA;IAC5B,MAAM,CAAC,EAAE,aAAa,EAAE,CAAA;IACxB,OAAO,CAAC,EAAE,eAAe,CAAA;IACzB,KAAK,CAAC,EAAE,YAAY,CAAA;IACpB,IAAI,CAAC,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,CAAA;IAC7B,SAAS,CAAC,EAAE,MAAM,CAAA;IAClB,QAAQ,CAAC,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,CAAA;CACnC,GAAG,kBAAkB,CAkBrB;AAED,wBAAgB,mBAAmB,CACjC,UAAU,EAAE,IAAI,CAAC,kBAAkB,EAAE,IAAI,GAAG,WAAW,GAAG,YAAY,GAAG,MAAM,CAAC,EAChF,MAAM,GAAE,mBAAwB,GAC/B,YAAY,CAYd;AAED,wBAAgB,yBAAyB,CACvC,UAAU,EAAE,kBAAkB,EAC9B,MAAM,CAAC,EAAE,mBAAmB,GAC3B,kBAAkB,CAKpB;AAED,wBAAgB,mCAAmC,CACjD,UAAU,EAAE,kBAAkB,GAC7B,eAAe,CAYjB;AAED,wBAAgB,sCAAsC,CACpD,YAAY,EAAE,kBAAkB,EAAE,GACjC,eAAe,EAAE,CAEnB;AAED,wBAAgB,gCAAgC,CAC9C,UAAU,EAAE,kBAAkB,GAC7B,oBAAoB,CAgBtB;AAED,wBAAgB,mCAAmC,CACjD,YAAY,EAAE,kBAAkB,EAAE,GACjC,oBAAoB,EAAE,CAExB;AAED,wBAAsB,wBAAwB,CAC5C,UAAU,EAAE,kBAAkB,EAC9B,OAAO,EAAE,qBAAqB,GAC7B,OAAO,CAAC,oBAAoB,CAAC,CAgC/B;AAED,wBAAsB,0BAA0B,CAC9C,YAAY,EAAE,kBAAkB,EAAE,EAClC,OAAO,EAAE,qBAAqB,GAC7B,OAAO,CAAC,oBAAoB,EAAE,CAAC,CAMjC;AAED,wBAAgB,yBAAyB,CACvC,YAAY,EAAE,kBAAkB,EAAE,EAClC,OAAO,GAAE;IAAE,UAAU,CAAC,EAAE,MAAM,CAAA;CAAO,GACpC,qBAAqB,EAAE,CAyBzB;AAED,wBAAgB,8BAA8B,CAAC,OAAO,EAAE,qBAAqB,EAAE,GAAG,MAAM,CASvF;AAED,wBAAgB,kCAAkC,CAAC,YAAY,EAAE,kBAAkB,EAAE,GAAG,MAAM,CAM7F;AAED,wBAAgB,8BAA8B,CAAC,KAAK,EAAE,MAAM,GAAG,kBAAkB,EAAE,CAOlF;AAED,wBAAgB,8BAA8B,CAAC,MAAM,EAAE,OAAO,EAAE,aAAa,EAC3E,GAAG,EAAE,gBAAgB,CAAC,MAAM,EAAE,OAAO,EAAE,aAAa,CAAC,EACrD,OAAO,GAAE;IACP,SAAS,CAAC,EAAE,MAAM,CAAA;IAClB,UAAU,CAAC,EAAE,MAAM,CAAA;IACnB,YAAY,CAAC,EAAE,oBAAoB,CAAA;IACnC,gBAAgB,CAAC,EAAE,CAAC,IAAI,EAAE,WAAW,CAAC,MAAM,EAAE,OAAO,EAAE,aAAa,CAAC,KAAK,OAAO,CAAA;IACjF,sBAAsB,CAAC,EAAE,CACvB,IAAI,EAAE,WAAW,CAAC,MAAM,EAAE,OAAO,EAAE,aAAa,CAAC,KAC9C,kBAAkB,GAAG,SAAS,CAAA;IACnC,SAAS,CAAC,EAAE,MAAM,CAAA;CACd,GACL,kBAAkB,CA6CpB"}
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"feedback-trajectory.test.d.ts","sourceRoot":"","sources":["../src/feedback-trajectory.test.ts"],"names":[],"mappings":""}
|
|
@@ -0,0 +1,90 @@
|
|
|
1
|
+
/**
|
|
2
|
+
* Flow layer — drive a previewed app through a scripted user walk.
|
|
3
|
+
*
|
|
4
|
+
* The MultiLayerVerifier already had a `flow` slot wired in
|
|
5
|
+
* VerticalBench's verification-harness, but the layer module was
|
|
6
|
+
* always-skipped ("flow layer module not yet wired"). This adds the
|
|
7
|
+
* module: a Layer<Env> that takes a {@link FlowSpec} (URL + steps),
|
|
8
|
+
* boots a preview server via the supplied {@link FlowRunner}, executes
|
|
9
|
+
* each step, and returns a LayerResult whose `findings` enumerate
|
|
10
|
+
* which step failed.
|
|
11
|
+
*
|
|
12
|
+
* The runner is injected so this module can swap between:
|
|
13
|
+
* - production: agent-browser CLI (a11y-tree based steps)
|
|
14
|
+
* - test: in-memory mock that returns canned step outcomes
|
|
15
|
+
* - future: Playwright, Puppeteer, custom scrapers
|
|
16
|
+
*
|
|
17
|
+
* Paired with {@link runIntentMatchJudge}: intent-match catches "wrong
|
|
18
|
+
* app entirely"; flow-layer catches "right app but the buttons don't work."
|
|
19
|
+
*/
|
|
20
|
+
import type { Layer, Severity } from './multi-layer-verifier';
|
|
21
|
+
export type FlowAction = 'navigate' | 'click' | 'fill' | 'expect-text' | 'expect-element' | 'expect-url' | 'wait';
|
|
22
|
+
export interface FlowStep {
|
|
23
|
+
/** What this step does. */
|
|
24
|
+
action: FlowAction;
|
|
25
|
+
/** Human-readable description for findings. */
|
|
26
|
+
describe?: string;
|
|
27
|
+
/**
|
|
28
|
+
* For navigate/expect-url: full URL. For click/fill/expect-element:
|
|
29
|
+
* accessible-name selector or CSS selector.
|
|
30
|
+
* For expect-text: substring expected on the page.
|
|
31
|
+
* For wait: ignored (use `value` for ms).
|
|
32
|
+
*/
|
|
33
|
+
target?: string;
|
|
34
|
+
/** For fill: text to enter. For wait: ms. */
|
|
35
|
+
value?: string;
|
|
36
|
+
/** Severity of a failure. Default `major`. */
|
|
37
|
+
severity?: Severity;
|
|
38
|
+
}
|
|
39
|
+
export interface FlowSpec {
|
|
40
|
+
/** Initial URL the runner should open. */
|
|
41
|
+
url: string;
|
|
42
|
+
/** Ordered steps. Stops at the first failure unless `continueOnFail: true`. */
|
|
43
|
+
steps: FlowStep[];
|
|
44
|
+
/** When true, execute every step even after a failure (collect all findings). */
|
|
45
|
+
continueOnFail?: boolean;
|
|
46
|
+
/** Per-step wall cap (ms). Default 15s. */
|
|
47
|
+
stepTimeoutMs?: number;
|
|
48
|
+
}
|
|
49
|
+
export interface FlowRunnerStepResult {
|
|
50
|
+
ok: boolean;
|
|
51
|
+
/** Concrete observation: matched text snippet, captured URL, error message. */
|
|
52
|
+
evidence?: string;
|
|
53
|
+
/** Wall-clock duration of the step. */
|
|
54
|
+
durationMs?: number;
|
|
55
|
+
}
|
|
56
|
+
export interface FlowRunner {
|
|
57
|
+
/** Open the target URL. Returns when the page is interactable. */
|
|
58
|
+
open(url: string): Promise<FlowRunnerStepResult>;
|
|
59
|
+
/** Execute one step. The runner owns interpretation of `target`. */
|
|
60
|
+
step(step: FlowStep): Promise<FlowRunnerStepResult>;
|
|
61
|
+
/** Tear down browser, free resources. Always called once per layer.run. */
|
|
62
|
+
close(): Promise<void>;
|
|
63
|
+
}
|
|
64
|
+
export interface FlowLayerEnv {
|
|
65
|
+
/** Optional override per-call. Defaults supplied by the layer factory. */
|
|
66
|
+
flowSpec?: FlowSpec;
|
|
67
|
+
}
|
|
68
|
+
export interface FlowLayerFactoryInput {
|
|
69
|
+
/** Static spec (used when env doesn't supply one). */
|
|
70
|
+
flowSpec?: FlowSpec;
|
|
71
|
+
/** Build the runner per call (lets the layer create + tear down per leaf). */
|
|
72
|
+
runner: () => FlowRunner | Promise<FlowRunner>;
|
|
73
|
+
/** Layer name. Default `flow`. */
|
|
74
|
+
name?: string;
|
|
75
|
+
/** Layer dependencies — default `['serve']` so a non-booting preview skips us. */
|
|
76
|
+
dependsOn?: string[];
|
|
77
|
+
/** Layer weight for blendedScore (0..1+). Default 1. */
|
|
78
|
+
weight?: number;
|
|
79
|
+
/** Cap for the entire flow run (ms). Default 60s. */
|
|
80
|
+
capMs?: number;
|
|
81
|
+
}
|
|
82
|
+
/**
|
|
83
|
+
* Build a flow layer that scripts a user walk via the supplied runner.
|
|
84
|
+
*
|
|
85
|
+
* Score: 1.0 when every step passed; otherwise 1 - (failedSteps / totalSteps).
|
|
86
|
+
* Status: `pass` iff every step passed; `fail` if any step failed; `error`
|
|
87
|
+
* on runner setup error; `skipped` when no flowSpec is available.
|
|
88
|
+
*/
|
|
89
|
+
export declare function flowLayer<Env extends FlowLayerEnv = FlowLayerEnv>(input: FlowLayerFactoryInput): Layer<Env>;
|
|
90
|
+
//# sourceMappingURL=flow-layer.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"flow-layer.d.ts","sourceRoot":"","sources":["../src/flow-layer.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;GAkBG;AAEH,OAAO,KAAK,EAAE,KAAK,EAAe,QAAQ,EAAE,MAAM,wBAAwB,CAAA;AAI1E,MAAM,MAAM,UAAU,GAClB,UAAU,GACV,OAAO,GACP,MAAM,GACN,aAAa,GACb,gBAAgB,GAChB,YAAY,GACZ,MAAM,CAAA;AAEV,MAAM,WAAW,QAAQ;IACvB,2BAA2B;IAC3B,MAAM,EAAE,UAAU,CAAA;IAClB,+CAA+C;IAC/C,QAAQ,CAAC,EAAE,MAAM,CAAA;IACjB;;;;;OAKG;IACH,MAAM,CAAC,EAAE,MAAM,CAAA;IACf,6CAA6C;IAC7C,KAAK,CAAC,EAAE,MAAM,CAAA;IACd,8CAA8C;IAC9C,QAAQ,CAAC,EAAE,QAAQ,CAAA;CACpB;AAED,MAAM,WAAW,QAAQ;IACvB,0CAA0C;IAC1C,GAAG,EAAE,MAAM,CAAA;IACX,+EAA+E;IAC/E,KAAK,EAAE,QAAQ,EAAE,CAAA;IACjB,iFAAiF;IACjF,cAAc,CAAC,EAAE,OAAO,CAAA;IACxB,2CAA2C;IAC3C,aAAa,CAAC,EAAE,MAAM,CAAA;CACvB;AAED,MAAM,WAAW,oBAAoB;IACnC,EAAE,EAAE,OAAO,CAAA;IACX,+EAA+E;IAC/E,QAAQ,CAAC,EAAE,MAAM,CAAA;IACjB,uCAAuC;IACvC,UAAU,CAAC,EAAE,MAAM,CAAA;CACpB;AAED,MAAM,WAAW,UAAU;IACzB,kEAAkE;IAClE,IAAI,CAAC,GAAG,EAAE,MAAM,GAAG,OAAO,CAAC,oBAAoB,CAAC,CAAA;IAChD,oEAAoE;IACpE,IAAI,CAAC,IAAI,EAAE,QAAQ,GAAG,OAAO,CAAC,oBAAoB,CAAC,CAAA;IACnD,2EAA2E;IAC3E,KAAK,IAAI,OAAO,CAAC,IAAI,CAAC,CAAA;CACvB;AAED,MAAM,WAAW,YAAY;IAC3B,0EAA0E;IAC1E,QAAQ,CAAC,EAAE,QAAQ,CAAA;CACpB;AAED,MAAM,WAAW,qBAAqB;IACpC,sDAAsD;IACtD,QAAQ,CAAC,EAAE,QAAQ,CAAA;IACnB,8EAA8E;IAC9E,MAAM,EAAE,MAAM,UAAU,GAAG,OAAO,CAAC,UAAU,CAAC,CAAA;IAC9C,kCAAkC;IAClC,IAAI,CAAC,EAAE,MAAM,CAAA;IACb,kFAAkF;IAClF,SAAS,CAAC,EAAE,MAAM,EAAE,CAAA;IACpB,wDAAwD;IACxD,MAAM,CAAC,EAAE,MAAM,CAAA;IACf,qDAAqD;IACrD,KAAK,CAAC,EAAE,MAAM,CAAA;CACf;AAID;;;;;;GAMG;AACH,wBAAgB,SAAS,CAAC,GAAG,SAAS,YAAY,GAAG,YAAY,EAC/D,KAAK,EAAE,qBAAqB,GAC3B,KAAK,CAAC,GAAG,CAAC,CAqHZ"}
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"flow-layer.test.d.ts","sourceRoot":"","sources":["../src/flow-layer.test.ts"],"names":[],"mappings":""}
|
|
@@ -0,0 +1,46 @@
|
|
|
1
|
+
/**
|
|
2
|
+
* The capsule — the artifact every exploration produces.
|
|
3
|
+
*
|
|
4
|
+
* `buildCapsule` assembles coverage + verified findings + the QD archive into a
|
|
5
|
+
* pure `CapsuleData` (no clock, no I/O — deterministic and snapshot-testable).
|
|
6
|
+
* `renderCapsuleHtml` turns it into a standalone page: the input-cell heat-map
|
|
7
|
+
* (planned vs covered), per-dimension weakness chips, and the minimized finding
|
|
8
|
+
* exemplars. One artifact — the hardening map and the shareable proof object.
|
|
9
|
+
*/
|
|
10
|
+
import type { EvalRecord } from './cube';
|
|
11
|
+
import type { ArchiveEntry, CapsuleData, Cell, Finding } from './types';
|
|
12
|
+
export interface BuildCapsuleInput<S> {
|
|
13
|
+
target: string;
|
|
14
|
+
objective: string;
|
|
15
|
+
cells: Cell[];
|
|
16
|
+
log: EvalRecord[];
|
|
17
|
+
/** The objective's notable threshold — drives findingRate. */
|
|
18
|
+
threshold: number;
|
|
19
|
+
archive: ArchiveEntry<S>[];
|
|
20
|
+
findings: Finding<S>[];
|
|
21
|
+
candidateFindings: number;
|
|
22
|
+
runsUsed: number;
|
|
23
|
+
/** Known-dollar / unknown-run split — present only when cost tracking was
|
|
24
|
+
* wired; the capsule never fabricates a $0 total. */
|
|
25
|
+
cost?: {
|
|
26
|
+
costUsd: number;
|
|
27
|
+
costUnknownRuns: number;
|
|
28
|
+
};
|
|
29
|
+
/** Evaluations that threw — infra outcomes, never folded into robustness. */
|
|
30
|
+
evalErrors: number;
|
|
31
|
+
/** Set when the consecutive-error circuit breaker stopped the run early. */
|
|
32
|
+
stoppedEarly?: {
|
|
33
|
+
reason: 'eval-errors';
|
|
34
|
+
detail: string;
|
|
35
|
+
};
|
|
36
|
+
}
|
|
37
|
+
export declare function buildCapsule<S>(input: BuildCapsuleInput<S>): CapsuleData<S>;
|
|
38
|
+
export interface RenderCapsuleOptions {
|
|
39
|
+
/** Max finding exemplars to show. Default 8. */
|
|
40
|
+
maxFindings?: number;
|
|
41
|
+
/** ISO timestamp to stamp into the page (keeps the pure capsule clock-free). */
|
|
42
|
+
generatedAt?: string;
|
|
43
|
+
}
|
|
44
|
+
/** Render a self-contained HTML capsule — heat-map + per-dimension chips + verified findings. */
|
|
45
|
+
export declare function renderCapsuleHtml<S>(capsule: CapsuleData<S>, opts?: RenderCapsuleOptions): string;
|
|
46
|
+
//# sourceMappingURL=capsule.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"capsule.d.ts","sourceRoot":"","sources":["../../src/fuzz/capsule.ts"],"names":[],"mappings":"AAAA;;;;;;;;GAQG;AAEH,OAAO,KAAK,EAAE,UAAU,EAAE,MAAM,QAAQ,CAAA;AAExC,OAAO,KAAK,EAAE,YAAY,EAAE,WAAW,EAAE,IAAI,EAA8B,OAAO,EAAE,MAAM,SAAS,CAAA;AAEnG,MAAM,WAAW,iBAAiB,CAAC,CAAC;IAClC,MAAM,EAAE,MAAM,CAAA;IACd,SAAS,EAAE,MAAM,CAAA;IACjB,KAAK,EAAE,IAAI,EAAE,CAAA;IACb,GAAG,EAAE,UAAU,EAAE,CAAA;IACjB,8DAA8D;IAC9D,SAAS,EAAE,MAAM,CAAA;IACjB,OAAO,EAAE,YAAY,CAAC,CAAC,CAAC,EAAE,CAAA;IAC1B,QAAQ,EAAE,OAAO,CAAC,CAAC,CAAC,EAAE,CAAA;IACtB,iBAAiB,EAAE,MAAM,CAAA;IACzB,QAAQ,EAAE,MAAM,CAAA;IAChB;0DACsD;IACtD,IAAI,CAAC,EAAE;QAAE,OAAO,EAAE,MAAM,CAAC;QAAC,eAAe,EAAE,MAAM,CAAA;KAAE,CAAA;IACnD,6EAA6E;IAC7E,UAAU,EAAE,MAAM,CAAA;IAClB,4EAA4E;IAC5E,YAAY,CAAC,EAAE;QAAE,MAAM,EAAE,aAAa,CAAC;QAAC,MAAM,EAAE,MAAM,CAAA;KAAE,CAAA;CACzD;AAED,wBAAgB,YAAY,CAAC,CAAC,EAAE,KAAK,EAAE,iBAAiB,CAAC,CAAC,CAAC,GAAG,WAAW,CAAC,CAAC,CAAC,CAiC3E;AAqGD,MAAM,WAAW,oBAAoB;IACnC,gDAAgD;IAChD,WAAW,CAAC,EAAE,MAAM,CAAA;IACpB,gFAAgF;IAChF,WAAW,CAAC,EAAE,MAAM,CAAA;CACrB;AAED,iGAAiG;AACjG,wBAAgB,iBAAiB,CAAC,CAAC,EACjC,OAAO,EAAE,WAAW,CAAC,CAAC,CAAC,EACvB,IAAI,GAAE,oBAAyB,GAC9B,MAAM,CAgFR"}
|
|
@@ -0,0 +1,36 @@
|
|
|
1
|
+
/**
|
|
2
|
+
* Input-space tiling + coverage projection.
|
|
3
|
+
*
|
|
4
|
+
* Cells are the cartesian product of the input axes — the stratification plan,
|
|
5
|
+
* enumerable up front so the planned-vs-covered denominator is honest. Coverage
|
|
6
|
+
* is projected from the evaluation log: per cell, the full DISTRIBUTION of the
|
|
7
|
+
* headline score, of each scored dimension, and of evaluation latency — a bare
|
|
8
|
+
* mean hides outliers, so every aggregate carries its spread. Per-cell cost is
|
|
9
|
+
* split known-dollars vs unknown-runs, never folded into a fabricated $0.
|
|
10
|
+
*/
|
|
11
|
+
import type { BehaviorSpace, Cell, CoverageCell, Distribution, Evaluation } from './types';
|
|
12
|
+
/** One recorded evaluation — the unit coverage and the capsule are built from. */
|
|
13
|
+
export interface EvalRecord {
|
|
14
|
+
cell: Cell;
|
|
15
|
+
ev: Evaluation;
|
|
16
|
+
/** The objective's interest score for this evaluation. */
|
|
17
|
+
interest: number;
|
|
18
|
+
/** Evaluation wall-clock — engine-measured unless `ev.latencyMs` overrode it. */
|
|
19
|
+
latencyMs: number;
|
|
20
|
+
/** Known dollars for this run. `null` = cost tracking was wired but this
|
|
21
|
+
* run's cost was unknowable (counted apart). Absent = not tracked at all. */
|
|
22
|
+
costUsd?: number | null;
|
|
23
|
+
}
|
|
24
|
+
/** Enumerate every input cell (cartesian product of the axes), in stable order. */
|
|
25
|
+
export declare function enumerateCells(space: BehaviorSpace): Cell[];
|
|
26
|
+
/** Deterministic id for a coordinate map, e.g. `matterType=nda|difficulty=hard`. */
|
|
27
|
+
export declare function cellId(space: BehaviorSpace, coords: Record<string, string>): string;
|
|
28
|
+
/** Summarize a sample. Throws on an empty sample — callers represent "no data"
|
|
29
|
+
* as `null`, never as a zeroed distribution. */
|
|
30
|
+
export declare function distribution(values: number[]): Distribution;
|
|
31
|
+
/**
|
|
32
|
+
* Project the evaluation log into the per-input-cell coverage map. A cell with
|
|
33
|
+
* no evaluations reports `score: null` (honestly uncovered), never zeros.
|
|
34
|
+
*/
|
|
35
|
+
export declare function buildCoverage(cells: Cell[], log: EvalRecord[], threshold: number): CoverageCell[];
|
|
36
|
+
//# sourceMappingURL=cube.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"cube.d.ts","sourceRoot":"","sources":["../../src/fuzz/cube.ts"],"names":[],"mappings":"AAAA;;;;;;;;;GASG;AAEH,OAAO,KAAK,EAAE,aAAa,EAAE,IAAI,EAAE,YAAY,EAAE,YAAY,EAAE,UAAU,EAAE,MAAM,SAAS,CAAA;AAE1F,kFAAkF;AAClF,MAAM,WAAW,UAAU;IACzB,IAAI,EAAE,IAAI,CAAA;IACV,EAAE,EAAE,UAAU,CAAA;IACd,0DAA0D;IAC1D,QAAQ,EAAE,MAAM,CAAA;IAChB,iFAAiF;IACjF,SAAS,EAAE,MAAM,CAAA;IACjB;kFAC8E;IAC9E,OAAO,CAAC,EAAE,MAAM,GAAG,IAAI,CAAA;CACxB;AAED,mFAAmF;AACnF,wBAAgB,cAAc,CAAC,KAAK,EAAE,aAAa,GAAG,IAAI,EAAE,CAW3D;AAED,oFAAoF;AACpF,wBAAgB,MAAM,CAAC,KAAK,EAAE,aAAa,EAAE,MAAM,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,GAAG,MAAM,CAEnF;AAQD;iDACiD;AACjD,wBAAgB,YAAY,CAAC,MAAM,EAAE,MAAM,EAAE,GAAG,YAAY,CAa3D;AAED;;;GAGG;AACH,wBAAgB,aAAa,CAAC,KAAK,EAAE,IAAI,EAAE,EAAE,GAAG,EAAE,UAAU,EAAE,EAAE,SAAS,EAAE,MAAM,GAAG,YAAY,EAAE,CA0CjG"}
|