@tangle-network/agent-eval 0.95.0 → 0.96.0
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 +17 -0
- package/dist/adapters/http.d.ts +17 -10
- package/dist/adapters/langchain.d.ts +14 -7
- package/dist/adapters/otel.d.ts +25 -13
- package/dist/{rl/adversarial.d.ts → adversarial-DIVcDoI_.d.ts} +7 -6
- package/dist/analyst/index.d.ts +236 -28
- package/dist/analyst/index.js +1 -1
- package/dist/{trace-analyst/analyst.d.ts → analyst-C8HHvfJp.d.ts} +14 -12
- package/dist/{contract/analyze-runs.d.ts → analyze-runs-DtT6F_6T.d.ts} +10 -9
- package/dist/authenticity/index.d.ts +16 -15
- package/dist/{baseline.d.ts → baseline-Bbid3WoO.d.ts} +44 -8
- package/dist/belief-state/index.d.ts +605 -14
- package/dist/benchmarks/index.d.ts +5 -23
- package/dist/builder-eval/index.d.ts +250 -5
- package/dist/calibration-BPmzuVPk.d.ts +101 -0
- package/dist/campaign/index.d.ts +1452 -38
- package/dist/{chunk-AQ5WQAIV.js → chunk-3NHEO6ZC.js} +2 -2
- package/dist/chunk-3NHEO6ZC.js.map +1 -0
- package/dist/cli.d.ts +0 -2
- package/dist/{contract/intake/code-agent-session.d.ts → code-agent-session-CPHRCb4-.d.ts} +17 -15
- package/dist/contract/index.d.ts +824 -107
- package/dist/contract/index.js +145 -1
- package/dist/contract/index.js.map +1 -1
- package/dist/control-Doncu-B_.d.ts +259 -0
- package/dist/{control-runtime.d.ts → control-runtime-Acf9CGhw.d.ts} +26 -23
- package/dist/control.d.ts +10 -11
- package/dist/corpus-D4YW9UoJ.d.ts +560 -0
- package/dist/{cost-ledger.d.ts → cost-ledger-DuSqlw5B.d.ts} +11 -10
- package/dist/{counterfactual.d.ts → counterfactual-DlOz8PBx.d.ts} +13 -12
- package/dist/{dataset.d.ts → dataset-BbGkaN2I.d.ts} +14 -11
- package/dist/{analyst/registry.d.ts → default-registry-GyE8X5SP.d.ts} +37 -8
- package/dist/diagnose.d.ts +252 -1
- package/dist/{trace/emitter.d.ts → emitter-C2rqGH_l.d.ts} +12 -9
- package/dist/{errors.d.ts → errors-CzMUYo7b.d.ts} +11 -10
- package/dist/failure-cluster-DH9Flgcf.d.ts +76 -0
- package/dist/{feedback-trajectory.d.ts → feedback-trajectory-BxY0cKfs.d.ts} +38 -36
- package/dist/fuzz.d.ts +547 -1
- package/dist/gepa-Dprxvz8r.d.ts +414 -0
- package/dist/governance/index.d.ts +135 -5
- package/dist/harness-optimizer-mOl9XX_O.d.ts +106 -0
- package/dist/hosted/index.d.ts +239 -10
- package/dist/index-_Y4oNOOb.d.ts +159 -0
- package/dist/index.d.ts +5660 -277
- package/dist/index.js +1 -1
- package/dist/{contract/insight-report.d.ts → insight-report-BnRjTibG.d.ts} +19 -16
- package/dist/{trace/integrity.d.ts → integrity-D2t12mMw.d.ts} +14 -11
- package/dist/{judge-calibration.d.ts → judge-calibration-0p2QcWNE.d.ts} +17 -16
- package/dist/{analyst/kind-factory.d.ts → kind-factory-X3eDYbKn.d.ts} +61 -10
- package/dist/knowledge/index.d.ts +103 -3
- package/dist/{llm-client.d.ts → llm-client-Bj7g0rqu.d.ts} +23 -21
- package/dist/matrix/index.d.ts +30 -12
- package/dist/meta-eval/index.d.ts +182 -6
- package/dist/{multi-layer-verifier.d.ts → multi-layer-verifier-DUZXrPDA.d.ts} +15 -12
- package/dist/multishot/index.d.ts +290 -7
- package/dist/{rl/off-policy.d.ts → off-policy-DiwuKKg7.d.ts} +9 -8
- package/dist/openapi.json +1 -1
- package/dist/{meta-eval/outcome-store.d.ts → outcome-store-rnXLEqSn.d.ts} +8 -7
- package/dist/{pareto.d.ts → pareto-E-pembql.d.ts} +10 -9
- package/dist/perf/index.d.ts +119 -13
- package/dist/pipelines/index.d.ts +173 -8
- package/dist/pre-registration-nfUdc9EQ.d.ts +483 -0
- package/dist/prm/index.d.ts +104 -5
- package/dist/provenance-CIxfBnkl.d.ts +426 -0
- package/dist/query-B7GGjRox.d.ts +32 -0
- package/dist/{trace/raw-provider-sink.d.ts → raw-provider-sink-C46HDghv.d.ts} +14 -13
- package/dist/{red-team.d.ts → red-team-BWdoyleI.d.ts} +16 -14
- package/dist/{trace/redact.d.ts → redact-B40YG2M_.d.ts} +8 -7
- package/dist/release-report-pidWUMZ2.d.ts +233 -0
- package/dist/reporting.d.ts +16 -15
- package/dist/{eval-campaign.d.ts → researcher-Jr8ME1dZ.d.ts} +156 -34
- package/dist/rl.d.ts +1193 -1
- package/dist/{prm/rubric.d.ts → rubric-Cc6UHvUb.d.ts} +13 -10
- package/dist/{meta-eval/rubric-predictive-validity.d.ts → rubric-predictive-validity-C2hDKM8Z.d.ts} +11 -8
- package/dist/run-critic-CmMf05uV.d.ts +56 -0
- package/dist/{run-record.d.ts → run-record-CP2ObebC.d.ts} +117 -15
- package/dist/runtime-trajectory-BOUUjI0y.d.ts +49 -0
- package/dist/{trace/schema.d.ts → schema-m0gsnbt3.d.ts} +30 -29
- package/dist/semantic-concept-judge-DSBB2Cfp.d.ts +624 -0
- package/dist/{sequential.d.ts → sequential-5iSVfzl2.d.ts} +10 -9
- package/dist/{series-convergence.d.ts → series-convergence-D5OWMBg6.d.ts} +5 -4
- package/dist/sink-fetch-B1Yg4Til.d.ts +101 -0
- package/dist/{statistics.d.ts → statistics-CCJpTGOS.d.ts} +48 -46
- package/dist/{trace/store.d.ts → store-BcFXE6LG.d.ts} +12 -21
- package/dist/{trace-analyst/types.d.ts → store-C1YxJDEK.d.ts} +74 -18
- package/dist/storyboard/index.d.ts +81 -16
- package/dist/{summary-report.d.ts → summary-report-CInXwsza.d.ts} +160 -23
- package/dist/telemetry/{sink-file.d.ts → file.d.ts} +7 -5
- package/dist/telemetry/index.d.ts +35 -17
- package/dist/{sandbox-harness.d.ts → test-graded-scenario-DeODGLra.d.ts} +60 -15
- package/dist/{locked-jsonl-appender.d.ts → testing-C21CHsq2.d.ts} +4 -3
- package/dist/testing.d.ts +1 -5
- package/dist/traces.d.ts +976 -4
- package/dist/{trajectory.d.ts → trajectory-2TkpSEVh.d.ts} +9 -8
- package/dist/{analyst/types.d.ts → types-B5x54y6n.d.ts} +112 -19
- package/dist/{campaign/types.d.ts → types-BMahhhio.d.ts} +47 -44
- package/dist/{matrix/types.d.ts → types-BUxNaJ8c.d.ts} +11 -9
- package/dist/{types.d.ts → types-C7DGg5ex.d.ts} +33 -31
- package/dist/{verdict.d.ts → verdict-C9MlYujm.d.ts} +3 -2
- package/dist/wire/index.d.ts +570 -13
- package/dist/workflow/index.d.ts +496 -22
- package/package.json +2 -2
- package/dist/action-policy.d.ts +0 -24
- package/dist/action-policy.d.ts.map +0 -1
- package/dist/action-policy.test.d.ts +0 -2
- package/dist/action-policy.test.d.ts.map +0 -1
- package/dist/active-learning.d.ts +0 -41
- package/dist/active-learning.d.ts.map +0 -1
- package/dist/adapters/http.d.ts.map +0 -1
- package/dist/adapters/langchain.d.ts.map +0 -1
- package/dist/adapters/otel.d.ts.map +0 -1
- package/dist/agent-profile-cell.d.ts +0 -101
- package/dist/agent-profile-cell.d.ts.map +0 -1
- package/dist/agent-profile.d.ts +0 -27
- package/dist/agent-profile.d.ts.map +0 -1
- package/dist/agent-profile.test.d.ts +0 -2
- package/dist/agent-profile.test.d.ts.map +0 -1
- package/dist/analyst/adapters.d.ts +0 -62
- package/dist/analyst/adapters.d.ts.map +0 -1
- package/dist/analyst/analyst.test.d.ts +0 -2
- package/dist/analyst/analyst.test.d.ts.map +0 -1
- package/dist/analyst/ax-service.d.ts +0 -27
- package/dist/analyst/ax-service.d.ts.map +0 -1
- package/dist/analyst/behavioral-analyst.d.ts +0 -28
- package/dist/analyst/behavioral-analyst.d.ts.map +0 -1
- package/dist/analyst/chat-client.d.ts +0 -91
- package/dist/analyst/chat-client.d.ts.map +0 -1
- package/dist/analyst/default-registry.d.ts +0 -27
- package/dist/analyst/default-registry.d.ts.map +0 -1
- package/dist/analyst/default-registry.test.d.ts +0 -2
- package/dist/analyst/default-registry.test.d.ts.map +0 -1
- package/dist/analyst/finding-signature.d.ts +0 -48
- package/dist/analyst/finding-signature.d.ts.map +0 -1
- package/dist/analyst/finding-subject.d.ts +0 -146
- package/dist/analyst/finding-subject.d.ts.map +0 -1
- package/dist/analyst/finding-subject.test.d.ts +0 -2
- package/dist/analyst/finding-subject.test.d.ts.map +0 -1
- package/dist/analyst/findings-store.d.ts +0 -75
- package/dist/analyst/findings-store.d.ts.map +0 -1
- package/dist/analyst/index.d.ts.map +0 -1
- package/dist/analyst/kind-factory.d.ts.map +0 -1
- package/dist/analyst/kinds/failure-mode.d.ts +0 -19
- package/dist/analyst/kinds/failure-mode.d.ts.map +0 -1
- package/dist/analyst/kinds/improvement.d.ts +0 -23
- package/dist/analyst/kinds/improvement.d.ts.map +0 -1
- package/dist/analyst/kinds/index.d.ts +0 -22
- package/dist/analyst/kinds/index.d.ts.map +0 -1
- package/dist/analyst/kinds/kinds.test.d.ts +0 -2
- package/dist/analyst/kinds/kinds.test.d.ts.map +0 -1
- package/dist/analyst/kinds/knowledge-gap.d.ts +0 -28
- package/dist/analyst/kinds/knowledge-gap.d.ts.map +0 -1
- package/dist/analyst/kinds/knowledge-poisoning.d.ts +0 -22
- package/dist/analyst/kinds/knowledge-poisoning.d.ts.map +0 -1
- package/dist/analyst/kinds/skill-usage.d.ts +0 -84
- package/dist/analyst/kinds/skill-usage.d.ts.map +0 -1
- package/dist/analyst/kinds/skill-usage.test.d.ts +0 -2
- package/dist/analyst/kinds/skill-usage.test.d.ts.map +0 -1
- package/dist/analyst/parse-tolerant.d.ts +0 -26
- package/dist/analyst/parse-tolerant.d.ts.map +0 -1
- package/dist/analyst/parse-tolerant.test.d.ts +0 -2
- package/dist/analyst/parse-tolerant.test.d.ts.map +0 -1
- package/dist/analyst/registry.budget.test.d.ts +0 -2
- package/dist/analyst/registry.budget.test.d.ts.map +0 -1
- package/dist/analyst/registry.d.ts.map +0 -1
- package/dist/analyst/steer-firewall.d.ts +0 -35
- package/dist/analyst/steer-firewall.d.ts.map +0 -1
- package/dist/analyst/steer-firewall.test.d.ts +0 -2
- package/dist/analyst/steer-firewall.test.d.ts.map +0 -1
- package/dist/analyst/structure-findings.d.ts +0 -37
- package/dist/analyst/structure-findings.d.ts.map +0 -1
- package/dist/analyst/structure-findings.test.d.ts +0 -2
- package/dist/analyst/structure-findings.test.d.ts.map +0 -1
- package/dist/analyst/tool-groups.d.ts +0 -34
- package/dist/analyst/tool-groups.d.ts.map +0 -1
- package/dist/analyst/types.d.ts.map +0 -1
- package/dist/anti-slop.d.ts +0 -59
- package/dist/anti-slop.d.ts.map +0 -1
- package/dist/artifact-validator.d.ts +0 -74
- package/dist/artifact-validator.d.ts.map +0 -1
- package/dist/attestation.d.ts +0 -63
- package/dist/attestation.d.ts.map +0 -1
- package/dist/attestation.test.d.ts +0 -2
- package/dist/attestation.test.d.ts.map +0 -1
- package/dist/authenticity/index.d.ts.map +0 -1
- package/dist/authenticity/index.test.d.ts +0 -2
- package/dist/authenticity/index.test.d.ts.map +0 -1
- package/dist/auto-pr.d.ts +0 -120
- package/dist/auto-pr.d.ts.map +0 -1
- package/dist/baseline.d.ts.map +0 -1
- package/dist/behavior-dsl.d.ts +0 -73
- package/dist/behavior-dsl.d.ts.map +0 -1
- package/dist/belief-state/calibration.d.ts +0 -10
- package/dist/belief-state/calibration.d.ts.map +0 -1
- package/dist/belief-state/calibration.test.d.ts +0 -2
- package/dist/belief-state/calibration.test.d.ts.map +0 -1
- package/dist/belief-state/code-agent-corpus.d.ts +0 -66
- package/dist/belief-state/code-agent-corpus.d.ts.map +0 -1
- package/dist/belief-state/code-agent-corpus.test.d.ts +0 -2
- package/dist/belief-state/code-agent-corpus.test.d.ts.map +0 -1
- package/dist/belief-state/code-agent-evidence.d.ts +0 -22
- package/dist/belief-state/code-agent-evidence.d.ts.map +0 -1
- package/dist/belief-state/code-agent-evidence.test.d.ts +0 -2
- package/dist/belief-state/code-agent-evidence.test.d.ts.map +0 -1
- package/dist/belief-state/extract.d.ts +0 -7
- package/dist/belief-state/extract.d.ts.map +0 -1
- package/dist/belief-state/extract.test.d.ts +0 -2
- package/dist/belief-state/extract.test.d.ts.map +0 -1
- package/dist/belief-state/index.d.ts.map +0 -1
- package/dist/belief-state/ope.d.ts +0 -17
- package/dist/belief-state/ope.d.ts.map +0 -1
- package/dist/belief-state/ope.test.d.ts +0 -2
- package/dist/belief-state/ope.test.d.ts.map +0 -1
- package/dist/belief-state/phase0-measurement.d.ts +0 -55
- package/dist/belief-state/phase0-measurement.d.ts.map +0 -1
- package/dist/belief-state/report.d.ts +0 -17
- package/dist/belief-state/report.d.ts.map +0 -1
- package/dist/belief-state/report.test.d.ts +0 -2
- package/dist/belief-state/report.test.d.ts.map +0 -1
- package/dist/belief-state/research-evidence.d.ts +0 -23
- package/dist/belief-state/research-evidence.d.ts.map +0 -1
- package/dist/belief-state/research-evidence.test.d.ts +0 -2
- package/dist/belief-state/research-evidence.test.d.ts.map +0 -1
- package/dist/belief-state/runtime-benchmark-corpus.d.ts +0 -32
- package/dist/belief-state/runtime-benchmark-corpus.d.ts.map +0 -1
- package/dist/belief-state/runtime-hooks.d.ts +0 -87
- package/dist/belief-state/runtime-hooks.d.ts.map +0 -1
- package/dist/belief-state/runtime-hooks.test.d.ts +0 -2
- package/dist/belief-state/runtime-hooks.test.d.ts.map +0 -1
- package/dist/belief-state/selective.d.ts +0 -15
- package/dist/belief-state/selective.d.ts.map +0 -1
- package/dist/belief-state/selective.test.d.ts +0 -2
- package/dist/belief-state/selective.test.d.ts.map +0 -1
- package/dist/belief-state/shadow-probe.d.ts +0 -80
- package/dist/belief-state/shadow-probe.d.ts.map +0 -1
- package/dist/belief-state/shadow-probe.test.d.ts +0 -2
- package/dist/belief-state/shadow-probe.test.d.ts.map +0 -1
- package/dist/belief-state/types.d.ts +0 -195
- package/dist/belief-state/types.d.ts.map +0 -1
- package/dist/belief-state/types.test.d.ts +0 -2
- package/dist/belief-state/types.test.d.ts.map +0 -1
- package/dist/benchmark.d.ts +0 -14
- package/dist/benchmark.d.ts.map +0 -1
- package/dist/benchmarks/index.d.ts.map +0 -1
- package/dist/benchmarks/routing/dataset.d.ts +0 -34
- package/dist/benchmarks/routing/dataset.d.ts.map +0 -1
- package/dist/benchmarks/routing/index.d.ts +0 -34
- package/dist/benchmarks/routing/index.d.ts.map +0 -1
- package/dist/benchmarks/types.d.ts +0 -49
- package/dist/benchmarks/types.d.ts.map +0 -1
- package/dist/bisector.d.ts +0 -81
- package/dist/bisector.d.ts.map +0 -1
- package/dist/budget-guard.d.ts +0 -31
- package/dist/budget-guard.d.ts.map +0 -1
- package/dist/builder-eval/builder-session.d.ts +0 -111
- package/dist/builder-eval/builder-session.d.ts.map +0 -1
- package/dist/builder-eval/correlation.d.ts +0 -32
- package/dist/builder-eval/correlation.d.ts.map +0 -1
- package/dist/builder-eval/index.d.ts.map +0 -1
- package/dist/builder-eval/project-registry.d.ts +0 -51
- package/dist/builder-eval/project-registry.d.ts.map +0 -1
- package/dist/builder-eval/three-layer-eval.d.ts +0 -55
- package/dist/builder-eval/three-layer-eval.d.ts.map +0 -1
- package/dist/campaign/analyst-surface.d.ts +0 -108
- package/dist/campaign/analyst-surface.d.ts.map +0 -1
- package/dist/campaign/analyst-surface.test.d.ts +0 -2
- package/dist/campaign/analyst-surface.test.d.ts.map +0 -1
- package/dist/campaign/auto-pr.d.ts +0 -46
- package/dist/campaign/auto-pr.d.ts.map +0 -1
- package/dist/campaign/distillation/agreement-judge.d.ts +0 -69
- package/dist/campaign/distillation/agreement-judge.d.ts.map +0 -1
- package/dist/campaign/distillation/cli.d.ts +0 -35
- package/dist/campaign/distillation/cli.d.ts.map +0 -1
- package/dist/campaign/distillation/distillation.test.d.ts +0 -2
- package/dist/campaign/distillation/distillation.test.d.ts.map +0 -1
- package/dist/campaign/distillation/gold-scenarios.d.ts +0 -54
- package/dist/campaign/distillation/gold-scenarios.d.ts.map +0 -1
- package/dist/campaign/distillation/run-distillation.d.ts +0 -119
- package/dist/campaign/distillation/run-distillation.d.ts.map +0 -1
- package/dist/campaign/gates/compose.d.ts +0 -12
- package/dist/campaign/gates/compose.d.ts.map +0 -1
- package/dist/campaign/gates/default-production-gate.d.ts +0 -58
- package/dist/campaign/gates/default-production-gate.d.ts.map +0 -1
- package/dist/campaign/gates/heldout-gate.d.ts +0 -12
- package/dist/campaign/gates/heldout-gate.d.ts.map +0 -1
- package/dist/campaign/gates/promotion-policy.d.ts +0 -125
- package/dist/campaign/gates/promotion-policy.d.ts.map +0 -1
- package/dist/campaign/gates/promotion-policy.test.d.ts +0 -2
- package/dist/campaign/gates/promotion-policy.test.d.ts.map +0 -1
- package/dist/campaign/gates/sequential.d.ts +0 -146
- package/dist/campaign/gates/sequential.d.ts.map +0 -1
- package/dist/campaign/gates/sequential.test.d.ts +0 -2
- package/dist/campaign/gates/sequential.test.d.ts.map +0 -1
- package/dist/campaign/gates/statistical-heldout.d.ts +0 -99
- package/dist/campaign/gates/statistical-heldout.d.ts.map +0 -1
- package/dist/campaign/gates/statistical-heldout.test.d.ts +0 -2
- package/dist/campaign/gates/statistical-heldout.test.d.ts.map +0 -1
- package/dist/campaign/index.d.ts.map +0 -1
- package/dist/campaign/labeled-store/fs-adapter.d.ts +0 -59
- package/dist/campaign/labeled-store/fs-adapter.d.ts.map +0 -1
- package/dist/campaign/presets/compare-proposers.d.ts +0 -146
- package/dist/campaign/presets/compare-proposers.d.ts.map +0 -1
- package/dist/campaign/presets/playback.d.ts +0 -120
- package/dist/campaign/presets/playback.d.ts.map +0 -1
- package/dist/campaign/presets/playback.test.d.ts +0 -2
- package/dist/campaign/presets/playback.test.d.ts.map +0 -1
- package/dist/campaign/presets/run-eval.d.ts +0 -14
- package/dist/campaign/presets/run-eval.d.ts.map +0 -1
- package/dist/campaign/presets/run-improvement-loop.d.ts +0 -63
- package/dist/campaign/presets/run-improvement-loop.d.ts.map +0 -1
- package/dist/campaign/presets/run-improvement-loop.test.d.ts +0 -2
- package/dist/campaign/presets/run-improvement-loop.test.d.ts.map +0 -1
- package/dist/campaign/presets/run-optimization.d.ts +0 -92
- package/dist/campaign/presets/run-optimization.d.ts.map +0 -1
- package/dist/campaign/presets/run-profile-matrix.d.ts +0 -151
- package/dist/campaign/presets/run-profile-matrix.d.ts.map +0 -1
- package/dist/campaign/presets/run-skill-opt.d.ts +0 -96
- package/dist/campaign/presets/run-skill-opt.d.ts.map +0 -1
- package/dist/campaign/proposers/_findings-text.d.ts +0 -22
- package/dist/campaign/proposers/_findings-text.d.ts.map +0 -1
- package/dist/campaign/proposers/ace.d.ts +0 -33
- package/dist/campaign/proposers/ace.d.ts.map +0 -1
- package/dist/campaign/proposers/ace.test.d.ts +0 -2
- package/dist/campaign/proposers/ace.test.d.ts.map +0 -1
- package/dist/campaign/proposers/analysis-edit.d.ts +0 -32
- package/dist/campaign/proposers/analysis-edit.d.ts.map +0 -1
- package/dist/campaign/proposers/evolutionary.d.ts +0 -20
- package/dist/campaign/proposers/evolutionary.d.ts.map +0 -1
- package/dist/campaign/proposers/fapo.d.ts +0 -120
- package/dist/campaign/proposers/fapo.d.ts.map +0 -1
- package/dist/campaign/proposers/gepa.d.ts +0 -86
- package/dist/campaign/proposers/gepa.d.ts.map +0 -1
- package/dist/campaign/proposers/halo.d.ts +0 -44
- package/dist/campaign/proposers/halo.d.ts.map +0 -1
- package/dist/campaign/proposers/halo.test.d.ts +0 -2
- package/dist/campaign/proposers/halo.test.d.ts.map +0 -1
- package/dist/campaign/proposers/memory.d.ts +0 -47
- package/dist/campaign/proposers/memory.d.ts.map +0 -1
- package/dist/campaign/proposers/memory.test.d.ts +0 -2
- package/dist/campaign/proposers/memory.test.d.ts.map +0 -1
- package/dist/campaign/proposers/skill-opt.d.ts +0 -88
- package/dist/campaign/proposers/skill-opt.d.ts.map +0 -1
- package/dist/campaign/proposers/trace-analyst.d.ts +0 -48
- package/dist/campaign/proposers/trace-analyst.d.ts.map +0 -1
- package/dist/campaign/proposers/trace-analyst.test.d.ts +0 -2
- package/dist/campaign/proposers/trace-analyst.test.d.ts.map +0 -1
- package/dist/campaign/provenance.d.ts +0 -185
- package/dist/campaign/provenance.d.ts.map +0 -1
- package/dist/campaign/run-campaign.d.ts +0 -90
- package/dist/campaign/run-campaign.d.ts.map +0 -1
- package/dist/campaign/score-utils.d.ts +0 -26
- package/dist/campaign/score-utils.d.ts.map +0 -1
- package/dist/campaign/skill-patch.d.ts +0 -62
- package/dist/campaign/skill-patch.d.ts.map +0 -1
- package/dist/campaign/storage.d.ts +0 -38
- package/dist/campaign/storage.d.ts.map +0 -1
- package/dist/campaign/types.d.ts.map +0 -1
- package/dist/campaign/worktree/index.d.ts +0 -53
- package/dist/campaign/worktree/index.d.ts.map +0 -1
- package/dist/canary.d.ts +0 -101
- package/dist/canary.d.ts.map +0 -1
- package/dist/causal-attribution.d.ts +0 -45
- package/dist/causal-attribution.d.ts.map +0 -1
- package/dist/chunk-AQ5WQAIV.js.map +0 -1
- package/dist/ci-gate.d.ts +0 -44
- package/dist/ci-gate.d.ts.map +0 -1
- package/dist/cli.d.ts.map +0 -1
- package/dist/client.d.ts +0 -77
- package/dist/client.d.ts.map +0 -1
- package/dist/client.test.d.ts +0 -2
- package/dist/client.test.d.ts.map +0 -1
- package/dist/command-runner.d.ts +0 -74
- package/dist/command-runner.d.ts.map +0 -1
- package/dist/command-runner.test.d.ts +0 -2
- package/dist/command-runner.test.d.ts.map +0 -1
- package/dist/completion-verifier.d.ts +0 -147
- package/dist/completion-verifier.d.ts.map +0 -1
- package/dist/completion-verifier.test.d.ts +0 -9
- package/dist/completion-verifier.test.d.ts.map +0 -1
- package/dist/concurrency.d.ts +0 -23
- package/dist/concurrency.d.ts.map +0 -1
- package/dist/contamination-guard.d.ts +0 -81
- package/dist/contamination-guard.d.ts.map +0 -1
- package/dist/contract/analyze-runs.d.ts.map +0 -1
- package/dist/contract/define-agent-eval.d.ts +0 -52
- package/dist/contract/define-agent-eval.d.ts.map +0 -1
- package/dist/contract/diff.d.ts +0 -114
- package/dist/contract/diff.d.ts.map +0 -1
- package/dist/contract/index.d.ts.map +0 -1
- package/dist/contract/insight-report.d.ts.map +0 -1
- package/dist/contract/insight-types-fwd.d.ts +0 -7
- package/dist/contract/insight-types-fwd.d.ts.map +0 -1
- package/dist/contract/intake/agent-trace.d.ts +0 -97
- package/dist/contract/intake/agent-trace.d.ts.map +0 -1
- package/dist/contract/intake/code-agent-session.d.ts.map +0 -1
- package/dist/contract/intake/feedback-table.d.ts +0 -87
- package/dist/contract/intake/feedback-table.d.ts.map +0 -1
- package/dist/contract/intake/index.d.ts +0 -22
- package/dist/contract/intake/index.d.ts.map +0 -1
- package/dist/contract/intake/otel-spans.d.ts +0 -33
- package/dist/contract/intake/otel-spans.d.ts.map +0 -1
- package/dist/contract/self-improve.d.ts +0 -284
- package/dist/contract/self-improve.d.ts.map +0 -1
- package/dist/control-runtime.d.ts.map +0 -1
- package/dist/control-runtime.test.d.ts +0 -2
- package/dist/control-runtime.test.d.ts.map +0 -1
- package/dist/control.d.ts.map +0 -1
- package/dist/convergence.d.ts +0 -29
- package/dist/convergence.d.ts.map +0 -1
- package/dist/cost-ledger.d.ts.map +0 -1
- package/dist/cost-ledger.test.d.ts +0 -2
- package/dist/cost-ledger.test.d.ts.map +0 -1
- package/dist/cost-report.d.ts +0 -43
- package/dist/cost-report.d.ts.map +0 -1
- package/dist/cost-report.test.d.ts +0 -2
- package/dist/cost-report.test.d.ts.map +0 -1
- package/dist/cost-tracker.d.ts +0 -76
- package/dist/cost-tracker.d.ts.map +0 -1
- package/dist/counterfactual.d.ts.map +0 -1
- package/dist/cross-trace-diff.d.ts +0 -56
- package/dist/cross-trace-diff.d.ts.map +0 -1
- package/dist/dataset.d.ts.map +0 -1
- package/dist/deploy-gate-layer.d.ts +0 -125
- package/dist/deploy-gate-layer.d.ts.map +0 -1
- package/dist/deploy-gate-layer.test.d.ts +0 -2
- package/dist/deploy-gate-layer.test.d.ts.map +0 -1
- package/dist/description-length-gate.d.ts +0 -119
- package/dist/description-length-gate.d.ts.map +0 -1
- package/dist/detectors/edge.test.d.ts +0 -2
- package/dist/detectors/edge.test.d.ts.map +0 -1
- package/dist/detectors/index.d.ts +0 -81
- package/dist/detectors/index.d.ts.map +0 -1
- package/dist/detectors/index.test.d.ts +0 -2
- package/dist/detectors/index.test.d.ts.map +0 -1
- package/dist/diagnose/causal-sweep.d.ts +0 -100
- package/dist/diagnose/causal-sweep.d.ts.map +0 -1
- package/dist/diagnose/index.d.ts +0 -36
- package/dist/diagnose/index.d.ts.map +0 -1
- package/dist/diagnose/remediation.d.ts +0 -68
- package/dist/diagnose/remediation.d.ts.map +0 -1
- package/dist/diagnose/repair.d.ts +0 -77
- package/dist/diagnose/repair.d.ts.map +0 -1
- package/dist/discover-personas.d.ts +0 -35
- package/dist/discover-personas.d.ts.map +0 -1
- package/dist/driver.d.ts +0 -95
- package/dist/driver.d.ts.map +0 -1
- package/dist/driver.test.d.ts +0 -8
- package/dist/driver.test.d.ts.map +0 -1
- package/dist/dual-agent-bench.d.ts +0 -81
- package/dist/dual-agent-bench.d.ts.map +0 -1
- package/dist/error-count-extractor.d.ts +0 -47
- package/dist/error-count-extractor.d.ts.map +0 -1
- package/dist/error-count-extractor.test.d.ts +0 -2
- package/dist/error-count-extractor.test.d.ts.map +0 -1
- package/dist/errors.d.ts.map +0 -1
- package/dist/eval-campaign.d.ts.map +0 -1
- package/dist/eval-campaign.test.d.ts +0 -2
- package/dist/eval-campaign.test.d.ts.map +0 -1
- package/dist/eval-tools.d.ts +0 -55
- package/dist/eval-tools.d.ts.map +0 -1
- package/dist/eval-trace-store.d.ts +0 -107
- package/dist/eval-trace-store.d.ts.map +0 -1
- package/dist/eval-trace-store.test.d.ts +0 -2
- package/dist/eval-trace-store.test.d.ts.map +0 -1
- package/dist/executor.d.ts +0 -38
- package/dist/executor.d.ts.map +0 -1
- package/dist/executor.test.d.ts +0 -10
- package/dist/executor.test.d.ts.map +0 -1
- package/dist/experiment-tracker.d.ts +0 -178
- package/dist/experiment-tracker.d.ts.map +0 -1
- package/dist/experiment-tracker.test.d.ts +0 -2
- package/dist/experiment-tracker.test.d.ts.map +0 -1
- package/dist/failure-taxonomy.d.ts +0 -38
- package/dist/failure-taxonomy.d.ts.map +0 -1
- package/dist/feedback-trajectory.d.ts.map +0 -1
- package/dist/feedback-trajectory.test.d.ts +0 -2
- package/dist/feedback-trajectory.test.d.ts.map +0 -1
- package/dist/flow-layer.d.ts +0 -90
- package/dist/flow-layer.d.ts.map +0 -1
- package/dist/flow-layer.test.d.ts +0 -2
- package/dist/flow-layer.test.d.ts.map +0 -1
- package/dist/fuzz/capsule.d.ts +0 -46
- package/dist/fuzz/capsule.d.ts.map +0 -1
- package/dist/fuzz/cube.d.ts +0 -36
- package/dist/fuzz/cube.d.ts.map +0 -1
- package/dist/fuzz/explorer-cost.test.d.ts +0 -2
- package/dist/fuzz/explorer-cost.test.d.ts.map +0 -1
- package/dist/fuzz/explorer.d.ts +0 -64
- package/dist/fuzz/explorer.d.ts.map +0 -1
- package/dist/fuzz/fuzz-agent.d.ts +0 -16
- package/dist/fuzz/fuzz-agent.d.ts.map +0 -1
- package/dist/fuzz/fuzz-agent.test.d.ts +0 -2
- package/dist/fuzz/fuzz-agent.test.d.ts.map +0 -1
- package/dist/fuzz/gates.d.ts +0 -33
- package/dist/fuzz/gates.d.ts.map +0 -1
- package/dist/fuzz/index.d.ts +0 -26
- package/dist/fuzz/index.d.ts.map +0 -1
- package/dist/fuzz/policies.d.ts +0 -28
- package/dist/fuzz/policies.d.ts.map +0 -1
- package/dist/fuzz/tools.d.ts +0 -20
- package/dist/fuzz/tools.d.ts.map +0 -1
- package/dist/fuzz/types.d.ts +0 -307
- package/dist/fuzz/types.d.ts.map +0 -1
- package/dist/golden-matcher.d.ts +0 -71
- package/dist/golden-matcher.d.ts.map +0 -1
- package/dist/governance/eu-ai-act.d.ts +0 -37
- package/dist/governance/eu-ai-act.d.ts.map +0 -1
- package/dist/governance/index.d.ts.map +0 -1
- package/dist/governance/nist-ai-rmf.d.ts +0 -15
- package/dist/governance/nist-ai-rmf.d.ts.map +0 -1
- package/dist/governance/soc2.d.ts +0 -12
- package/dist/governance/soc2.d.ts.map +0 -1
- package/dist/governance/types.d.ts +0 -66
- package/dist/governance/types.d.ts.map +0 -1
- package/dist/harness-optimizer.d.ts +0 -82
- package/dist/harness-optimizer.d.ts.map +0 -1
- package/dist/held-out-gate.d.ts +0 -135
- package/dist/held-out-gate.d.ts.map +0 -1
- package/dist/hosted/client.d.ts +0 -73
- package/dist/hosted/client.d.ts.map +0 -1
- package/dist/hosted/from-env.test.d.ts +0 -8
- package/dist/hosted/from-env.test.d.ts.map +0 -1
- package/dist/hosted/index.d.ts.map +0 -1
- package/dist/hosted/types.d.ts +0 -159
- package/dist/hosted/types.d.ts.map +0 -1
- package/dist/index.d.ts.map +0 -1
- package/dist/integrity/backend-integrity.d.ts +0 -71
- package/dist/integrity/backend-integrity.d.ts.map +0 -1
- package/dist/integrity/preflight.d.ts +0 -72
- package/dist/integrity/preflight.d.ts.map +0 -1
- package/dist/integrity/preflight.test.d.ts +0 -2
- package/dist/integrity/preflight.test.d.ts.map +0 -1
- package/dist/integrity/single-backend.d.ts +0 -67
- package/dist/integrity/single-backend.d.ts.map +0 -1
- package/dist/intent-match-judge.d.ts +0 -69
- package/dist/intent-match-judge.d.ts.map +0 -1
- package/dist/intent-match-judge.test.d.ts +0 -2
- package/dist/intent-match-judge.test.d.ts.map +0 -1
- package/dist/judge-calibration.d.ts.map +0 -1
- package/dist/judge-ensemble.d.ts +0 -66
- package/dist/judge-ensemble.d.ts.map +0 -1
- package/dist/judge-ensemble.test.d.ts +0 -8
- package/dist/judge-ensemble.test.d.ts.map +0 -1
- package/dist/judge-families.d.ts +0 -38
- package/dist/judge-families.d.ts.map +0 -1
- package/dist/judge-panel.d.ts +0 -65
- package/dist/judge-panel.d.ts.map +0 -1
- package/dist/judge-retry.d.ts +0 -70
- package/dist/judge-retry.d.ts.map +0 -1
- package/dist/judge-runner.d.ts +0 -36
- package/dist/judge-runner.d.ts.map +0 -1
- package/dist/judge-runner.test.d.ts +0 -2
- package/dist/judge-runner.test.d.ts.map +0 -1
- package/dist/judges.d.ts +0 -74
- package/dist/judges.d.ts.map +0 -1
- package/dist/keyword-coverage-judge.d.ts +0 -89
- package/dist/keyword-coverage-judge.d.ts.map +0 -1
- package/dist/keyword-coverage-judge.test.d.ts +0 -2
- package/dist/keyword-coverage-judge.test.d.ts.map +0 -1
- package/dist/knowledge/index.d.ts.map +0 -1
- package/dist/knowledge/readiness.d.ts +0 -26
- package/dist/knowledge/readiness.d.ts.map +0 -1
- package/dist/knowledge/types.d.ts +0 -75
- package/dist/knowledge/types.d.ts.map +0 -1
- package/dist/live-proof.d.ts +0 -62
- package/dist/live-proof.d.ts.map +0 -1
- package/dist/llm-client.d.ts.map +0 -1
- package/dist/llm-client.test.d.ts +0 -2
- package/dist/llm-client.test.d.ts.map +0 -1
- package/dist/locked-jsonl-appender.d.ts.map +0 -1
- package/dist/matrix/aggregation.d.ts +0 -16
- package/dist/matrix/aggregation.d.ts.map +0 -1
- package/dist/matrix/index.d.ts.map +0 -1
- package/dist/matrix/runner.d.ts +0 -15
- package/dist/matrix/runner.d.ts.map +0 -1
- package/dist/matrix/types.d.ts.map +0 -1
- package/dist/meta-eval/calibration.d.ts +0 -47
- package/dist/meta-eval/calibration.d.ts.map +0 -1
- package/dist/meta-eval/correlation-study.d.ts +0 -53
- package/dist/meta-eval/correlation-study.d.ts.map +0 -1
- package/dist/meta-eval/index.d.ts.map +0 -1
- package/dist/meta-eval/outcome-store.d.ts.map +0 -1
- package/dist/meta-eval/rubric-predictive-validity.d.ts.map +0 -1
- package/dist/meta-eval/sentinel.d.ts +0 -169
- package/dist/meta-eval/sentinel.d.ts.map +0 -1
- package/dist/metrics.d.ts +0 -63
- package/dist/metrics.d.ts.map +0 -1
- package/dist/model-seats.d.ts +0 -71
- package/dist/model-seats.d.ts.map +0 -1
- package/dist/model-seats.test.d.ts +0 -2
- package/dist/model-seats.test.d.ts.map +0 -1
- package/dist/muffled-gate-scanner.d.ts +0 -102
- package/dist/muffled-gate-scanner.d.ts.map +0 -1
- package/dist/multi-layer-verifier.d.ts.map +0 -1
- package/dist/multi-layer-verifier.test.d.ts +0 -2
- package/dist/multi-layer-verifier.test.d.ts.map +0 -1
- package/dist/multi-toolchain-layer.d.ts +0 -80
- package/dist/multi-toolchain-layer.d.ts.map +0 -1
- package/dist/multi-toolchain-layer.test.d.ts +0 -2
- package/dist/multi-toolchain-layer.test.d.ts.map +0 -1
- package/dist/multishot/default-tools.d.ts +0 -34
- package/dist/multishot/default-tools.d.ts.map +0 -1
- package/dist/multishot/index.d.ts.map +0 -1
- package/dist/multishot/judges.d.ts +0 -32
- package/dist/multishot/judges.d.ts.map +0 -1
- package/dist/multishot/matrix.d.ts +0 -107
- package/dist/multishot/matrix.d.ts.map +0 -1
- package/dist/multishot/multishot.d.ts +0 -23
- package/dist/multishot/multishot.d.ts.map +0 -1
- package/dist/multishot/router.d.ts +0 -37
- package/dist/multishot/router.d.ts.map +0 -1
- package/dist/multishot/types.d.ts +0 -60
- package/dist/multishot/types.d.ts.map +0 -1
- package/dist/observability.d.ts +0 -71
- package/dist/observability.d.ts.map +0 -1
- package/dist/oracle.d.ts +0 -55
- package/dist/oracle.d.ts.map +0 -1
- package/dist/orthogonality.d.ts +0 -35
- package/dist/orthogonality.d.ts.map +0 -1
- package/dist/otel-pipeline.d.ts +0 -31
- package/dist/otel-pipeline.d.ts.map +0 -1
- package/dist/paraphrase.d.ts +0 -107
- package/dist/paraphrase.d.ts.map +0 -1
- package/dist/pareto.d.ts.map +0 -1
- package/dist/partition-held-out.d.ts +0 -70
- package/dist/partition-held-out.d.ts.map +0 -1
- package/dist/partition-held-out.test.d.ts +0 -2
- package/dist/partition-held-out.test.d.ts.map +0 -1
- package/dist/perf/index.d.ts.map +0 -1
- package/dist/perf/integrity.d.ts +0 -30
- package/dist/perf/integrity.d.ts.map +0 -1
- package/dist/perf/journey.d.ts +0 -45
- package/dist/perf/journey.d.ts.map +0 -1
- package/dist/perf/ratchet.d.ts +0 -47
- package/dist/perf/ratchet.d.ts.map +0 -1
- package/dist/pipelines/budget-breach.d.ts +0 -31
- package/dist/pipelines/budget-breach.d.ts.map +0 -1
- package/dist/pipelines/budget-breach.test.d.ts +0 -2
- package/dist/pipelines/budget-breach.test.d.ts.map +0 -1
- package/dist/pipelines/failure-cluster.d.ts +0 -38
- package/dist/pipelines/failure-cluster.d.ts.map +0 -1
- package/dist/pipelines/failure-cluster.test.d.ts +0 -2
- package/dist/pipelines/failure-cluster.test.d.ts.map +0 -1
- package/dist/pipelines/first-divergence.d.ts +0 -26
- package/dist/pipelines/first-divergence.d.ts.map +0 -1
- package/dist/pipelines/first-divergence.test.d.ts +0 -2
- package/dist/pipelines/first-divergence.test.d.ts.map +0 -1
- package/dist/pipelines/index.d.ts.map +0 -1
- package/dist/pipelines/judge-agreement.d.ts +0 -26
- package/dist/pipelines/judge-agreement.d.ts.map +0 -1
- package/dist/pipelines/judge-agreement.test.d.ts +0 -2
- package/dist/pipelines/judge-agreement.test.d.ts.map +0 -1
- package/dist/pipelines/regression.d.ts +0 -23
- package/dist/pipelines/regression.d.ts.map +0 -1
- package/dist/pipelines/regression.test.d.ts +0 -2
- package/dist/pipelines/regression.test.d.ts.map +0 -1
- package/dist/pipelines/stuck-loop.d.ts +0 -32
- package/dist/pipelines/stuck-loop.d.ts.map +0 -1
- package/dist/pipelines/stuck-loop.test.d.ts +0 -2
- package/dist/pipelines/stuck-loop.test.d.ts.map +0 -1
- package/dist/pipelines/tool-waste.d.ts +0 -34
- package/dist/pipelines/tool-waste.d.ts.map +0 -1
- package/dist/pipelines/tool-waste.test.d.ts +0 -2
- package/dist/pipelines/tool-waste.test.d.ts.map +0 -1
- package/dist/playbook.d.ts +0 -16
- package/dist/playbook.d.ts.map +0 -1
- package/dist/pr-review-benchmark.d.ts +0 -88
- package/dist/pr-review-benchmark.d.ts.map +0 -1
- package/dist/pr-review-benchmark.test.d.ts +0 -2
- package/dist/pr-review-benchmark.test.d.ts.map +0 -1
- package/dist/pre-registration.d.ts +0 -125
- package/dist/pre-registration.d.ts.map +0 -1
- package/dist/prm/builtin-rubrics.d.ts +0 -33
- package/dist/prm/builtin-rubrics.d.ts.map +0 -1
- package/dist/prm/index.d.ts.map +0 -1
- package/dist/prm/inference.d.ts +0 -29
- package/dist/prm/inference.d.ts.map +0 -1
- package/dist/prm/inference.test.d.ts +0 -2
- package/dist/prm/inference.test.d.ts.map +0 -1
- package/dist/prm/rubric.d.ts.map +0 -1
- package/dist/prm/training-export.d.ts +0 -38
- package/dist/prm/training-export.d.ts.map +0 -1
- package/dist/produced-state.d.ts +0 -63
- package/dist/produced-state.d.ts.map +0 -1
- package/dist/produced-state.test.d.ts +0 -8
- package/dist/produced-state.test.d.ts.map +0 -1
- package/dist/profile/baselines.d.ts +0 -37
- package/dist/profile/baselines.d.ts.map +0 -1
- package/dist/profile/index.d.ts +0 -105
- package/dist/profile/index.d.ts.map +0 -1
- package/dist/promotion-gate.d.ts +0 -93
- package/dist/promotion-gate.d.ts.map +0 -1
- package/dist/prompt-registry.d.ts +0 -41
- package/dist/prompt-registry.d.ts.map +0 -1
- package/dist/propose-review-control.d.ts +0 -49
- package/dist/propose-review-control.d.ts.map +0 -1
- package/dist/propose-review-control.test.d.ts +0 -2
- package/dist/propose-review-control.test.d.ts.map +0 -1
- package/dist/propose-review.d.ts +0 -155
- package/dist/propose-review.d.ts.map +0 -1
- package/dist/red-team.d.ts.map +0 -1
- package/dist/reference-replay-steering.d.ts +0 -11
- package/dist/reference-replay-steering.d.ts.map +0 -1
- package/dist/reference-replay.d.ts +0 -176
- package/dist/reference-replay.d.ts.map +0 -1
- package/dist/reflective-mutation.d.ts +0 -79
- package/dist/reflective-mutation.d.ts.map +0 -1
- package/dist/registry.d.ts +0 -31
- package/dist/registry.d.ts.map +0 -1
- package/dist/release-confidence.d.ts +0 -128
- package/dist/release-confidence.d.ts.map +0 -1
- package/dist/release-report.d.ts +0 -11
- package/dist/release-report.d.ts.map +0 -1
- package/dist/replay.d.ts +0 -120
- package/dist/replay.d.ts.map +0 -1
- package/dist/reporter.d.ts +0 -14
- package/dist/reporter.d.ts.map +0 -1
- package/dist/reporting.d.ts.map +0 -1
- package/dist/researcher.d.ts +0 -140
- package/dist/researcher.d.ts.map +0 -1
- package/dist/reviewer.d.ts +0 -118
- package/dist/reviewer.d.ts.map +0 -1
- package/dist/reviewer.test.d.ts +0 -2
- package/dist/reviewer.test.d.ts.map +0 -1
- package/dist/reward-model-export.d.ts +0 -60
- package/dist/reward-model-export.d.ts.map +0 -1
- package/dist/rl/active-curriculum.d.ts +0 -110
- package/dist/rl/active-curriculum.d.ts.map +0 -1
- package/dist/rl/adaptation-eval.d.ts +0 -109
- package/dist/rl/adaptation-eval.d.ts.map +0 -1
- package/dist/rl/adversarial.d.ts.map +0 -1
- package/dist/rl/compute-curves.d.ts +0 -127
- package/dist/rl/compute-curves.d.ts.map +0 -1
- package/dist/rl/contamination.d.ts +0 -117
- package/dist/rl/contamination.d.ts.map +0 -1
- package/dist/rl/corpus.d.ts +0 -55
- package/dist/rl/corpus.d.ts.map +0 -1
- package/dist/rl/corpus.test.d.ts +0 -2
- package/dist/rl/corpus.test.d.ts.map +0 -1
- package/dist/rl/dataset.d.ts +0 -102
- package/dist/rl/dataset.d.ts.map +0 -1
- package/dist/rl/dataset.test.d.ts +0 -2
- package/dist/rl/dataset.test.d.ts.map +0 -1
- package/dist/rl/exporters.d.ts +0 -141
- package/dist/rl/exporters.d.ts.map +0 -1
- package/dist/rl/index.d.ts +0 -49
- package/dist/rl/index.d.ts.map +0 -1
- package/dist/rl/off-policy.d.ts.map +0 -1
- package/dist/rl/predictive-validity-researcher.d.ts +0 -69
- package/dist/rl/predictive-validity-researcher.d.ts.map +0 -1
- package/dist/rl/preferences.d.ts +0 -141
- package/dist/rl/preferences.d.ts.map +0 -1
- package/dist/rl/process-reward.d.ts +0 -122
- package/dist/rl/process-reward.d.ts.map +0 -1
- package/dist/rl/reward-hacking.d.ts +0 -104
- package/dist/rl/reward-hacking.d.ts.map +0 -1
- package/dist/rl/rl-campaign.d.ts +0 -85
- package/dist/rl/rl-campaign.d.ts.map +0 -1
- package/dist/rl/run-record-adapters.d.ts +0 -56
- package/dist/rl/run-record-adapters.d.ts.map +0 -1
- package/dist/rl/sim-fidelity.d.ts +0 -166
- package/dist/rl/sim-fidelity.d.ts.map +0 -1
- package/dist/rl/sim-fidelity.test.d.ts +0 -2
- package/dist/rl/sim-fidelity.test.d.ts.map +0 -1
- package/dist/rl/tournament.d.ts +0 -115
- package/dist/rl/tournament.d.ts.map +0 -1
- package/dist/rl/verifiable-reward.d.ts +0 -124
- package/dist/rl/verifiable-reward.d.ts.map +0 -1
- package/dist/run-critic.d.ts +0 -23
- package/dist/run-critic.d.ts.map +0 -1
- package/dist/run-evidence.d.ts +0 -32
- package/dist/run-evidence.d.ts.map +0 -1
- package/dist/run-record.d.ts.map +0 -1
- package/dist/run-record.test.d.ts +0 -2
- package/dist/run-record.test.d.ts.map +0 -1
- package/dist/run-score.d.ts +0 -31
- package/dist/run-score.d.ts.map +0 -1
- package/dist/runtime-trajectory.d.ts +0 -47
- package/dist/runtime-trajectory.d.ts.map +0 -1
- package/dist/sandbox-harness.d.ts.map +0 -1
- package/dist/sandbox-harness.test.d.ts +0 -2
- package/dist/sandbox-harness.test.d.ts.map +0 -1
- package/dist/sandbox-pool.d.ts +0 -74
- package/dist/sandbox-pool.d.ts.map +0 -1
- package/dist/sandbox-pool.test.d.ts +0 -2
- package/dist/sandbox-pool.test.d.ts.map +0 -1
- package/dist/scorecard.d.ts +0 -133
- package/dist/scorecard.d.ts.map +0 -1
- package/dist/scorecard.test.d.ts +0 -2
- package/dist/scorecard.test.d.ts.map +0 -1
- package/dist/self-play.d.ts +0 -69
- package/dist/self-play.d.ts.map +0 -1
- package/dist/semantic-concept-judge.d.ts +0 -135
- package/dist/semantic-concept-judge.d.ts.map +0 -1
- package/dist/semantic-concept-judge.test.d.ts +0 -2
- package/dist/semantic-concept-judge.test.d.ts.map +0 -1
- package/dist/sequential.d.ts.map +0 -1
- package/dist/series-convergence.d.ts.map +0 -1
- package/dist/slo.d.ts +0 -48
- package/dist/slo.d.ts.map +0 -1
- package/dist/state-continuity.d.ts +0 -47
- package/dist/state-continuity.d.ts.map +0 -1
- package/dist/statistics.d.ts.map +0 -1
- package/dist/statistics.test.d.ts +0 -2
- package/dist/statistics.test.d.ts.map +0 -1
- package/dist/steering-optimizer.d.ts +0 -58
- package/dist/steering-optimizer.d.ts.map +0 -1
- package/dist/steering.d.ts +0 -24
- package/dist/steering.d.ts.map +0 -1
- package/dist/storyboard/code-edit.d.ts +0 -64
- package/dist/storyboard/code-edit.d.ts.map +0 -1
- package/dist/storyboard/code-edit.test.d.ts +0 -2
- package/dist/storyboard/code-edit.test.d.ts.map +0 -1
- package/dist/storyboard/index.d.ts.map +0 -1
- package/dist/storyboard/index.test.d.ts +0 -2
- package/dist/storyboard/index.test.d.ts.map +0 -1
- package/dist/summary-report.d.ts.map +0 -1
- package/dist/telemetry/client.d.ts +0 -35
- package/dist/telemetry/client.d.ts.map +0 -1
- package/dist/telemetry/index.d.ts.map +0 -1
- package/dist/telemetry/schema.d.ts +0 -61
- package/dist/telemetry/schema.d.ts.map +0 -1
- package/dist/telemetry/sink-fetch.d.ts +0 -39
- package/dist/telemetry/sink-fetch.d.ts.map +0 -1
- package/dist/telemetry/sink-file.d.ts.map +0 -1
- package/dist/test-graded-scenario.d.ts +0 -42
- package/dist/test-graded-scenario.d.ts.map +0 -1
- package/dist/testing.d.ts.map +0 -1
- package/dist/tool-use-metrics.d.ts +0 -35
- package/dist/tool-use-metrics.d.ts.map +0 -1
- package/dist/trace/capture-fetch.d.ts +0 -48
- package/dist/trace/capture-fetch.d.ts.map +0 -1
- package/dist/trace/capture-fetch.test.d.ts +0 -2
- package/dist/trace/capture-fetch.test.d.ts.map +0 -1
- package/dist/trace/emitter.d.ts.map +0 -1
- package/dist/trace/extract-usage.d.ts +0 -46
- package/dist/trace/extract-usage.d.ts.map +0 -1
- package/dist/trace/extract-usage.test.d.ts +0 -2
- package/dist/trace/extract-usage.test.d.ts.map +0 -1
- package/dist/trace/index.d.ts +0 -15
- package/dist/trace/index.d.ts.map +0 -1
- package/dist/trace/integrity.d.ts.map +0 -1
- package/dist/trace/otel-bridge.d.ts +0 -29
- package/dist/trace/otel-bridge.d.ts.map +0 -1
- package/dist/trace/otel-export.d.ts +0 -52
- package/dist/trace/otel-export.d.ts.map +0 -1
- package/dist/trace/otel.d.ts +0 -57
- package/dist/trace/otel.d.ts.map +0 -1
- package/dist/trace/otlp-attributes.d.ts +0 -17
- package/dist/trace/otlp-attributes.d.ts.map +0 -1
- package/dist/trace/query.d.ts +0 -29
- package/dist/trace/query.d.ts.map +0 -1
- package/dist/trace/query.test.d.ts +0 -2
- package/dist/trace/query.test.d.ts.map +0 -1
- package/dist/trace/raw-provider-sink.d.ts.map +0 -1
- package/dist/trace/redact.d.ts.map +0 -1
- package/dist/trace/schema.d.ts.map +0 -1
- package/dist/trace/store-to-otlp.d.ts +0 -72
- package/dist/trace/store-to-otlp.d.ts.map +0 -1
- package/dist/trace/store-to-otlp.test.d.ts +0 -2
- package/dist/trace/store-to-otlp.test.d.ts.map +0 -1
- package/dist/trace/store.d.ts.map +0 -1
- package/dist/trace/store.test.d.ts +0 -2
- package/dist/trace/store.test.d.ts.map +0 -1
- package/dist/trace-analyst/analyst.d.ts.map +0 -1
- package/dist/trace-analyst/analyst.test.d.ts +0 -2
- package/dist/trace-analyst/analyst.test.d.ts.map +0 -1
- package/dist/trace-analyst/behavioral-metrics.d.ts +0 -40
- package/dist/trace-analyst/behavioral-metrics.d.ts.map +0 -1
- package/dist/trace-analyst/behavioral-metrics.test.d.ts +0 -2
- package/dist/trace-analyst/behavioral-metrics.test.d.ts.map +0 -1
- package/dist/trace-analyst/hook.d.ts +0 -55
- package/dist/trace-analyst/hook.d.ts.map +0 -1
- package/dist/trace-analyst/index.d.ts +0 -18
- package/dist/trace-analyst/index.d.ts.map +0 -1
- package/dist/trace-analyst/insights.d.ts +0 -71
- package/dist/trace-analyst/insights.d.ts.map +0 -1
- package/dist/trace-analyst/insights.test.d.ts +0 -2
- package/dist/trace-analyst/insights.test.d.ts.map +0 -1
- package/dist/trace-analyst/otlp-flatten.d.ts +0 -42
- package/dist/trace-analyst/otlp-flatten.d.ts.map +0 -1
- package/dist/trace-analyst/otlp-span.d.ts +0 -85
- package/dist/trace-analyst/otlp-span.d.ts.map +0 -1
- package/dist/trace-analyst/otlp-span.test.d.ts +0 -2
- package/dist/trace-analyst/otlp-span.test.d.ts.map +0 -1
- package/dist/trace-analyst/otlp-to-run-records.d.ts +0 -115
- package/dist/trace-analyst/otlp-to-run-records.d.ts.map +0 -1
- package/dist/trace-analyst/otlp-to-run-records.test.d.ts +0 -2
- package/dist/trace-analyst/otlp-to-run-records.test.d.ts.map +0 -1
- package/dist/trace-analyst/otlp-to-run-records.timestamps.test.d.ts +0 -2
- package/dist/trace-analyst/otlp-to-run-records.timestamps.test.d.ts.map +0 -1
- package/dist/trace-analyst/prompts.d.ts +0 -6
- package/dist/trace-analyst/prompts.d.ts.map +0 -1
- package/dist/trace-analyst/store-otlp.d.ts +0 -126
- package/dist/trace-analyst/store-otlp.d.ts.map +0 -1
- package/dist/trace-analyst/store-otlp.test.d.ts +0 -8
- package/dist/trace-analyst/store-otlp.test.d.ts.map +0 -1
- package/dist/trace-analyst/store-otlp.timestamps.test.d.ts +0 -2
- package/dist/trace-analyst/store-otlp.timestamps.test.d.ts.map +0 -1
- package/dist/trace-analyst/store.d.ts +0 -63
- package/dist/trace-analyst/store.d.ts.map +0 -1
- package/dist/trace-analyst/tools.d.ts +0 -44
- package/dist/trace-analyst/tools.d.ts.map +0 -1
- package/dist/trace-analyst/tools.test.d.ts +0 -10
- package/dist/trace-analyst/tools.test.d.ts.map +0 -1
- package/dist/trace-analyst/types.d.ts.map +0 -1
- package/dist/trace-contracts.d.ts +0 -180
- package/dist/trace-contracts.d.ts.map +0 -1
- package/dist/traced-analyst.d.ts +0 -26
- package/dist/traced-analyst.d.ts.map +0 -1
- package/dist/traced-judges.d.ts +0 -27
- package/dist/traced-judges.d.ts.map +0 -1
- package/dist/traces.d.ts.map +0 -1
- package/dist/trajectory.d.ts.map +0 -1
- package/dist/types.d.ts.map +0 -1
- package/dist/ui-finding.d.ts +0 -104
- package/dist/ui-finding.d.ts.map +0 -1
- package/dist/verdict-cache.d.ts +0 -78
- package/dist/verdict-cache.d.ts.map +0 -1
- package/dist/verdict-cache.test.d.ts +0 -2
- package/dist/verdict-cache.test.d.ts.map +0 -1
- package/dist/verdict.d.ts.map +0 -1
- package/dist/visual-diff.d.ts +0 -32
- package/dist/visual-diff.d.ts.map +0 -1
- package/dist/wire/handlers.d.ts +0 -54
- package/dist/wire/handlers.d.ts.map +0 -1
- package/dist/wire/index.d.ts.map +0 -1
- package/dist/wire/openapi.d.ts +0 -3
- package/dist/wire/openapi.d.ts.map +0 -1
- package/dist/wire/rpc.d.ts +0 -21
- package/dist/wire/rpc.d.ts.map +0 -1
- package/dist/wire/rubrics.d.ts +0 -34
- package/dist/wire/rubrics.d.ts.map +0 -1
- package/dist/wire/schemas.d.ts +0 -410
- package/dist/wire/schemas.d.ts.map +0 -1
- package/dist/wire/server.d.ts +0 -60
- package/dist/wire/server.d.ts.map +0 -1
- package/dist/workflow/event-schema.d.ts +0 -5
- package/dist/workflow/event-schema.d.ts.map +0 -1
- package/dist/workflow/feedback-pack.d.ts +0 -99
- package/dist/workflow/feedback-pack.d.ts.map +0 -1
- package/dist/workflow/index.d.ts.map +0 -1
- package/dist/workflow/intelligence-export.d.ts +0 -62
- package/dist/workflow/intelligence-export.d.ts.map +0 -1
- package/dist/workflow/partner-report.d.ts +0 -49
- package/dist/workflow/partner-report.d.ts.map +0 -1
- package/dist/workflow/phase-graph.d.ts +0 -43
- package/dist/workflow/phase-graph.d.ts.map +0 -1
- package/dist/workflow/promotion-gate.d.ts +0 -61
- package/dist/workflow/promotion-gate.d.ts.map +0 -1
- package/dist/workflow/run-record.d.ts +0 -12
- package/dist/workflow/run-record.d.ts.map +0 -1
- package/dist/workflow/runtime-adapter.d.ts +0 -20
- package/dist/workflow/runtime-adapter.d.ts.map +0 -1
- package/dist/workflow/sanitize.d.ts +0 -21
- package/dist/workflow/sanitize.d.ts.map +0 -1
- package/dist/workflow/schema.d.ts +0 -5
- package/dist/workflow/schema.d.ts.map +0 -1
- package/dist/workflow/summary.d.ts +0 -43
- package/dist/workflow/summary.d.ts.map +0 -1
- package/dist/workflow/trace-event-fields.d.ts +0 -6
- package/dist/workflow/trace-event-fields.d.ts.map +0 -1
- package/dist/workflow/trajectory.d.ts +0 -15
- package/dist/workflow/trajectory.d.ts.map +0 -1
- package/dist/workflow/types.d.ts +0 -68
- package/dist/workflow/types.d.ts.map +0 -1
- package/dist/workspace-inspector.d.ts +0 -67
- package/dist/workspace-inspector.d.ts.map +0 -1
- package/dist/wrangler-deploy-runner.test.d.ts +0 -2
- package/dist/wrangler-deploy-runner.test.d.ts.map +0 -1
|
@@ -1,185 +0,0 @@
|
|
|
1
|
-
/**
|
|
2
|
-
* Loop provenance — the durable, queryable record of WHAT a self-improvement
|
|
3
|
-
* loop did and WHY, plus the OTel spans that let an OTLP collector pivot from
|
|
4
|
-
* an eval-run to the underlying candidate→cell→gate→promote chain.
|
|
5
|
-
*
|
|
6
|
-
* Two artifacts, one source of truth:
|
|
7
|
-
*
|
|
8
|
-
* 1. `LoopProvenanceRecord` — a structured JSON record capturing every
|
|
9
|
-
* candidate (surfaceHash + label + rationale), its measured composite,
|
|
10
|
-
* the gate decision + reasons + delta, the held-out lift, the explicit
|
|
11
|
-
* baseline→candidate diff, and BACKEND PROVENANCE (the
|
|
12
|
-
* `assertRealBackend` verdict + worker call count + model). This is the
|
|
13
|
-
* ingestable audit artifact: the +lift recomputes from it, the "because
|
|
14
|
-
* Z" rationale survives in it, and a stub backend is detectable from it.
|
|
15
|
-
*
|
|
16
|
-
* 2. `loopProvenanceSpans()` — the same chain emitted as OTLP-ingestable
|
|
17
|
-
* `TraceSpanEvent`s, pivoted on the substrate's standard
|
|
18
|
-
* `tangle.runId` / `tangle.scenarioId` / `tangle.cellId` /
|
|
19
|
-
* `tangle.generation` attributes (the same pivots `/adapters/otel`
|
|
20
|
-
* reads). The hosted `/v1/ingest/traces` endpoint receives the FULL loop,
|
|
21
|
-
* not just the `cost.*` spans `runCampaign` already emits per cell.
|
|
22
|
-
*
|
|
23
|
-
* The record is built from the substrate's own loop result + the per-call
|
|
24
|
-
* `RunRecord`s the worker emitted — no new measurement, no recomputation that
|
|
25
|
-
* could drift from what the gate actually saw.
|
|
26
|
-
*/
|
|
27
|
-
import type { HostedClient } from '../hosted/client';
|
|
28
|
-
import type { TraceSpanEvent } from '../hosted/types';
|
|
29
|
-
import type { RunRecord } from '../run-record';
|
|
30
|
-
import type { CampaignStorage } from './storage';
|
|
31
|
-
import type { CampaignResult, GateDecision, GateResult, MutableSurface, Scenario } from './types';
|
|
32
|
-
/** Stable sha256 (full hex) of a surface's effective text. Code surfaces hash
|
|
33
|
-
* their worktree+base identity since the content lives in git. Distinct from
|
|
34
|
-
* `surfaceHash` (16-char content fingerprint used as a loop identity key);
|
|
35
|
-
* this is the byte-identical-verifiable content hash the provenance record +
|
|
36
|
-
* `RunRecord.promptHash` carry. */
|
|
37
|
-
export declare function surfaceContentHash(surface: MutableSurface): string;
|
|
38
|
-
export interface LoopProvenanceCandidate {
|
|
39
|
-
/** Generation index this candidate was proposed in. */
|
|
40
|
-
generation: number;
|
|
41
|
-
/** 16-char loop-identity fingerprint (matches `GenerationCandidate.surfaceHash`). */
|
|
42
|
-
surfaceHash: string;
|
|
43
|
-
/** Full sha256 content hash — byte-identical-verifiable. */
|
|
44
|
-
contentHash: string;
|
|
45
|
-
/** Proposer label, when the proposer returned a `ProposedCandidate`. */
|
|
46
|
-
label?: string;
|
|
47
|
-
/** Proposer rationale — the "because Z". When the proposer returned a bare
|
|
48
|
-
* surface (blind mutator) this is absent. */
|
|
49
|
-
rationale?: string;
|
|
50
|
-
/** Mean composite this candidate scored on the search split. */
|
|
51
|
-
composite: number;
|
|
52
|
-
/** Whether this candidate was promoted out of its generation. */
|
|
53
|
-
promoted: boolean;
|
|
54
|
-
}
|
|
55
|
-
export interface LoopProvenanceBackend {
|
|
56
|
-
/** `assertRealBackend`-grade verdict over the worker call records. */
|
|
57
|
-
verdict: 'real' | 'mixed' | 'stub';
|
|
58
|
-
/** Number of worker LLM calls captured (the audit's "worker call count"). */
|
|
59
|
-
workerCallCount: number;
|
|
60
|
-
/** Distinct model ids observed across worker calls. */
|
|
61
|
-
models: string[];
|
|
62
|
-
totalInputTokens: number;
|
|
63
|
-
totalOutputTokens: number;
|
|
64
|
-
totalCostUsd: number;
|
|
65
|
-
}
|
|
66
|
-
/**
|
|
67
|
-
* The durable provenance record. Aligns to the hosted `EvalRunEvent` path but
|
|
68
|
-
* ADDS the rationale + the explicit baseline→candidate diff (both omitted from
|
|
69
|
-
* the bare hosted event) + backend provenance.
|
|
70
|
-
*/
|
|
71
|
-
export interface LoopProvenanceRecord {
|
|
72
|
-
schema: 'tangle.loop-provenance.v1';
|
|
73
|
-
runId: string;
|
|
74
|
-
runDir: string;
|
|
75
|
-
timestamp: string;
|
|
76
|
-
/** Baseline + winner surface content hashes — distinguishable, byte-verifiable. */
|
|
77
|
-
baselineContentHash: string;
|
|
78
|
-
winnerContentHash: string;
|
|
79
|
-
/** Proposer label/rationale for the promoted change. Absent ⇒ winner == baseline. */
|
|
80
|
-
winnerLabel?: string;
|
|
81
|
-
winnerRationale?: string;
|
|
82
|
-
/** The explicit baseline→winner unified diff the gate decided on. */
|
|
83
|
-
diff: string;
|
|
84
|
-
/** Every candidate across every generation, each carrying its rationale. */
|
|
85
|
-
candidates: LoopProvenanceCandidate[];
|
|
86
|
-
/** The gate verdict — decision + reasons + contributing gates + delta. */
|
|
87
|
-
gate: {
|
|
88
|
-
decision: GateDecision;
|
|
89
|
-
reasons: string[];
|
|
90
|
-
delta?: number;
|
|
91
|
-
contributingGates: Array<{
|
|
92
|
-
name: string;
|
|
93
|
-
passed: boolean;
|
|
94
|
-
}>;
|
|
95
|
-
};
|
|
96
|
-
/** baseline-on-holdout composite mean. */
|
|
97
|
-
baselineHoldoutComposite: number;
|
|
98
|
-
/** winner-on-holdout composite mean. */
|
|
99
|
-
winnerHoldoutComposite: number;
|
|
100
|
-
/** winnerHoldout - baselineHoldout — RECOMPUTABLE from this record. */
|
|
101
|
-
heldOutLift: number;
|
|
102
|
-
/** Backend provenance: stub-vs-real verdict + worker call count + models. */
|
|
103
|
-
backend: LoopProvenanceBackend;
|
|
104
|
-
totalCostUsd: number;
|
|
105
|
-
totalDurationMs: number;
|
|
106
|
-
}
|
|
107
|
-
export interface BuildLoopProvenanceArgs<TArtifact, TScenario extends Scenario> {
|
|
108
|
-
runId: string;
|
|
109
|
-
runDir: string;
|
|
110
|
-
timestamp: string;
|
|
111
|
-
baselineSurface: MutableSurface;
|
|
112
|
-
winnerSurface: MutableSurface;
|
|
113
|
-
winnerLabel?: string;
|
|
114
|
-
winnerRationale?: string;
|
|
115
|
-
diff: string;
|
|
116
|
-
/** Per-generation candidate records straight off the loop result. */
|
|
117
|
-
generations: Array<{
|
|
118
|
-
generationIndex: number;
|
|
119
|
-
candidates: Array<{
|
|
120
|
-
surfaceHash: string;
|
|
121
|
-
composite: number;
|
|
122
|
-
label?: string;
|
|
123
|
-
rationale?: string;
|
|
124
|
-
}>;
|
|
125
|
-
promoted: string[];
|
|
126
|
-
/** Surfaces measured this generation, keyed positionally to candidates so
|
|
127
|
-
* the content hash can be computed from the real surface text. */
|
|
128
|
-
surfaces: Array<{
|
|
129
|
-
surfaceHash: string;
|
|
130
|
-
surface: MutableSurface;
|
|
131
|
-
}>;
|
|
132
|
-
}>;
|
|
133
|
-
gate: GateResult;
|
|
134
|
-
baselineOnHoldout: CampaignResult<TArtifact, TScenario>;
|
|
135
|
-
winnerOnHoldout: CampaignResult<TArtifact, TScenario>;
|
|
136
|
-
/** Worker call records — the source for backend provenance. */
|
|
137
|
-
workerRecords: ReadonlyArray<RunRecord>;
|
|
138
|
-
totalCostUsd: number;
|
|
139
|
-
totalDurationMs: number;
|
|
140
|
-
}
|
|
141
|
-
/** Build the durable provenance record from a completed loop result. */
|
|
142
|
-
export declare function buildLoopProvenanceRecord<TArtifact, TScenario extends Scenario>(args: BuildLoopProvenanceArgs<TArtifact, TScenario>): LoopProvenanceRecord;
|
|
143
|
-
/**
|
|
144
|
-
* Build the loop's OTLP-ingestable spans from a provenance record. One root
|
|
145
|
-
* span per loop (`tangle.runId`), one span per generation, one span per
|
|
146
|
-
* candidate (carrying its surfaceHash + label), and one span for the gate
|
|
147
|
-
* decision (carrying reasons + delta + lift). Candidate + gate spans pivot on
|
|
148
|
-
* the same `tangle.runId` / `tangle.generation` attributes `/adapters/otel`
|
|
149
|
-
* reads, so the hosted collector reconstructs the full tree.
|
|
150
|
-
*
|
|
151
|
-
* Times are synthesized monotonically off a single base so the span tree is
|
|
152
|
-
* orderable; the substrate does not retain per-candidate wall-clock starts.
|
|
153
|
-
*/
|
|
154
|
-
export declare function loopProvenanceSpans(record: LoopProvenanceRecord, opts?: {
|
|
155
|
-
baseTimeMs?: number;
|
|
156
|
-
}): TraceSpanEvent[];
|
|
157
|
-
/** Canonical durable paths under the run dir. */
|
|
158
|
-
export declare function provenanceRecordPath(runDir: string): string;
|
|
159
|
-
export declare function provenanceSpansPath(runDir: string): string;
|
|
160
|
-
export interface EmitLoopProvenanceResult {
|
|
161
|
-
record: LoopProvenanceRecord;
|
|
162
|
-
spans: TraceSpanEvent[];
|
|
163
|
-
/** Absolute paths the record + spans were written to, when storage persists. */
|
|
164
|
-
recordPath: string;
|
|
165
|
-
spansPath: string;
|
|
166
|
-
}
|
|
167
|
-
export interface EmitLoopProvenanceArgs<TArtifact, TScenario extends Scenario> extends BuildLoopProvenanceArgs<TArtifact, TScenario> {
|
|
168
|
-
/** Storage the record + spans are written through. */
|
|
169
|
-
storage: CampaignStorage;
|
|
170
|
-
/** When set, the spans are also shipped to the hosted `/v1/ingest/traces`
|
|
171
|
-
* endpoint so the collector receives the full loop, not just `cost.*`. */
|
|
172
|
-
hostedClient?: HostedClient;
|
|
173
|
-
}
|
|
174
|
-
/**
|
|
175
|
-
* Build the provenance record + OTel spans and persist them durably under the
|
|
176
|
-
* run dir (and ship spans to a hosted collector when one is wired). Returns
|
|
177
|
-
* both artifacts so the caller can assert on / re-derive from them.
|
|
178
|
-
*
|
|
179
|
-
* Fail-loud: the durable write throws on storage failure (a swallowed write is
|
|
180
|
-
* exactly the "emitted but lost" failure this closes). The hosted span ship is
|
|
181
|
-
* the one best-effort leg — its failure is logged, not thrown, so an offline
|
|
182
|
-
* collector never fails the loop (the durable artifact is the source of truth).
|
|
183
|
-
*/
|
|
184
|
-
export declare function emitLoopProvenance<TArtifact, TScenario extends Scenario>(args: EmitLoopProvenanceArgs<TArtifact, TScenario>): Promise<EmitLoopProvenanceResult>;
|
|
185
|
-
//# sourceMappingURL=provenance.d.ts.map
|
|
@@ -1 +0,0 @@
|
|
|
1
|
-
{"version":3,"file":"provenance.d.ts","sourceRoot":"","sources":["../../src/campaign/provenance.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;GAyBG;AAIH,OAAO,KAAK,EAAE,YAAY,EAAE,MAAM,kBAAkB,CAAA;AACpD,OAAO,KAAK,EAIV,cAAc,EACf,MAAM,iBAAiB,CAAA;AAExB,OAAO,KAAK,EAAE,SAAS,EAAE,MAAM,eAAe,CAAA;AAC9C,OAAO,KAAK,EAAE,eAAe,EAAE,MAAM,WAAW,CAAA;AAChD,OAAO,KAAK,EAAE,cAAc,EAAE,YAAY,EAAE,UAAU,EAAE,cAAc,EAAE,QAAQ,EAAE,MAAM,SAAS,CAAA;AAEjG;;;;oCAIoC;AACpC,wBAAgB,kBAAkB,CAAC,OAAO,EAAE,cAAc,GAAG,MAAM,CAUlE;AAED,MAAM,WAAW,uBAAuB;IACtC,uDAAuD;IACvD,UAAU,EAAE,MAAM,CAAA;IAClB,qFAAqF;IACrF,WAAW,EAAE,MAAM,CAAA;IACnB,4DAA4D;IAC5D,WAAW,EAAE,MAAM,CAAA;IACnB,wEAAwE;IACxE,KAAK,CAAC,EAAE,MAAM,CAAA;IACd;kDAC8C;IAC9C,SAAS,CAAC,EAAE,MAAM,CAAA;IAClB,gEAAgE;IAChE,SAAS,EAAE,MAAM,CAAA;IACjB,iEAAiE;IACjE,QAAQ,EAAE,OAAO,CAAA;CAClB;AAED,MAAM,WAAW,qBAAqB;IACpC,sEAAsE;IACtE,OAAO,EAAE,MAAM,GAAG,OAAO,GAAG,MAAM,CAAA;IAClC,6EAA6E;IAC7E,eAAe,EAAE,MAAM,CAAA;IACvB,uDAAuD;IACvD,MAAM,EAAE,MAAM,EAAE,CAAA;IAChB,gBAAgB,EAAE,MAAM,CAAA;IACxB,iBAAiB,EAAE,MAAM,CAAA;IACzB,YAAY,EAAE,MAAM,CAAA;CACrB;AAED;;;;GAIG;AACH,MAAM,WAAW,oBAAoB;IACnC,MAAM,EAAE,2BAA2B,CAAA;IACnC,KAAK,EAAE,MAAM,CAAA;IACb,MAAM,EAAE,MAAM,CAAA;IACd,SAAS,EAAE,MAAM,CAAA;IACjB,mFAAmF;IACnF,mBAAmB,EAAE,MAAM,CAAA;IAC3B,iBAAiB,EAAE,MAAM,CAAA;IACzB,qFAAqF;IACrF,WAAW,CAAC,EAAE,MAAM,CAAA;IACpB,eAAe,CAAC,EAAE,MAAM,CAAA;IACxB,qEAAqE;IACrE,IAAI,EAAE,MAAM,CAAA;IACZ,4EAA4E;IAC5E,UAAU,EAAE,uBAAuB,EAAE,CAAA;IACrC,0EAA0E;IAC1E,IAAI,EAAE;QACJ,QAAQ,EAAE,YAAY,CAAA;QACtB,OAAO,EAAE,MAAM,EAAE,CAAA;QACjB,KAAK,CAAC,EAAE,MAAM,CAAA;QACd,iBAAiB,EAAE,KAAK,CAAC;YAAE,IAAI,EAAE,MAAM,CAAC;YAAC,MAAM,EAAE,OAAO,CAAA;SAAE,CAAC,CAAA;KAC5D,CAAA;IACD,0CAA0C;IAC1C,wBAAwB,EAAE,MAAM,CAAA;IAChC,wCAAwC;IACxC,sBAAsB,EAAE,MAAM,CAAA;IAC9B,uEAAuE;IACvE,WAAW,EAAE,MAAM,CAAA;IACnB,6EAA6E;IAC7E,OAAO,EAAE,qBAAqB,CAAA;IAC9B,YAAY,EAAE,MAAM,CAAA;IACpB,eAAe,EAAE,MAAM,CAAA;CACxB;AAED,MAAM,WAAW,uBAAuB,CAAC,SAAS,EAAE,SAAS,SAAS,QAAQ;IAC5E,KAAK,EAAE,MAAM,CAAA;IACb,MAAM,EAAE,MAAM,CAAA;IACd,SAAS,EAAE,MAAM,CAAA;IACjB,eAAe,EAAE,cAAc,CAAA;IAC/B,aAAa,EAAE,cAAc,CAAA;IAC7B,WAAW,CAAC,EAAE,MAAM,CAAA;IACpB,eAAe,CAAC,EAAE,MAAM,CAAA;IACxB,IAAI,EAAE,MAAM,CAAA;IACZ,qEAAqE;IACrE,WAAW,EAAE,KAAK,CAAC;QACjB,eAAe,EAAE,MAAM,CAAA;QACvB,UAAU,EAAE,KAAK,CAAC;YAChB,WAAW,EAAE,MAAM,CAAA;YACnB,SAAS,EAAE,MAAM,CAAA;YACjB,KAAK,CAAC,EAAE,MAAM,CAAA;YACd,SAAS,CAAC,EAAE,MAAM,CAAA;SACnB,CAAC,CAAA;QACF,QAAQ,EAAE,MAAM,EAAE,CAAA;QAClB;2EACmE;QACnE,QAAQ,EAAE,KAAK,CAAC;YAAE,WAAW,EAAE,MAAM,CAAC;YAAC,OAAO,EAAE,cAAc,CAAA;SAAE,CAAC,CAAA;KAClE,CAAC,CAAA;IACF,IAAI,EAAE,UAAU,CAAA;IAChB,iBAAiB,EAAE,cAAc,CAAC,SAAS,EAAE,SAAS,CAAC,CAAA;IACvD,eAAe,EAAE,cAAc,CAAC,SAAS,EAAE,SAAS,CAAC,CAAA;IACrD,+DAA+D;IAC/D,aAAa,EAAE,aAAa,CAAC,SAAS,CAAC,CAAA;IACvC,YAAY,EAAE,MAAM,CAAA;IACpB,eAAe,EAAE,MAAM,CAAA;CACxB;AAcD,wEAAwE;AACxE,wBAAgB,yBAAyB,CAAC,SAAS,EAAE,SAAS,SAAS,QAAQ,EAC7E,IAAI,EAAE,uBAAuB,CAAC,SAAS,EAAE,SAAS,CAAC,GAClD,oBAAoB,CA8DtB;AAgBD;;;;;;;;;;GAUG;AACH,wBAAgB,mBAAmB,CACjC,MAAM,EAAE,oBAAoB,EAC5B,IAAI,GAAE;IAAE,UAAU,CAAC,EAAE,MAAM,CAAA;CAAO,GACjC,cAAc,EAAE,CA4GlB;AAID,iDAAiD;AACjD,wBAAgB,oBAAoB,CAAC,MAAM,EAAE,MAAM,GAAG,MAAM,CAE3D;AACD,wBAAgB,mBAAmB,CAAC,MAAM,EAAE,MAAM,GAAG,MAAM,CAE1D;AAED,MAAM,WAAW,wBAAwB;IACvC,MAAM,EAAE,oBAAoB,CAAA;IAC5B,KAAK,EAAE,cAAc,EAAE,CAAA;IACvB,gFAAgF;IAChF,UAAU,EAAE,MAAM,CAAA;IAClB,SAAS,EAAE,MAAM,CAAA;CAClB;AAED,MAAM,WAAW,sBAAsB,CAAC,SAAS,EAAE,SAAS,SAAS,QAAQ,CAC3E,SAAQ,uBAAuB,CAAC,SAAS,EAAE,SAAS,CAAC;IACrD,sDAAsD;IACtD,OAAO,EAAE,eAAe,CAAA;IACxB;+EAC2E;IAC3E,YAAY,CAAC,EAAE,YAAY,CAAA;CAC5B;AAuED;;;;;;;;;GASG;AACH,wBAAsB,kBAAkB,CAAC,SAAS,EAAE,SAAS,SAAS,QAAQ,EAC5E,IAAI,EAAE,sBAAsB,CAAC,SAAS,EAAE,SAAS,CAAC,GACjD,OAAO,CAAC,wBAAwB,CAAC,CAgCnC"}
|
|
@@ -1,90 +0,0 @@
|
|
|
1
|
-
/**
|
|
2
|
-
* `runCampaign` — Pass A substrate primitive. ONE function that orchestrates
|
|
3
|
-
* scenarios → dispatch → artifacts → judges → aggregates, with full
|
|
4
|
-
* reproducibility (seed + manifest hash), cell-level resumability, bootstrap
|
|
5
|
-
* CIs, and the `LabeledScenarioStore` capture flywheel.
|
|
6
|
-
*
|
|
7
|
-
* Improvement loops (optimizer / gate / autoOnPromote) ride on top of this
|
|
8
|
-
* primitive but live in `presets/run-improvement-loop.ts`. This file keeps
|
|
9
|
-
* the core orchestrator minimal — Phase 1 of the Pass A track.
|
|
10
|
-
*/
|
|
11
|
-
import { type CampaignStorage } from './storage';
|
|
12
|
-
import type { CampaignResult, CampaignTraceWriter, DispatchFn, JudgeConfig, LabeledScenarioStore, Scenario } from './types';
|
|
13
|
-
export interface RunCampaignOptions<TScenario extends Scenario, TArtifact> {
|
|
14
|
-
scenarios: TScenario[];
|
|
15
|
-
dispatch: DispatchFn<TScenario, TArtifact>;
|
|
16
|
-
judges?: JudgeConfig<TArtifact, TScenario>[];
|
|
17
|
-
/** Required for reproducibility. Default 42. */
|
|
18
|
-
seed?: number;
|
|
19
|
-
/** Per-scenario replicates for CI bands. Default 1; raise to 5+ for
|
|
20
|
-
* bootstrap-tight intervals on critical eval. */
|
|
21
|
-
reps?: number;
|
|
22
|
-
/** When true (default), completed cells are cached by
|
|
23
|
-
* (manifestHash, scenarioId, rep, generation). Re-runs skip cached cells. */
|
|
24
|
-
resumable?: boolean;
|
|
25
|
-
/** Optional store — when present, every artifact + judge score is captured
|
|
26
|
-
* with the configured `captureSource`. Capture is default ON; pass `'off'`
|
|
27
|
-
* to disable. */
|
|
28
|
-
labeledStore?: LabeledScenarioStore | 'off';
|
|
29
|
-
captureSource?: 'production-trace' | 'eval-run' | 'manual' | 'red-team' | 'synthetic';
|
|
30
|
-
captureSourceVersionHash?: string;
|
|
31
|
-
/** Wall-clock cost cap across all cells. Cells beyond ceiling are skipped. */
|
|
32
|
-
costCeiling?: number;
|
|
33
|
-
/** Max concurrent cells. Default 2. */
|
|
34
|
-
maxConcurrency?: number;
|
|
35
|
-
/**
|
|
36
|
-
* Per-cell dispatch deadline in ms. A `dispatch` that neither resolves nor
|
|
37
|
-
* rejects within this window is a hang (a stalled model request, an
|
|
38
|
-
* exhausted runtime resource, a backend that never closes its stream). When
|
|
39
|
-
* set, the cell's `ctx.signal` is aborted and the cell is recorded as a LOUD
|
|
40
|
-
* error (`dispatch exceeded <N>ms`) so the campaign proceeds and the failure
|
|
41
|
-
* is visible — instead of one wedged cell silently hanging the whole run (and
|
|
42
|
-
* every loop/CI job above it) forever. `undefined`/`0` = unbounded (legacy).
|
|
43
|
-
*/
|
|
44
|
-
dispatchTimeoutMs?: number;
|
|
45
|
-
/** Required: where artifacts + traces land. */
|
|
46
|
-
runDir: string;
|
|
47
|
-
/** Tracing posture. Default is the substrate's `FileSystemTraceStore` rooted
|
|
48
|
-
* at `<runDir>/traces/`. `'off'` disables capture entirely — substrate
|
|
49
|
-
* refuses this when the caller wires `autoOnPromote !== 'none'`. */
|
|
50
|
-
tracing?: 'on' | 'off';
|
|
51
|
-
/**
|
|
52
|
-
* Per-cell usage expectation — the early, fine-grained sibling of the
|
|
53
|
-
* batch `assertRealBackend` guard. A cell that produced an artifact (no
|
|
54
|
-
* error) but reported `costUsd === 0` AND zero tokens is a stub: the
|
|
55
|
-
* dispatch never reported LLM activity via `ctx.cost`. Modes:
|
|
56
|
-
* - `'warn'` (default) — log the offending cell loudly, keep going.
|
|
57
|
-
* - `'assert'` — throw `BackendIntegrityError` on the first such cell
|
|
58
|
-
* (fail-fast; recommended for CI campaigns expecting real LLM calls).
|
|
59
|
-
* - `'off'` — no check (replay / deterministic-only / offline analysis).
|
|
60
|
-
*/
|
|
61
|
-
expectUsage?: 'assert' | 'warn' | 'off';
|
|
62
|
-
/** Test seam — override the wall clock for deterministic tests. */
|
|
63
|
-
now?: () => Date;
|
|
64
|
-
/** Test seam — override per-cell trace writer factory. */
|
|
65
|
-
buildTraceWriter?: (cellId: string, dir: string) => CampaignTraceWriter;
|
|
66
|
-
/** Storage backend for run/cell dirs, the resumability cache, artifacts,
|
|
67
|
-
* and trace spans. Default: the Node filesystem (`fsCampaignStorage`).
|
|
68
|
-
* Pass `inMemoryCampaignStorage()` to run in a filesystem-less runtime
|
|
69
|
-
* (Cloudflare Workers, Deno, edge) — the `CampaignResult` is still
|
|
70
|
-
* produced; artifacts/traces just aren't persisted to disk. */
|
|
71
|
-
storage?: CampaignStorage;
|
|
72
|
-
/**
|
|
73
|
-
* Optional per-cell placement strategy. Returns an opaque string the
|
|
74
|
-
* substrate forwards as `ctx.placement` to the Dispatch — placement-aware
|
|
75
|
-
* Dispatches (e.g. `httpDispatch` from `/adapters/http`) use it to route
|
|
76
|
-
* each cell to the right worker, region, or sandbox. When unset, every
|
|
77
|
-
* cell receives `ctx.placement = undefined` and behaves identically to
|
|
78
|
-
* the in-process case.
|
|
79
|
-
*
|
|
80
|
-
* @example
|
|
81
|
-
* cellPlacement: ({ scenario }) => scenario.tags?.includes('eu') ? 'eu-west' : 'us-east'
|
|
82
|
-
*/
|
|
83
|
-
cellPlacement?: (input: {
|
|
84
|
-
scenario: TScenario;
|
|
85
|
-
rep: number;
|
|
86
|
-
generation?: number;
|
|
87
|
-
}) => string | undefined;
|
|
88
|
-
}
|
|
89
|
-
export declare function runCampaign<TScenario extends Scenario, TArtifact>(opts: RunCampaignOptions<TScenario, TArtifact>): Promise<CampaignResult<TArtifact, TScenario>>;
|
|
90
|
-
//# sourceMappingURL=run-campaign.d.ts.map
|
|
@@ -1 +0,0 @@
|
|
|
1
|
-
{"version":3,"file":"run-campaign.d.ts","sourceRoot":"","sources":["../../src/campaign/run-campaign.ts"],"names":[],"mappings":"AAAA;;;;;;;;;GASG;AAMH,OAAO,EAAE,KAAK,eAAe,EAAqB,MAAM,WAAW,CAAA;AACnE,OAAO,KAAK,EAKV,cAAc,EAEd,mBAAmB,EAEnB,UAAU,EAEV,WAAW,EAEX,oBAAoB,EACpB,QAAQ,EAGT,MAAM,SAAS,CAAA;AAEhB,MAAM,WAAW,kBAAkB,CAAC,SAAS,SAAS,QAAQ,EAAE,SAAS;IACvE,SAAS,EAAE,SAAS,EAAE,CAAA;IACtB,QAAQ,EAAE,UAAU,CAAC,SAAS,EAAE,SAAS,CAAC,CAAA;IAC1C,MAAM,CAAC,EAAE,WAAW,CAAC,SAAS,EAAE,SAAS,CAAC,EAAE,CAAA;IAC5C,gDAAgD;IAChD,IAAI,CAAC,EAAE,MAAM,CAAA;IACb;sDACkD;IAClD,IAAI,CAAC,EAAE,MAAM,CAAA;IACb;kFAC8E;IAC9E,SAAS,CAAC,EAAE,OAAO,CAAA;IACnB;;sBAEkB;IAClB,YAAY,CAAC,EAAE,oBAAoB,GAAG,KAAK,CAAA;IAC3C,aAAa,CAAC,EAAE,kBAAkB,GAAG,UAAU,GAAG,QAAQ,GAAG,UAAU,GAAG,WAAW,CAAA;IACrF,wBAAwB,CAAC,EAAE,MAAM,CAAA;IACjC,8EAA8E;IAC9E,WAAW,CAAC,EAAE,MAAM,CAAA;IACpB,uCAAuC;IACvC,cAAc,CAAC,EAAE,MAAM,CAAA;IACvB;;;;;;;;OAQG;IACH,iBAAiB,CAAC,EAAE,MAAM,CAAA;IAC1B,+CAA+C;IAC/C,MAAM,EAAE,MAAM,CAAA;IACd;;yEAEqE;IACrE,OAAO,CAAC,EAAE,IAAI,GAAG,KAAK,CAAA;IACtB;;;;;;;;;OASG;IACH,WAAW,CAAC,EAAE,QAAQ,GAAG,MAAM,GAAG,KAAK,CAAA;IACvC,mEAAmE;IACnE,GAAG,CAAC,EAAE,MAAM,IAAI,CAAA;IAChB,0DAA0D;IAC1D,gBAAgB,CAAC,EAAE,CAAC,MAAM,EAAE,MAAM,EAAE,GAAG,EAAE,MAAM,KAAK,mBAAmB,CAAA;IACvE;;;;oEAIgE;IAChE,OAAO,CAAC,EAAE,eAAe,CAAA;IACzB;;;;;;;;;;OAUG;IACH,aAAa,CAAC,EAAE,CAAC,KAAK,EAAE;QACtB,QAAQ,EAAE,SAAS,CAAA;QACnB,GAAG,EAAE,MAAM,CAAA;QACX,UAAU,CAAC,EAAE,MAAM,CAAA;KACpB,KAAK,MAAM,GAAG,SAAS,CAAA;CACzB;AAED,wBAAsB,WAAW,CAAC,SAAS,SAAS,QAAQ,EAAE,SAAS,EACrE,IAAI,EAAE,kBAAkB,CAAC,SAAS,EAAE,SAAS,CAAC,GAC7C,OAAO,CAAC,cAAc,CAAC,SAAS,EAAE,SAAS,CAAC,CAAC,CAyH/C"}
|
|
@@ -1,26 +0,0 @@
|
|
|
1
|
-
/**
|
|
2
|
-
* Shared campaign-score reductions used by every optimizer preset
|
|
3
|
-
* (`runOptimization`, `runSkillOpt`, `compareProposers`). ONE definition of
|
|
4
|
-
* "composite of a campaign" and "per-scenario / per-dimension breakdown" so
|
|
5
|
-
* the optimizers cannot drift on how a surface's score is computed.
|
|
6
|
-
*/
|
|
7
|
-
import type { CampaignResult, Scenario } from './types';
|
|
8
|
-
/** Mean composite across a campaign: per cell, the mean of its judges'
|
|
9
|
-
* composites; then the mean across cells. Cells with no judge scores are
|
|
10
|
-
* skipped. Empty ⇒ 0. */
|
|
11
|
-
export declare function campaignMeanComposite<TArtifact, TScenario extends Scenario>(campaign: CampaignResult<TArtifact, TScenario>): number;
|
|
12
|
-
export interface CampaignBreakdown {
|
|
13
|
-
/** Mean score per judge dimension across all cells. */
|
|
14
|
-
dimensions: Record<string, number>;
|
|
15
|
-
/** Per-scenario composite (mean over reps + judges) + the judge's free-form
|
|
16
|
-
* `notes` for that scenario (the "why" a reflective proposer grounds on). */
|
|
17
|
-
scenarios: Array<{
|
|
18
|
-
scenarioId: string;
|
|
19
|
-
composite: number;
|
|
20
|
-
notes?: string;
|
|
21
|
-
}>;
|
|
22
|
-
}
|
|
23
|
-
/** Per-candidate evidence a reflective/patch proposer grounds its next proposal
|
|
24
|
-
* on: mean score per judge dimension + per-scenario composite. */
|
|
25
|
-
export declare function campaignBreakdown<TArtifact, TScenario extends Scenario>(campaign: CampaignResult<TArtifact, TScenario>): CampaignBreakdown;
|
|
26
|
-
//# sourceMappingURL=score-utils.d.ts.map
|
|
@@ -1 +0,0 @@
|
|
|
1
|
-
{"version":3,"file":"score-utils.d.ts","sourceRoot":"","sources":["../../src/campaign/score-utils.ts"],"names":[],"mappings":"AAAA;;;;;GAKG;AAEH,OAAO,KAAK,EAAE,cAAc,EAAE,QAAQ,EAAE,MAAM,SAAS,CAAA;AAEvD;;0BAE0B;AAC1B,wBAAgB,qBAAqB,CAAC,SAAS,EAAE,SAAS,SAAS,QAAQ,EACzE,QAAQ,EAAE,cAAc,CAAC,SAAS,EAAE,SAAS,CAAC,GAC7C,MAAM,CASR;AAED,MAAM,WAAW,iBAAiB;IAChC,uDAAuD;IACvD,UAAU,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,CAAA;IAClC;kFAC8E;IAC9E,SAAS,EAAE,KAAK,CAAC;QAAE,UAAU,EAAE,MAAM,CAAC;QAAC,SAAS,EAAE,MAAM,CAAC;QAAC,KAAK,CAAC,EAAE,MAAM,CAAA;KAAE,CAAC,CAAA;CAC5E;AAED;mEACmE;AACnE,wBAAgB,iBAAiB,CAAC,SAAS,EAAE,SAAS,SAAS,QAAQ,EACrE,QAAQ,EAAE,cAAc,CAAC,SAAS,EAAE,SAAS,CAAC,GAC7C,iBAAiB,CA4CnB"}
|
|
@@ -1,62 +0,0 @@
|
|
|
1
|
-
/**
|
|
2
|
-
* SkillOpt patch primitives (Microsoft, arXiv:2605.23904 — "Executive
|
|
3
|
-
* Strategy for Self-Evolving Agent Skills"). Where GEPA regenerates a surface
|
|
4
|
-
* by reflection, SkillOpt emits BOUNDED, anchored edits to ONE skill document
|
|
5
|
-
* — add / delete / replace — and accepts an edit only if it strictly improves
|
|
6
|
-
* a held-out score. Bounded edits are the "textual learning rate": small,
|
|
7
|
-
* reversible, and cheap to accept/reject, so a good rule introduced earlier is
|
|
8
|
-
* not overwritten by a later sweeping rewrite.
|
|
9
|
-
*
|
|
10
|
-
* This module applies a patch deterministically and reports, per op, what
|
|
11
|
-
* applied and what could not (a missing anchor is a rejected op, never a
|
|
12
|
-
* silently dropped one). Pure, no I/O.
|
|
13
|
-
*/
|
|
14
|
-
/** A single bounded edit against a skill surface.
|
|
15
|
-
* - `add` — insert `text` after the first line containing `after`
|
|
16
|
-
* (append to the end when `after` is absent/empty).
|
|
17
|
-
* - `delete` — remove the first line containing `anchor`.
|
|
18
|
-
* - `replace` — replace the first line containing `anchor` with `text`.
|
|
19
|
-
* `text` may be multi-line; it is spliced in as multiple lines. Anchors match
|
|
20
|
-
* the FIRST line that contains the substring (deterministic; SkillOpt is
|
|
21
|
-
* expected to anchor on unique text). */
|
|
22
|
-
export type SkillPatchOp = {
|
|
23
|
-
op: 'add';
|
|
24
|
-
after?: string;
|
|
25
|
-
text: string;
|
|
26
|
-
} | {
|
|
27
|
-
op: 'delete';
|
|
28
|
-
anchor: string;
|
|
29
|
-
} | {
|
|
30
|
-
op: 'replace';
|
|
31
|
-
anchor: string;
|
|
32
|
-
text: string;
|
|
33
|
-
};
|
|
34
|
-
/** A named, attributable bundle of ops the optimizer proposes as one edit. */
|
|
35
|
-
export interface SkillPatch {
|
|
36
|
-
label: string;
|
|
37
|
-
rationale: string;
|
|
38
|
-
ops: SkillPatchOp[];
|
|
39
|
-
}
|
|
40
|
-
export interface SkillPatchRejection {
|
|
41
|
-
op: SkillPatchOp;
|
|
42
|
-
reason: string;
|
|
43
|
-
}
|
|
44
|
-
export interface ApplySkillPatchResult {
|
|
45
|
-
surface: string;
|
|
46
|
-
/** Count of ops that mutated the surface. */
|
|
47
|
-
applied: number;
|
|
48
|
-
/** Ops that could not apply (unanchored / empty), with the reason. The
|
|
49
|
-
* surface still reflects every APPLIED op — partial application is honest,
|
|
50
|
-
* and the caller decides whether a partial patch is worth scoring. */
|
|
51
|
-
rejected: SkillPatchRejection[];
|
|
52
|
-
}
|
|
53
|
-
/**
|
|
54
|
-
* Apply a SkillOpt patch to a text surface. Ops apply in array order against
|
|
55
|
-
* the evolving line buffer (an `add after X` followed by a `delete X` sees the
|
|
56
|
-
* inserted lines). A missing anchor rejects only that op; the rest still apply.
|
|
57
|
-
*/
|
|
58
|
-
export declare function applySkillPatch(surface: string, patch: SkillPatch): ApplySkillPatchResult;
|
|
59
|
-
/** Total ops in a patch — the edit-budget axis (SkillOpt's "textual learning
|
|
60
|
-
* rate" caps this per epoch). */
|
|
61
|
-
export declare function patchEditCount(patch: SkillPatch): number;
|
|
62
|
-
//# sourceMappingURL=skill-patch.d.ts.map
|
|
@@ -1 +0,0 @@
|
|
|
1
|
-
{"version":3,"file":"skill-patch.d.ts","sourceRoot":"","sources":["../../src/campaign/skill-patch.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;GAYG;AAEH;;;;;;;0CAO0C;AAC1C,MAAM,MAAM,YAAY,GACpB;IAAE,EAAE,EAAE,KAAK,CAAC;IAAC,KAAK,CAAC,EAAE,MAAM,CAAC;IAAC,IAAI,EAAE,MAAM,CAAA;CAAE,GAC3C;IAAE,EAAE,EAAE,QAAQ,CAAC;IAAC,MAAM,EAAE,MAAM,CAAA;CAAE,GAChC;IAAE,EAAE,EAAE,SAAS,CAAC;IAAC,MAAM,EAAE,MAAM,CAAC;IAAC,IAAI,EAAE,MAAM,CAAA;CAAE,CAAA;AAEnD,8EAA8E;AAC9E,MAAM,WAAW,UAAU;IACzB,KAAK,EAAE,MAAM,CAAA;IACb,SAAS,EAAE,MAAM,CAAA;IACjB,GAAG,EAAE,YAAY,EAAE,CAAA;CACpB;AAED,MAAM,WAAW,mBAAmB;IAClC,EAAE,EAAE,YAAY,CAAA;IAChB,MAAM,EAAE,MAAM,CAAA;CACf;AAED,MAAM,WAAW,qBAAqB;IACpC,OAAO,EAAE,MAAM,CAAA;IACf,6CAA6C;IAC7C,OAAO,EAAE,MAAM,CAAA;IACf;;2EAEuE;IACvE,QAAQ,EAAE,mBAAmB,EAAE,CAAA;CAChC;AAED;;;;GAIG;AACH,wBAAgB,eAAe,CAAC,OAAO,EAAE,MAAM,EAAE,KAAK,EAAE,UAAU,GAAG,qBAAqB,CAkDzF;AAED;kCACkC;AAClC,wBAAgB,cAAc,CAAC,KAAK,EAAE,UAAU,GAAG,MAAM,CAExD"}
|
|
@@ -1,38 +0,0 @@
|
|
|
1
|
-
/**
|
|
2
|
-
* `CampaignStorage` — the filesystem seam `runCampaign` writes through
|
|
3
|
-
* (run/cell dirs, the resumability cache, per-cell artifacts, trace spans).
|
|
4
|
-
*
|
|
5
|
-
* The default (`fsCampaignStorage`) is the Node filesystem — identical
|
|
6
|
-
* behavior to the inline `node:fs` calls it replaces, so existing CLI
|
|
7
|
-
* consumers are unaffected. `inMemoryCampaignStorage` keeps everything in a
|
|
8
|
-
* `Map`, so the substrate runs in environments WITHOUT a filesystem
|
|
9
|
-
* (Cloudflare Workers, Deno Deploy, other edge runtimes) — the campaign
|
|
10
|
-
* still produces its `CampaignResult` (cells + aggregates) in memory;
|
|
11
|
-
* artifacts/traces simply aren't persisted to disk.
|
|
12
|
-
*
|
|
13
|
-
* Paths are opaque keys to the in-memory adapter — it does not parse them,
|
|
14
|
-
* so the same `join(...)`-built paths work unchanged across both adapters.
|
|
15
|
-
*/
|
|
16
|
-
export interface CampaignStorage {
|
|
17
|
-
/** Ensure a directory exists (recursive). No-op for in-memory. */
|
|
18
|
-
ensureDir(dir: string): void;
|
|
19
|
-
/** Does this path exist (as a written file or an ensured dir)? */
|
|
20
|
-
exists(path: string): boolean;
|
|
21
|
-
/** Read a UTF-8 file; `undefined` when missing or unreadable. */
|
|
22
|
-
read(path: string): string | undefined;
|
|
23
|
-
/** Write a file (string or bytes). Parent dir is assumed ensured. */
|
|
24
|
-
write(path: string, content: string | Uint8Array): void;
|
|
25
|
-
}
|
|
26
|
-
/** Node-filesystem storage — the default. Lazily requires `node:fs` so the
|
|
27
|
-
* module imports cleanly in non-Node runtimes (where the caller passes
|
|
28
|
-
* `inMemoryCampaignStorage` instead and never constructs this).
|
|
29
|
-
*
|
|
30
|
-
* `createRequire(import.meta.url)` is the ESM-native lazy require — a bare
|
|
31
|
-
* `require` is a ReferenceError under `"type": "module"`, which is exactly
|
|
32
|
-
* the shape this package publishes. */
|
|
33
|
-
export declare function fsCampaignStorage(): CampaignStorage;
|
|
34
|
-
/** In-memory storage for filesystem-less runtimes. Artifacts + trace spans
|
|
35
|
-
* live in a `Map` for the duration of the run; the `CampaignResult` is
|
|
36
|
-
* fully populated, but nothing is persisted to disk. */
|
|
37
|
-
export declare function inMemoryCampaignStorage(): CampaignStorage;
|
|
38
|
-
//# sourceMappingURL=storage.d.ts.map
|
|
@@ -1 +0,0 @@
|
|
|
1
|
-
{"version":3,"file":"storage.d.ts","sourceRoot":"","sources":["../../src/campaign/storage.ts"],"names":[],"mappings":"AAEA;;;;;;;;;;;;;;GAcG;AACH,MAAM,WAAW,eAAe;IAC9B,kEAAkE;IAClE,SAAS,CAAC,GAAG,EAAE,MAAM,GAAG,IAAI,CAAA;IAC5B,kEAAkE;IAClE,MAAM,CAAC,IAAI,EAAE,MAAM,GAAG,OAAO,CAAA;IAC7B,iEAAiE;IACjE,IAAI,CAAC,IAAI,EAAE,MAAM,GAAG,MAAM,GAAG,SAAS,CAAA;IACtC,qEAAqE;IACrE,KAAK,CAAC,IAAI,EAAE,MAAM,EAAE,OAAO,EAAE,MAAM,GAAG,UAAU,GAAG,IAAI,CAAA;CACxD;AAED;;;;;;wCAMwC;AACxC,wBAAgB,iBAAiB,IAAI,eAAe,CAuBnD;AAED;;yDAEyD;AACzD,wBAAgB,uBAAuB,IAAI,eAAe,CAmBzD"}
|
|
@@ -1 +0,0 @@
|
|
|
1
|
-
{"version":3,"file":"types.d.ts","sourceRoot":"","sources":["../../src/campaign/types.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;GAgBG;AAEH,OAAO,KAAK,EAAE,aAAa,EAAE,MAAM,eAAe,CAAA;AAElD;8EAC8E;AAC9E,MAAM,WAAW,QAAQ;IACvB,EAAE,EAAE,MAAM,CAAA;IACV,IAAI,EAAE,MAAM,CAAA;IACZ,IAAI,CAAC,EAAE,MAAM,EAAE,CAAA;CAChB;AAED;;0DAE0D;AAC1D,MAAM,WAAW,eAAe;IAC9B,MAAM,EAAE,MAAM,CAAA;IACd,GAAG,EAAE,MAAM,CAAA;IACX,UAAU,CAAC,EAAE,MAAM,CAAA;IACnB,IAAI,EAAE,MAAM,CAAA;IACZ,MAAM,EAAE,WAAW,CAAA;IACnB,KAAK,EAAE,mBAAmB,CAAA;IAC1B,SAAS,EAAE,sBAAsB,CAAA;IACjC,IAAI,EAAE,iBAAiB,CAAA;IACvB,yEAAyE;IACzE,OAAO,CAAC,EAAE,MAAM,CAAA;IAChB,mEAAmE;IACnE,WAAW,CAAC,EAAE,MAAM,CAAA;IACpB;;;;;;OAMG;IACH,SAAS,CAAC,EAAE,MAAM,CAAA;CACnB;AAED;+EAC+E;AAC/E,MAAM,MAAM,UAAU,CAAC,SAAS,SAAS,QAAQ,EAAE,SAAS,IAAI,CAC9D,QAAQ,EAAE,SAAS,EACnB,GAAG,EAAE,eAAe,KACjB,OAAO,CAAC,SAAS,CAAC,CAAA;AAIvB;;iDAEiD;AACjD,MAAM,WAAW,aAAa,CAAC,SAAS,EAAE,SAAS;IACjD,EAAE,EAAE,MAAM,CAAA;IACV,MAAM,EAAE,MAAM,CAAA;IACd,QAAQ,CAAC,EAAE,MAAM,CAAA;IACjB,sEAAsE;IACtE,gBAAgB,CAAC,EAAE,OAAO,CAAA;IAC1B;wEACoE;IACpE,kBAAkB,CAAC,EAAE,CAAC,QAAQ,EAAE,SAAS,EAAE,YAAY,EAAE,MAAM,EAAE,QAAQ,EAAE,SAAS,KAAK,SAAS,CAAA;CACnG;AAID,MAAM,WAAW,cAAc;IAC7B,mCAAmC;IACnC,GAAG,EAAE,MAAM,CAAA;IACX,oDAAoD;IACpD,WAAW,EAAE,MAAM,CAAA;CACpB;AAED;;;;;;kFAMkF;AAClF,MAAM,WAAW,WAAW,CAAC,SAAS,EAAE,SAAS,SAAS,QAAQ,GAAG,QAAQ;IAC3E,IAAI,EAAE,MAAM,CAAA;IACZ,UAAU,EAAE,cAAc,EAAE,CAAA;IAC5B;0DACsD;IACtD,KAAK,CAAC,KAAK,EAAE;QACX,QAAQ,EAAE,SAAS,CAAA;QACnB,QAAQ,EAAE,SAAS,CAAA;QACnB,MAAM,EAAE,WAAW,CAAA;KACpB,GAAG,UAAU,GAAG,OAAO,CAAC,UAAU,CAAC,CAAA;IACpC,SAAS,CAAC,EAAE,CAAC,QAAQ,EAAE,SAAS,KAAK,OAAO,CAAA;CAC7C;AAED;;;;;;;;wDAQwD;AACxD,MAAM,WAAW,UAAU;IACzB,UAAU,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,CAAA;IAClC,SAAS,EAAE,MAAM,CAAA;IACjB,KAAK,EAAE,MAAM,CAAA;IACb;;uEAEmE;IACnE,MAAM,CAAC,EAAE,IAAI,CAAA;IACb;mEAC+D;IAC/D,eAAe,CAAC,EAAE,MAAM,CAAA;IACxB,8DAA8D;IAC9D,YAAY,CAAC,EAAE,MAAM,EAAE,CAAA;IACvB,oEAAoE;IACpE,QAAQ,CAAC,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC,CAAA;CAClD;AAID;;;;+CAI+C;AAC/C,MAAM,WAAW,WAAW;IAC1B,IAAI,EAAE,MAAM,CAAA;IACZ;2EACuE;IACvE,WAAW,EAAE,MAAM,CAAA;IACnB,yEAAyE;IACzE,OAAO,CAAC,EAAE,MAAM,CAAA;IAChB,sEAAsE;IACtE,OAAO,CAAC,EAAE,MAAM,CAAA;CACjB;AAED;;;;;;qBAMqB;AACrB,MAAM,MAAM,cAAc,GAAG,MAAM,GAAG,WAAW,CAAA;AAEjD;;;;;iFAKiF;AACjF,MAAM,WAAW,iBAAiB;IAChC,OAAO,EAAE,cAAc,CAAA;IACvB,6DAA6D;IAC7D,KAAK,EAAE,MAAM,CAAA;IACb;;qCAEiC;IACjC,SAAS,EAAE,MAAM,CAAA;CAClB;AAED;4EAC4E;AAC5E,wBAAgB,mBAAmB,CACjC,KAAK,EAAE,cAAc,GAAG,iBAAiB,GACxC,KAAK,IAAI,iBAAiB,CAQ5B;AAED;;;;;;;kEAOkE;AAClE,MAAM,WAAW,YAAY;IAC3B,OAAO,EAAE,cAAc,CAAA;IACvB,WAAW,EAAE,MAAM,CAAA;IACnB;8CAC0C;IAC1C,UAAU,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,CAAA;IAClC;qDACiD;IACjD,SAAS,EAAE,MAAM,CAAA;IACjB,qEAAqE;IACrE,UAAU,EAAE,MAAM,CAAA;IAClB,KAAK,CAAC,EAAE,MAAM,CAAA;IACd,SAAS,CAAC,EAAE,MAAM,CAAA;CACnB;AAED;;;4DAG4D;AAC5D,MAAM,WAAW,OAAO,CAAC,SAAS,GAAG,OAAO;IAC1C,IAAI,EAAE,MAAM,CAAA;IACZ,MAAM,CAAC,IAAI,EAAE;QACX,QAAQ,EAAE,SAAS,EAAE,CAAA;QACrB,cAAc,EAAE,cAAc,CAAA;QAC9B,cAAc,EAAE,MAAM,CAAA;QACtB,MAAM,EAAE,WAAW,CAAA;KACpB,GAAG,OAAO,CAAC,KAAK,CAAC,cAAc,GAAG,iBAAiB,CAAC,CAAC,CAAA;CACvD;AAED;;;;;wCAKwC;AACxC,MAAM,WAAW,cAAc,CAAC,SAAS,GAAG,OAAO;IACjD,cAAc,EAAE,cAAc,CAAA;IAC9B,OAAO,EAAE,gBAAgB,EAAE,CAAA;IAC3B,QAAQ,EAAE,SAAS,EAAE,CAAA;IACrB,sEAAsE;IACtE,cAAc,EAAE,MAAM,CAAA;IACtB,UAAU,EAAE,MAAM,CAAA;IAClB,MAAM,EAAE,WAAW,CAAA;IACnB;wDACoD;IACpD,MAAM,CAAC,EAAE,OAAO,CAAA;IAChB;gDAC4C;IAC5C,OAAO,CAAC,EAAE,oBAAoB,CAAA;IAC9B;;+BAE2B;IAC3B,mBAAmB,CAAC,EAAE,MAAM,CAAA;IAC5B;;;;;8DAK0D;IAC1D,aAAa,CAAC,EAAE,YAAY,EAAE,CAAA;IAC9B;;;;;;gFAM4E;IAC5E,WAAW,CAAC,EAAE,KAAK,CAAA;CACpB;AAED;;;;;;;;;GASG;AACH,MAAM,WAAW,eAAe,CAAC,SAAS,GAAG,OAAO;IAClD,IAAI,EAAE,MAAM,CAAA;IACZ;;;2EAGuE;IACvE,OAAO,CAAC,GAAG,EAAE,cAAc,CAAC,SAAS,CAAC,GAAG,OAAO,CAAC,KAAK,CAAC,cAAc,GAAG,iBAAiB,CAAC,CAAC,CAAA;IAC3F;kEAC8D;IAC9D,MAAM,CAAC,CAAC,IAAI,EAAE;QAAE,OAAO,EAAE,gBAAgB,EAAE,CAAA;KAAE,GAAG;QAAE,IAAI,EAAE,OAAO,CAAC;QAAC,MAAM,CAAC,EAAE,MAAM,CAAA;KAAE,CAAA;CACnF;AAED;gCACgC;AAChC,MAAM,MAAM,oBAAoB,CAAC,SAAS,GAAG,OAAO,IAAI,eAAe,CAAC,SAAS,CAAC,CAAA;AAElF,MAAM,WAAW,mBAAmB;IAClC,cAAc,EAAE,MAAM,CAAA;IACtB,cAAc,EAAE,MAAM,CAAA;IACtB,gBAAgB,EAAE,CAAC,OAAO,EAAE,OAAO,KAAK,cAAc,CAAA;CACvD;AAED,MAAM,WAAW,eAAgB,SAAQ,mBAAmB;IAC1D,QAAQ,EAAE,eAAe,CAAA;CAC1B;AAID,2DAA2D;AAC3D,MAAM,MAAM,YAAY,GAAG,MAAM,GAAG,MAAM,GAAG,gBAAgB,GAAG,eAAe,GAAG,cAAc,CAAA;AAEhG,MAAM,WAAW,WAAW,CAAC,SAAS,EAAE,SAAS,SAAS,QAAQ;IAChE,kBAAkB,EAAE,GAAG,CAAC,MAAM,EAAE,SAAS,CAAC,CAAA;IAC1C,iBAAiB,CAAC,EAAE,GAAG,CAAC,MAAM,EAAE,SAAS,CAAC,CAAA;IAC1C,wDAAwD;IACxD,WAAW,EAAE,GAAG,CAAC,MAAM,EAAE,MAAM,CAAC,MAAM,EAAE,UAAU,CAAC,CAAC,CAAA;IACpD;;;+DAG2D;IAC3D,mBAAmB,CAAC,EAAE,GAAG,CAAC,MAAM,EAAE,MAAM,CAAC,MAAM,EAAE,UAAU,CAAC,CAAC,CAAA;IAC7D,SAAS,EAAE,SAAS,EAAE,CAAA;IACtB,IAAI,EAAE;QAAE,SAAS,EAAE,MAAM,CAAC;QAAC,QAAQ,EAAE,MAAM,CAAA;KAAE,CAAA;IAC7C,MAAM,EAAE,WAAW,CAAA;CACpB;AAED,MAAM,WAAW,UAAU;IACzB,QAAQ,EAAE,YAAY,CAAA;IACtB,OAAO,EAAE,MAAM,EAAE,CAAA;IACjB,iBAAiB,EAAE,KAAK,CAAC;QAAE,IAAI,EAAE,MAAM,CAAC;QAAC,MAAM,EAAE,OAAO,CAAC;QAAC,MAAM,EAAE,OAAO,CAAA;KAAE,CAAC,CAAA;IAC5E,KAAK,CAAC,EAAE,MAAM,CAAA;CACf;AAED,iCAAiC;AACjC,MAAM,WAAW,IAAI,CAAC,SAAS,GAAG,OAAO,EAAE,SAAS,SAAS,QAAQ,GAAG,QAAQ;IAC9E,IAAI,EAAE,MAAM,CAAA;IACZ,MAAM,CAAC,GAAG,EAAE,WAAW,CAAC,SAAS,EAAE,SAAS,CAAC,GAAG,OAAO,CAAC,UAAU,CAAC,CAAA;CACpE;AAID;4DAC4D;AAC5D,MAAM,WAAW,mBAAmB;IAClC,IAAI,CAAC,IAAI,EAAE,MAAM,EAAE,UAAU,CAAC,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,GAAG,SAAS,CAAA;IACnE,KAAK,IAAI,OAAO,CAAC,IAAI,CAAC,CAAA;CACvB;AAED,MAAM,WAAW,SAAS;IACxB,GAAG,CAAC,UAAU,CAAC,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,GAAG,IAAI,CAAA;IAC/C,YAAY,CAAC,GAAG,EAAE,MAAM,EAAE,KAAK,EAAE,OAAO,GAAG,IAAI,CAAA;CAChD;AAED;kCACkC;AAClC,MAAM,WAAW,sBAAsB;IACrC,KAAK,CAAC,IAAI,EAAE,MAAM,EAAE,OAAO,EAAE,MAAM,GAAG,UAAU,GAAG,OAAO,CAAC,MAAM,CAAC,CAAA;IAClE,SAAS,CAAC,IAAI,EAAE,MAAM,EAAE,KAAK,EAAE,OAAO,GAAG,OAAO,CAAC,MAAM,CAAC,CAAA;CACzD;AAED;;;mEAGmE;AACnE,MAAM,MAAM,kBAAkB,GAAG,aAAa,CAAA;AAE9C;;;;;;oDAMoD;AACpD,MAAM,WAAW,iBAAiB;IAChC,OAAO,CAAC,SAAS,EAAE,MAAM,EAAE,MAAM,EAAE,MAAM,GAAG,IAAI,CAAA;IAChD;;;;4DAIwD;IACxD,aAAa,CAAC,KAAK,EAAE,kBAAkB,GAAG,IAAI,CAAA;IAC9C,OAAO,IAAI,MAAM,CAAA;IACjB,uEAAuE;IACvE,MAAM,IAAI,kBAAkB,CAAA;CAC7B;AAID;;sDAEsD;AACtD,MAAM,MAAM,qBAAqB,GAC7B,kBAAkB,GAClB,UAAU,GACV,QAAQ,GACR,UAAU,GACV,WAAW,CAAA;AAEf,MAAM,MAAM,eAAe,GAAG,KAAK,GAAG,cAAc,GAAG,kBAAkB,GAAG,gBAAgB,CAAA;AAE5F;;;;;;;;;;;;;oBAaoB;AACpB,MAAM,MAAM,UAAU,GAAG,YAAY,GAAG,iBAAiB,GAAG,aAAa,CAAA;AAQzE,2EAA2E;AAC3E,wBAAgB,cAAc,CAAC,KAAK,EAAE,UAAU,GAAG,SAAS,GAAG,MAAM,CAEpE;AAED;;6DAE6D;AAC7D,MAAM,WAAW,oBAAoB,CAAC,SAAS,SAAS,QAAQ,GAAG,QAAQ,EAAE,SAAS,GAAG,OAAO;IAC9F,QAAQ,EAAE,SAAS,CAAA;IACnB,QAAQ,EAAE,SAAS,CAAA;IACnB,WAAW,EAAE,MAAM,CAAC,MAAM,EAAE,UAAU,CAAC,CAAA;IACvC,MAAM,EAAE,qBAAqB,CAAA;IAC7B,iBAAiB,EAAE,MAAM,CAAA;IACzB,UAAU,EAAE,MAAM,CAAA;IAClB,eAAe,EAAE,eAAe,CAAA;IAChC;;;0CAGsC;IACtC,UAAU,CAAC,EAAE,UAAU,CAAA;IACvB,uEAAuE;IACvE,eAAe,CAAC,EAAE,MAAM,CAAA;CACzB;AAED,MAAM,WAAW,qBAAqB,CAAC,SAAS,SAAS,QAAQ,GAAG,QAAQ,EAAE,SAAS,GAAG,OAAO,CAC/F,SAAQ,oBAAoB,CAAC,SAAS,EAAE,SAAS,CAAC;IAClD,2EAA2E;IAC3E,UAAU,EAAE,MAAM,CAAA;IAClB;iFAC6E;IAC7E,KAAK,EAAE,OAAO,GAAG,MAAM,CAAA;CACxB;AAED,MAAM,WAAW,yBAAyB;IACxC,KAAK,EAAE,MAAM,CAAA;IACb,wEAAwE;IACxE,KAAK,EAAE,OAAO,GAAG,MAAM,CAAA;IACvB;;2CAEuC;IACvC,cAAc,EAAE,MAAM,CAAA;IACtB,MAAM,CAAC,EAAE;QACP,IAAI,CAAC,EAAE,MAAM,CAAA;QACb,MAAM,CAAC,EAAE,qBAAqB,GAAG,qBAAqB,EAAE,CAAA;QACxD,YAAY,CAAC,EAAE,MAAM,CAAA;QACrB,YAAY,CAAC,EAAE,MAAM,CAAA;QACrB;;;mCAG2B;QAC3B,QAAQ,CAAC,EAAE,UAAU,CAAA;KACtB,CAAA;CACF;AAED,MAAM,WAAW,oBAAoB;IACnC,OAAO,CAAC,KAAK,EAAE,oBAAoB,GAAG,OAAO,CAAC,IAAI,CAAC,CAAA;IACnD,MAAM,CAAC,IAAI,EAAE,yBAAyB,GAAG,OAAO,CAAC,qBAAqB,EAAE,CAAC,CAAA;IACzE,IAAI,IAAI,OAAO,CAAC;QACd,KAAK,EAAE,MAAM,CAAA;QACb,IAAI,EAAE,MAAM,CAAA;QACZ,QAAQ,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,CAAA;QAChC;0CACkC;QAClC,OAAO,EAAE,MAAM,CAAC,UAAU,EAAE,MAAM,CAAC,CAAA;KACpC,CAAC,CAAA;CACH;AAID,MAAM,WAAW,kBAAkB,CAAC,SAAS;IAC3C,MAAM,EAAE,MAAM,CAAA;IACd,UAAU,EAAE,MAAM,CAAA;IAClB,GAAG,EAAE,MAAM,CAAA;IACX,UAAU,CAAC,EAAE,MAAM,CAAA;IACnB,QAAQ,EAAE,SAAS,CAAA;IACnB,WAAW,EAAE,MAAM,CAAC,MAAM,EAAE,UAAU,CAAC,CAAA;IACvC,OAAO,EAAE,MAAM,CAAA;IACf;;mDAE+C;IAC/C,UAAU,EAAE,kBAAkB,CAAA;IAC9B,UAAU,EAAE,MAAM,CAAA;IAClB,IAAI,EAAE,MAAM,CAAA;IACZ,MAAM,EAAE,OAAO,CAAA;IACf,KAAK,CAAC,EAAE,MAAM,CAAA;CACf;AAED,MAAM,WAAW,cAAc;IAC7B,IAAI,EAAE,MAAM,CAAA;IACZ,KAAK,EAAE,MAAM,CAAA;IACb,IAAI,EAAE,CAAC,MAAM,EAAE,MAAM,CAAC,CAAA;IACtB,CAAC,EAAE,MAAM,CAAA;CACV;AAED,MAAM,WAAW,iBAAiB;IAChC,aAAa,EAAE,MAAM,CAAA;IACrB,IAAI,EAAE,CAAC,MAAM,EAAE,MAAM,CAAC,CAAA;IACtB,CAAC,EAAE,MAAM,CAAA;CACV;AAED,MAAM,WAAW,gBAAgB;IAC/B,eAAe,EAAE,MAAM,CAAA;IACvB,UAAU,EAAE,mBAAmB,EAAE,CAAA;IACjC,QAAQ,EAAE,MAAM,EAAE,CAAA;CACnB;AAED;;;2DAG2D;AAC3D,MAAM,WAAW,mBAAmB;IAClC,WAAW,EAAE,MAAM,CAAA;IACnB,SAAS,EAAE,MAAM,CAAA;IACjB,IAAI,EAAE,CAAC,MAAM,EAAE,MAAM,CAAC,CAAA;IACtB;+CAC2C;IAC3C,UAAU,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,CAAA;IAClC;;;;;gDAK4C;IAC5C,SAAS,EAAE,KAAK,CAAC;QAAE,UAAU,EAAE,MAAM,CAAC;QAAC,SAAS,EAAE,MAAM,CAAC;QAAC,KAAK,CAAC,EAAE,MAAM,CAAA;KAAE,CAAC,CAAA;IAC3E;4EACwE;IACxE,KAAK,CAAC,EAAE,MAAM,CAAA;IACd;;oEAEgE;IAChE,SAAS,CAAC,EAAE,MAAM,CAAA;CACnB;AAED,MAAM,WAAW,kBAAkB;IACjC,OAAO,EAAE,MAAM,CAAC,MAAM,EAAE,cAAc,CAAC,CAAA;IACvC,UAAU,EAAE,MAAM,CAAC,MAAM,EAAE,iBAAiB,CAAC,CAAA;IAC7C,YAAY,EAAE,MAAM,CAAA;IACpB,aAAa,EAAE,MAAM,CAAA;IACrB,YAAY,EAAE,MAAM,CAAA;IACpB,WAAW,EAAE,MAAM,CAAA;IACnB,WAAW,EAAE,MAAM,CAAA;CACpB;AAED,MAAM,WAAW,cAAc,CAAC,SAAS,GAAG,OAAO,EAAE,SAAS,SAAS,QAAQ,GAAG,QAAQ;IACxF,0GAA0G;IAC1G,YAAY,EAAE,MAAM,CAAA;IACpB,IAAI,EAAE,MAAM,CAAA;IACZ,SAAS,EAAE,MAAM,CAAA;IACjB,OAAO,EAAE,MAAM,CAAA;IACf,UAAU,EAAE,MAAM,CAAA;IAClB,KAAK,EAAE,KAAK,CAAC,kBAAkB,CAAC,SAAS,CAAC,CAAC,CAAA;IAC3C,UAAU,EAAE,kBAAkB,CAAA;IAC9B,YAAY,CAAC,EAAE;QACb,WAAW,EAAE,gBAAgB,EAAE,CAAA;QAC/B,iBAAiB,CAAC,EAAE,MAAM,CAAA;KAC3B,CAAA;IACD,IAAI,CAAC,EAAE,UAAU,CAAA;IACjB,KAAK,CAAC,EAAE,MAAM,CAAA;IACd,MAAM,EAAE,MAAM,CAAA;IACd,eAAe,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,CAAA;IACvC;;;+DAG2D;IAC3D,SAAS,EAAE,KAAK,CAAC,IAAI,CAAC,SAAS,EAAE,IAAI,GAAG,MAAM,CAAC,CAAC,CAAA;CACjD"}
|
|
@@ -1,53 +0,0 @@
|
|
|
1
|
-
/**
|
|
2
|
-
* VCS-pluggable worktree adapter. One improvement = one worktree, PR-like
|
|
3
|
-
* (multiple commits allowed). A code-tier proposer's `propose()` creates a
|
|
4
|
-
* worktree, an agent commits the change into it, and `finalize()` returns a
|
|
5
|
-
* `CodeSurface{ worktreeRef }` the measurement checks out to run the worker
|
|
6
|
-
* against the changed code. On promotion the worktree becomes the PR branch.
|
|
7
|
-
*
|
|
8
|
-
* The interface is VCS-agnostic so a future `jj` ([jj-vcs](https://github.com/jj-vcs/jj))
|
|
9
|
-
* adapter can slot in without touching proposer code. Only the git adapter
|
|
10
|
-
* ships today. See `docs/design/loop-taxonomy.md`.
|
|
11
|
-
*/
|
|
12
|
-
import type { CodeSurface } from '../types';
|
|
13
|
-
export interface Worktree {
|
|
14
|
-
/** Absolute path to the checked-out worktree directory. */
|
|
15
|
-
path: string;
|
|
16
|
-
/** The branch the worktree is on (becomes the PR branch on promotion). */
|
|
17
|
-
branch: string;
|
|
18
|
-
/** The ref the worktree was forked from. */
|
|
19
|
-
baseRef: string;
|
|
20
|
-
}
|
|
21
|
-
export interface WorktreeAdapter {
|
|
22
|
-
/** Create an isolated worktree on a fresh branch off `baseRef`. */
|
|
23
|
-
create(opts: {
|
|
24
|
-
baseRef: string;
|
|
25
|
-
label: string;
|
|
26
|
-
}): Promise<Worktree>;
|
|
27
|
-
/** Commit any pending changes in the worktree, then return a CodeSurface
|
|
28
|
-
* pointing at it. The agent has already written its change into
|
|
29
|
-
* `worktree.path` by the time this is called. */
|
|
30
|
-
finalize(worktree: Worktree, summary: string): Promise<CodeSurface>;
|
|
31
|
-
/** Remove the worktree (and its branch) — called for losing candidates. */
|
|
32
|
-
discard(worktree: Worktree): Promise<void>;
|
|
33
|
-
}
|
|
34
|
-
export declare class WorktreeAdapterError extends Error {
|
|
35
|
-
readonly cause?: unknown | undefined;
|
|
36
|
-
constructor(message: string, cause?: unknown | undefined);
|
|
37
|
-
}
|
|
38
|
-
export interface GitWorktreeAdapterOptions {
|
|
39
|
-
/** Repo root the worktrees fork from. */
|
|
40
|
-
repoRoot: string;
|
|
41
|
-
/** Directory worktrees are created under. Default: `<repoRoot>/.worktrees`. */
|
|
42
|
-
worktreeDir?: string;
|
|
43
|
-
/** Branch-name prefix. Default: `improve`. */
|
|
44
|
-
branchPrefix?: string;
|
|
45
|
-
/** Test seam — defaults to a real `git` runner. */
|
|
46
|
-
git?: (args: string[], cwd: string) => string;
|
|
47
|
-
}
|
|
48
|
-
export declare function gitWorktreeAdapter(opts: GitWorktreeAdapterOptions): WorktreeAdapter;
|
|
49
|
-
/** Resolve a `CodeSurface`'s worktreeRef to a directory the measurement can
|
|
50
|
-
* run the worker in. A path ref is returned as-is; anything else is treated
|
|
51
|
-
* as a ref under the adapter's worktree dir. */
|
|
52
|
-
export declare function resolveWorktreePath(surface: CodeSurface, worktreeDir?: string): string;
|
|
53
|
-
//# sourceMappingURL=index.d.ts.map
|
|
@@ -1 +0,0 @@
|
|
|
1
|
-
{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../../../src/campaign/worktree/index.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;GAUG;AAKH,OAAO,KAAK,EAAE,WAAW,EAAE,MAAM,UAAU,CAAA;AAE3C,MAAM,WAAW,QAAQ;IACvB,2DAA2D;IAC3D,IAAI,EAAE,MAAM,CAAA;IACZ,0EAA0E;IAC1E,MAAM,EAAE,MAAM,CAAA;IACd,4CAA4C;IAC5C,OAAO,EAAE,MAAM,CAAA;CAChB;AAED,MAAM,WAAW,eAAe;IAC9B,mEAAmE;IACnE,MAAM,CAAC,IAAI,EAAE;QAAE,OAAO,EAAE,MAAM,CAAC;QAAC,KAAK,EAAE,MAAM,CAAA;KAAE,GAAG,OAAO,CAAC,QAAQ,CAAC,CAAA;IACnE;;sDAEkD;IAClD,QAAQ,CAAC,QAAQ,EAAE,QAAQ,EAAE,OAAO,EAAE,MAAM,GAAG,OAAO,CAAC,WAAW,CAAC,CAAA;IACnE,2EAA2E;IAC3E,OAAO,CAAC,QAAQ,EAAE,QAAQ,GAAG,OAAO,CAAC,IAAI,CAAC,CAAA;CAC3C;AAED,qBAAa,oBAAqB,SAAQ,KAAK;IAG3C,QAAQ,CAAC,KAAK,CAAC,EAAE,OAAO;gBADxB,OAAO,EAAE,MAAM,EACN,KAAK,CAAC,EAAE,OAAO,YAAA;CAK3B;AAED,MAAM,WAAW,yBAAyB;IACxC,yCAAyC;IACzC,QAAQ,EAAE,MAAM,CAAA;IAChB,+EAA+E;IAC/E,WAAW,CAAC,EAAE,MAAM,CAAA;IACpB,8CAA8C;IAC9C,YAAY,CAAC,EAAE,MAAM,CAAA;IACrB,mDAAmD;IACnD,GAAG,CAAC,EAAE,CAAC,IAAI,EAAE,MAAM,EAAE,EAAE,GAAG,EAAE,MAAM,KAAK,MAAM,CAAA;CAC9C;AAyBD,wBAAgB,kBAAkB,CAAC,IAAI,EAAE,yBAAyB,GAAG,eAAe,CAqCnF;AAED;;iDAEiD;AACjD,wBAAgB,mBAAmB,CAAC,OAAO,EAAE,WAAW,EAAE,WAAW,CAAC,EAAE,MAAM,GAAG,MAAM,CAItF"}
|
package/dist/canary.d.ts
DELETED
|
@@ -1,101 +0,0 @@
|
|
|
1
|
-
/**
|
|
2
|
-
* Liveness canaries — cheap statistical checks that catch the failure
|
|
3
|
-
* modes a green test suite never sees.
|
|
4
|
-
*
|
|
5
|
-
* Three canary types in this module:
|
|
6
|
-
*
|
|
7
|
-
* 1. **Silent judge fallback** — the judge degraded to a fallback
|
|
8
|
-
* path (rules-only / cached / heuristic) without anyone
|
|
9
|
-
* noticing. Signature: a string of consecutive runs whose
|
|
10
|
-
* `judgeMetadata.confidence` equals a known fallback constant
|
|
11
|
-
* (default 0.30) OR whose `judgeMetadata.fallback` is true.
|
|
12
|
-
*
|
|
13
|
-
* 2. **Judge calibration drift** — the judge's confidence
|
|
14
|
-
* distribution has drifted from a historical window. Two-sample
|
|
15
|
-
* Kolmogorov-Smirnov test on the recent vs historical confidences,
|
|
16
|
-
* with the empirical-CDF max-difference statistic.
|
|
17
|
-
*
|
|
18
|
-
* 3. **Eval-set distribution shift** — the mix of categories /
|
|
19
|
-
* buckets in the recent runs differs significantly from the
|
|
20
|
-
* historical mix. Chi-square test on the binned counts.
|
|
21
|
-
*
|
|
22
|
-
* Outputs are alerts. The canary does NOT fail loud the way a test
|
|
23
|
-
* does — failing tests are reserved for hard correctness violations.
|
|
24
|
-
* A canary that fires is a *signal* to investigate, not a verdict.
|
|
25
|
-
*
|
|
26
|
-
* Why this lives here rather than in `observability.ts`: that module
|
|
27
|
-
* exports already, and is a pure-fanout-to-Langfuse/Prometheus
|
|
28
|
-
* adapter. Canaries are statistical detectors, not adapters.
|
|
29
|
-
*/
|
|
30
|
-
import type { RunRecord } from './run-record';
|
|
31
|
-
export type CanaryKind = 'silent_judge_fallback' | 'judge_calibration_drift' | 'distribution_shift';
|
|
32
|
-
export type CanarySeverity = 'info' | 'warn' | 'error';
|
|
33
|
-
export interface CanaryAlert {
|
|
34
|
-
kind: CanaryKind;
|
|
35
|
-
severity: CanarySeverity;
|
|
36
|
-
message: string;
|
|
37
|
-
/** Numbers that informed the decision — drop straight into a
|
|
38
|
-
* dashboard / paper figure. */
|
|
39
|
-
evidence: Record<string, unknown>;
|
|
40
|
-
}
|
|
41
|
-
export interface CanaryReport {
|
|
42
|
-
alerts: CanaryAlert[];
|
|
43
|
-
/** Per-kind summary count. */
|
|
44
|
-
counts: Record<CanaryKind, number>;
|
|
45
|
-
}
|
|
46
|
-
export interface CanaryOptions {
|
|
47
|
-
/**
|
|
48
|
-
* Silent-fallback detection.
|
|
49
|
-
* - `constant`: confidence value treated as the fallback signal.
|
|
50
|
-
* Default 0.30 (matches the soft-fail default in
|
|
51
|
-
* `propose-review.ts`).
|
|
52
|
-
* - `consecutiveThreshold`: trip the alert after this many
|
|
53
|
-
* consecutive runs at `constant` (or `fallback === true`).
|
|
54
|
-
* Default 3.
|
|
55
|
-
*/
|
|
56
|
-
silentFallback?: {
|
|
57
|
-
constant?: number;
|
|
58
|
-
consecutiveThreshold?: number;
|
|
59
|
-
/** Floating-point tolerance when comparing against `constant`. */
|
|
60
|
-
epsilon?: number;
|
|
61
|
-
};
|
|
62
|
-
/**
|
|
63
|
-
* Calibration-drift detection.
|
|
64
|
-
* - `historyWindow`: number of past runs (oldest-first) treated as
|
|
65
|
-
* the historical baseline. Default 50.
|
|
66
|
-
* - `recentWindow`: number of recent runs (newest-first) compared
|
|
67
|
-
* against history. Default 20.
|
|
68
|
-
* - `ksAlpha`: alpha for the KS statistic vs critical value.
|
|
69
|
-
* Default 0.05.
|
|
70
|
-
* - `minRecent`: minimum recent runs required to even attempt the
|
|
71
|
-
* check. Default 10.
|
|
72
|
-
*/
|
|
73
|
-
calibrationDrift?: {
|
|
74
|
-
historyWindow?: number;
|
|
75
|
-
recentWindow?: number;
|
|
76
|
-
ksAlpha?: number;
|
|
77
|
-
minRecent?: number;
|
|
78
|
-
};
|
|
79
|
-
/**
|
|
80
|
-
* Distribution-shift detection.
|
|
81
|
-
* - `category`: function that maps a run to a categorical bucket.
|
|
82
|
-
* Required to enable this canary; if omitted the chi-square check
|
|
83
|
-
* is skipped entirely.
|
|
84
|
-
* - `chiSquareAlpha`: alpha. Default 0.05.
|
|
85
|
-
* - `historyWindow`, `recentWindow`, `minRecent`: like above.
|
|
86
|
-
*/
|
|
87
|
-
distributionShift?: {
|
|
88
|
-
category: (run: RunRecord) => string | null;
|
|
89
|
-
chiSquareAlpha?: number;
|
|
90
|
-
historyWindow?: number;
|
|
91
|
-
recentWindow?: number;
|
|
92
|
-
minRecent?: number;
|
|
93
|
-
};
|
|
94
|
-
}
|
|
95
|
-
/**
|
|
96
|
-
* Run all configured canaries against a chronological run list.
|
|
97
|
-
* Runs MUST be sorted oldest-to-newest by the caller — the order of
|
|
98
|
-
* the input is used to define "recent" vs "historical" windows.
|
|
99
|
-
*/
|
|
100
|
-
export declare function runCanaries(runs: RunRecord[], opts?: CanaryOptions): CanaryReport;
|
|
101
|
-
//# sourceMappingURL=canary.d.ts.map
|
package/dist/canary.d.ts.map
DELETED
|
@@ -1 +0,0 @@
|
|
|
1
|
-
{"version":3,"file":"canary.d.ts","sourceRoot":"","sources":["../src/canary.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;GA4BG;AAEH,OAAO,KAAK,EAAE,SAAS,EAAE,MAAM,cAAc,CAAA;AAE7C,MAAM,MAAM,UAAU,GAAG,uBAAuB,GAAG,yBAAyB,GAAG,oBAAoB,CAAA;AAEnG,MAAM,MAAM,cAAc,GAAG,MAAM,GAAG,MAAM,GAAG,OAAO,CAAA;AAEtD,MAAM,WAAW,WAAW;IAC1B,IAAI,EAAE,UAAU,CAAA;IAChB,QAAQ,EAAE,cAAc,CAAA;IACxB,OAAO,EAAE,MAAM,CAAA;IACf;oCACgC;IAChC,QAAQ,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,CAAA;CAClC;AAED,MAAM,WAAW,YAAY;IAC3B,MAAM,EAAE,WAAW,EAAE,CAAA;IACrB,8BAA8B;IAC9B,MAAM,EAAE,MAAM,CAAC,UAAU,EAAE,MAAM,CAAC,CAAA;CACnC;AAED,MAAM,WAAW,aAAa;IAC5B;;;;;;;;OAQG;IACH,cAAc,CAAC,EAAE;QACf,QAAQ,CAAC,EAAE,MAAM,CAAA;QACjB,oBAAoB,CAAC,EAAE,MAAM,CAAA;QAC7B,kEAAkE;QAClE,OAAO,CAAC,EAAE,MAAM,CAAA;KACjB,CAAA;IAED;;;;;;;;;;OAUG;IACH,gBAAgB,CAAC,EAAE;QACjB,aAAa,CAAC,EAAE,MAAM,CAAA;QACtB,YAAY,CAAC,EAAE,MAAM,CAAA;QACrB,OAAO,CAAC,EAAE,MAAM,CAAA;QAChB,SAAS,CAAC,EAAE,MAAM,CAAA;KACnB,CAAA;IAED;;;;;;;OAOG;IACH,iBAAiB,CAAC,EAAE;QAClB,QAAQ,EAAE,CAAC,GAAG,EAAE,SAAS,KAAK,MAAM,GAAG,IAAI,CAAA;QAC3C,cAAc,CAAC,EAAE,MAAM,CAAA;QACvB,aAAa,CAAC,EAAE,MAAM,CAAA;QACtB,YAAY,CAAC,EAAE,MAAM,CAAA;QACrB,SAAS,CAAC,EAAE,MAAM,CAAA;KACnB,CAAA;CACF;AAED;;;;GAIG;AACH,wBAAgB,WAAW,CAAC,IAAI,EAAE,SAAS,EAAE,EAAE,IAAI,GAAE,aAAkB,GAAG,YAAY,CAarF"}
|