@tangle-network/agent-eval 0.115.2 → 0.116.0

package/dist/campaign/index.d.ts

@@ -1,33 +1,32 @@
-import { P as PairedArmsComparison, S as SignedManifest, B as BackendIntegrityReport, C as CompletionRequirement, R as RuntimeEventLike, a as CompletionVerdict, b as ProducedState, c as CorrectnessChecker } from '../pre-registration-D8h7ZxNL.js';
-export { L as LlmJudgeDimension, d as LlmJudgeOptions, l as llmJudge } from '../pre-registration-D8h7ZxNL.js';
-import { A as AnalyzeTracesOptions, a as AnalyzeTracesInput, b as AnalyzeTracesResult } from '../analyst-C8HHvfJp.js';
-import { S as Scenario, M as MutableSurface, D as DispatchContext, b as JudgeConfig, g as Gate, e as GenerationRecord, J as JudgeScore, L as LabeledScenarioStore, s as LabeledScenarioWrite, t as LabeledScenarioSampleArgs, u as LabeledScenarioRecord, v as LabelTrust, f as SurfaceProposer, w as ProposedCandidate, x as ProposeContext, m as CodeSurface, y as LabeledScenarioSource, C as CampaignResult } from '../types-C5gJrOVT.js';
-export { i as CampaignAggregates, j as CampaignArtifactWriter, k as CampaignCellResult, l as CampaignCostMeter, z as CampaignTokenUsage, d as CampaignTraceWriter, c as DispatchFn, n as GateContext, h as GateDecision, G as GateResult, o as GenerationCandidate, A as JudgeAggregate, a as JudgeDimension, p as Mutator, O as OptimizationProposer, q as OptimizerConfig, P as ParetoParent, R as RedactionStatus, B as ScenarioAggregate, r as SessionScript, T as TraceSpan, E as isProposedCandidate, F as labelTrustRank } from '../types-C5gJrOVT.js';
-import { C as CampaignRunPlan, P as PlanCampaignRunOptions, b as RunCampaignOptions, c as RunImprovementLoopOptions } from '../gepa-dne9JDPL.js';
-export { f as CampaignRunPlanCell, h as GepaProposerConstraints, G as GepaProposerOptions, O as OpenAutoPrOptions, i as OpenAutoPrResult, a as RunImprovementLoopResult, R as RunOptimizationOptions, j as RunOptimizationResult, k as countSentenceEdits, l as defaultRenderDiff, m as extractH2Sections, g as gepaProposer, o as openAutoPr, p as planCampaignRun, r as runCampaign, d as runImprovementLoop, n as runOptimization } from '../gepa-dne9JDPL.js';
+import { L as LlmClientOptions, w as PolicyEditAdmissionOptions, v as PolicyEditAdmission, u as PolicyEdit, d as AnalystFinding, F as FindingToPolicyEditOptions } from '../policy-edit-Clb2v6Oa.js';
+export { s as POLICY_EDIT_CANDIDATE_RECORD_SCHEMA, P as PolicyEditCandidateRecord, a4 as validatePolicyEditCandidateRecord } from '../policy-edit-Clb2v6Oa.js';
+import { P as PairedArmsComparison, S as SignedManifest, B as BackendIntegrityReport, C as CompletionRequirement, R as RuntimeEventLike, a as CompletionVerdict, b as ProducedState, c as CorrectnessChecker } from '../pre-registration--vU0mMtD.js';
+export { L as LlmJudgeDimension, d as LlmJudgeOptions, l as llmJudge } from '../pre-registration--vU0mMtD.js';
+import { A as AnalyzeTracesOptions, a as AnalyzeTracesInput, b as AnalyzeTracesResult } from '../analyst-CFBc14Wc.js';
+import { S as Scenario, M as MutableSurface, D as DispatchContext, b as JudgeConfig, g as Gate, e as GenerationRecord, J as JudgeScore, L as LabeledScenarioStore, s as LabeledScenarioWrite, t as LabeledScenarioSampleArgs, u as LabeledScenarioRecord, v as LabelTrust, f as SurfaceProposer, w as ProposedCandidate, x as ProposeContext, m as CodeSurface, y as LabeledScenarioSource, C as CampaignResult } from '../types-Ca_63YSD.js';
+export { i as CampaignAggregates, j as CampaignArtifactWriter, k as CampaignCellResult, l as CampaignCostMeter, z as CampaignTokenUsage, d as CampaignTraceWriter, c as DispatchFn, n as GateContext, h as GateDecision, G as GateResult, o as GenerationCandidate, A as JudgeAggregate, a as JudgeDimension, p as Mutator, O as OptimizationProposer, q as OptimizerConfig, P as ParetoParent, R as RedactionStatus, B as ScenarioAggregate, E as ScoredSurfaceOutcome, r as SessionScript, T as TraceSpan, F as isProposedCandidate, H as labelTrustRank } from '../types-Ca_63YSD.js';
+import { C as CampaignRunPlan, P as PlanCampaignRunOptions, b as RunCampaignOptions, c as RunImprovementLoopOptions } from '../gepa-CQelRtuC.js';
+export { f as CampaignRunPlanCell, h as GepaProposerConstraints, G as GepaProposerOptions, O as OpenAutoPrOptions, i as OpenAutoPrResult, a as RunImprovementLoopResult, R as RunOptimizationOptions, j as RunOptimizationResult, k as countSentenceEdits, l as defaultRenderDiff, m as extractH2Sections, g as gepaProposer, o as openAutoPr, p as planCampaignRun, r as runCampaign, d as runImprovementLoop, n as runOptimization } from '../gepa-CQelRtuC.js';
 import { a as PairedBootstrapResult, E as EProcessState } from '../statistics-oUbOJe-S.js';
 import { C as CampaignStorage } from '../storage-Dw_f7WMt.js';
 export { f as fsCampaignStorage, i as inMemoryCampaignStorage } from '../storage-Dw_f7WMt.js';
-export { A as AxisEvidence, a as AxisVerdict, B as BuildEvidenceVectorOptions, l as BuildLoopProvenanceArgs, D as DefaultProductionGateOptions, m as EmitLoopProvenanceArgs, n as EmitLoopProvenanceResult, E as EvidenceVector, b as EvolutionaryProposerOptions, H as HeldOutGateOptions, o as LoopProvenanceBackend, q as LoopProvenanceCandidate, L as LoopProvenanceRecord, O as ObjectiveSource, c as ParetoSignificanceGateOptions, P as PowerPreflight, s as PowerPreflightOptions, d as PromotionObjective, e as PromotionPolicy, R as RunEvalOptions, f as buildEvidenceVector, t as buildLoopProvenanceRecord, g as composeGate, h as defaultProductionGate, u as emitLoopProvenance, i as evolutionaryProposer, j as heldOutGate, v as loopProvenanceSpans, p as paretoPolicy, k as paretoSignificanceGate, w as powerPreflight, x as provenanceRecordPath, y as provenanceSpansPath, r as runEval } from '../provenance-Bibyg1U9.js';
-import { L as LlmClientOptions } from '../llm-client-DyqEH4jH.js';
+export { A as AxisEvidence, a as AxisVerdict, B as BuildEvidenceVectorOptions, l as BuildLoopProvenanceArgs, D as DefaultProductionGateOptions, m as EmitLoopProvenanceArgs, n as EmitLoopProvenanceResult, E as EvidenceVector, b as EvolutionaryProposerOptions, H as HeldOutGateOptions, o as LoopProvenanceBackend, q as LoopProvenanceCandidate, L as LoopProvenanceRecord, O as ObjectiveSource, c as ParetoSignificanceGateOptions, P as PowerPreflight, s as PowerPreflightOptions, d as PromotionObjective, e as PromotionPolicy, R as RunEvalOptions, f as buildEvidenceVector, t as buildLoopProvenanceRecord, g as composeGate, h as defaultProductionGate, u as emitLoopProvenance, i as evolutionaryProposer, j as heldOutGate, v as loopProvenanceSpans, p as paretoPolicy, k as paretoSignificanceGate, w as powerPreflight, x as provenanceRecordPath, y as provenanceSpansPath, r as runEval } from '../provenance-BbVagC68.js';
 import { AgentProfile } from '@tangle-network/agent-interface';
 import { A as AgentEvalError, V as ValidationError } from '../errors-oeQrLqXC.js';
-import { b as RunSplitTag, R as RunRecord } from '../run-record-B7RTi_ix.js';
-import { b as PolicyEdit, F as FindingToPolicyEditOptions, d as PolicyEditAdmissionOptions, c as PolicyEditAdmission } from '../policy-edit-RLn8GWof.js';
-import { T as TraceAnalystKindSpec, A as AnalystFinding } from '../kind-factory-DcNg13sZ.js';
+import { a as RunSplitTag, R as RunRecord } from '../run-record-CZmcpWPo.js';
+import { T as TraceAnalystKindSpec } from '../kind-factory-DWOvXjR_.js';
+import '../store-9cAScOcb.js';
+import '../types-C7DGg5ex.js';
 import '@tangle-network/tcloud';
-import '../raw-provider-sink-C46HDghv.js';
 import '../verdict-C9MlYujm.js';
 import '@ax-llm/ax';
-import '../store-C1YxJDEK.js';
 import '../dataset-NENEzRgk.js';
 import '../store-BsVi7ncX.js';
 import '../schema-SGWcK9wa.js';
 import '../judge-calibration-7C-IDmKr.js';
-import '../types-C7DGg5ex.js';
 import '../hosted/index.js';
-import '../insight-report-IwwvqZZv.js';
-import '../summary-report-BJ5aNwZ1.js';
+import '../insight-report-oMVxDTxl.js';
+import '../summary-report-DTNgQycC.js';
 import '../failure-cluster-C48PiReX.js';
 import 'zod';
 
@@ -1989,8 +1988,8 @@ declare function parseSkillPatchResponse(raw: string, maxPatches: number, editBu
 
 /**
  * `runSkillOpt` — the SkillOpt epoch hill-climb (Microsoft, arXiv:2605.23904).
- * Unlike `runOptimization`'s population/promote-top-K search, SkillOpt is a
- * sequential, held-out-gated hill-climb on ONE skill document:
+ * Unlike `runOptimization`'s population search around one global incumbent,
+ * SkillOpt is a sequential, held-out-gated hill-climb on ONE skill document:
  *
  *   each epoch:
  *     1. reflect on the CURRENT surface's weakest TRAIN scenarios/dimensions
@@ -2208,6 +2207,167 @@ interface HaloProposerOptions {
 /** Wrap the real halo-engine CLI as a SurfaceProposer (prompt-tier). */
 declare function haloProposer(opts: HaloProposerOptions): SurfaceProposer;
 
+declare const JSON_POLICY_EDIT_TARGET_SURFACES: readonly ["prompt", "tool-contract", "runtime-config", "memory", "agent-profile"];
+type JsonPolicyEditTargetSurface = (typeof JSON_POLICY_EDIT_TARGET_SURFACES)[number];
+declare const DEFAULT_POLICY_EDIT_HISTORY_LIMITS: Readonly<{
+    generations: 4;
+    candidatesPerGeneration: 16;
+    scenariosPerCandidate: 12;
+    findings: 32;
+    authorContextChars: 200000;
+}>;
+interface PolicyEditObjective {
+    /** Stable objective key cited by forecasts, for example `search.composite`. */
+    key: string;
+    /** Steering objectives are search-only; fresh-task results never enter author context. */
+    split: 'search';
+    /** Search promotes only larger composite scores. */
+    direction: 'increase';
+    scale: {
+        min: number;
+        max: number;
+    };
+    /** Forecast amounts are absolute deltas on the declared score scale. */
+    unit: 'score';
+}
+type PolicyEditFindingSource = {
+    kind: 'surface';
+    surfaceHash: string;
+    generation: number;
+} | {
+    kind: 'global';
+    label: string;
+};
+/** Trace-derived findings must name the measured profile that produced them.
+ *  Cross-run doctrine can be explicitly global; an unwrapped finding is rejected. */
+interface PolicyEditFindingInput {
+    finding: AnalystFinding;
+    source: PolicyEditFindingSource;
+}
+interface PolicyEditHistoryProjectionOptions {
+    /** Number of most recent generations retained. Default: 4. */
+    maxGenerations?: number;
+    /** Number of candidates retained per generation. Default: 16. */
+    maxCandidatesPerGeneration?: number;
+    /** Scored tasks retained per candidate/outcome after deterministic extreme selection. */
+    maxScenariosPerCandidate?: number;
+    /** Optional pseudonymizer applied before scenario IDs enter author text. */
+    scenarioIdTransform?: (scenarioId: string) => string;
+    /** Objectives used to compute forecast residuals from measured composite deltas. */
+    objectives?: readonly PolicyEditObjective[];
+}
+interface PolicyEditCandidateSummary {
+    editId: string;
+    axis: PolicyEdit['axis'];
+    target: PolicyEdit['target'];
+    change: PolicyEdit['change'];
+    claim: string;
+    expectedGain: PolicyEdit['expectedGain'];
+    confidence: number;
+    risk: PolicyEdit['risk'];
+    sourceFindingIds: string[];
+    rationale: string | null;
+    validationPlan: string | null;
+}
+interface PolicyEditHistoryCandidateContext {
+    surfaceHash: string;
+    parentSurfaceHash: string | null;
+    parentComposite: number | null;
+    label: string | null;
+    rationale: string | null;
+    composite: number;
+    observedDeltaFromParent: number | null;
+    eligibleForPromotion: boolean | null;
+    coverage: {
+        expectedCells: number;
+        scorableCells: number;
+        unscorableCells: Array<{
+            reason: string;
+        }>;
+    } | null;
+    dimensions: Record<string, number>;
+    scenarios: Array<{
+        scenarioId: string;
+        composite: number;
+        notes: string | null;
+    }>;
+    candidateEdit: PolicyEditCandidateSummary | null;
+    forecastCalibration: {
+        objectiveKey: string;
+        predictedDelta: number;
+        observedDelta: number;
+        residual: number;
+    } | null;
+}
+interface PolicyEditOutcomeContext {
+    split: 'search';
+    generation: number;
+    surfaceHash: string;
+    composite: number;
+    dimensions: Record<string, number>;
+    scenarios: Array<{
+        scenarioId: string;
+        composite: number;
+        notes: string | null;
+    }>;
+    coverage: {
+        expectedCells: number;
+        scorableCells: number;
+    };
+}
+interface PolicyEditHistoryGenerationContext {
+    generationIndex: number;
+    promoted: string[];
+    candidates: PolicyEditHistoryCandidateContext[];
+}
+interface LlmPolicyEditProposerOptions {
+    llm: LlmClientOptions;
+    model: string;
+    /** Plain-language description of the JSON surface being improved. */
+    target: string;
+    /** PolicyEdit target surface every authored edit must retain. */
+    targetSurface: JsonPolicyEditTargetSurface;
+    /** Exact JSON paths the author may change. Prefix or fuzzy matches are not accepted. */
+    allowedJsonPaths: readonly string[];
+    /** Exact search objectives forecasts may name. Unknown keys or mismatched directions fail. */
+    objectives: readonly PolicyEditObjective[];
+    /** Default: evidence-only, so uncertain edits are measured rather than
+     *  suppressed by their own model-authored predictions. */
+    admissionMode?: 'evidence-only' | 'strict';
+    /** Readiness thresholds used only when admissionMode is explicitly strict. */
+    admission?: PolicyEditAdmissionOptions;
+    maxCandidates?: number;
+    temperature?: number;
+    maxTokens?: number;
+    timeoutMs?: number;
+    /** Number of most recent scored generations sent to the author. Default: 4. */
+    maxHistoryGenerations?: number;
+    /** Candidates retained per admitted generation. Default: 16. */
+    maxHistoryCandidatesPerGeneration?: number;
+    /** Scored tasks retained per candidate/outcome after deterministic extreme selection. */
+    maxScenariosPerCandidate?: number;
+    /** Evidence-bearing findings retained after deterministic severity/confidence ordering. */
+    maxFindings?: number;
+    /** Hard character limit over system + schema + serialized author context. */
+    maxAuthorContextChars?: number;
+    /** Optional one-to-one pseudonymizer applied to every author-visible evidence field. */
+    scenarioIdTransform?: (scenarioId: string) => string;
+    onAdmission?: (admission: PolicyEditAdmission) => void;
+}
+/**
+ * LLM-backed PolicyEdit author. It reads only the current JSON surface,
+ * evidence-bearing analyst findings, and scored generation history. Model output
+ * is validated and rebound to exact finding evidence before the deterministic
+ * policyEditProposer applies, admits, and deduplicates candidates.
+ */
+declare function llmPolicyEditProposer(opts: LlmPolicyEditProposerOptions): SurfaceProposer<PolicyEditFindingInput>;
+/**
+ * Projects scored history into the only fields a policy author may consume.
+ * It admits recent generations and a bounded candidate count, while retaining
+ * every dimension and scenario score for each admitted candidate.
+ */
+declare function projectPolicyEditHistory(history: readonly GenerationRecord[], options?: PolicyEditHistoryProjectionOptions): PolicyEditHistoryGenerationContext[];
+
 /**
  * `memoryCurationProposer` — a CURATOR `SurfaceProposer`, the complement to the
  * OPTIMIZER proposers (`gepaProposer` rewrites the prompt; this one BUILDS a
@@ -2281,6 +2441,33 @@ interface PolicyEditProposerOptions {
  */
 declare function policyEditProposer(opts?: PolicyEditProposerOptions): SurfaceProposer;
 
+/** One measured scenario row eligible for PolicyEdit author context. */
+interface PolicyEditAuthorScenarioRow {
+    scenarioId: string;
+    composite: number;
+}
+interface SelectPolicyEditAuthorRowsOptions {
+    /** Maximum returned rows. Must be a positive safe integer. */
+    limit: number;
+    /** Optional score to compare against, keyed by scenario ID. */
+    referenceByScenario?: ReadonlyMap<string, number>;
+}
+interface SerializedJsonBudget {
+    json: string;
+    actualChars: number;
+    maxChars: number;
+}
+/**
+ * Select a bounded, deterministic evidence slice for a PolicyEdit author.
+ *
+ * Rows are deduplicated by scenario ID, keeping the first measured row. The
+ * result then interleaves three ranked views: hardest score, largest regression,
+ * and largest improvement. A row selected by multiple views appears once.
+ */
+declare function selectPolicyEditAuthorRows<T extends PolicyEditAuthorScenarioRow>(rows: readonly T[], options: SelectPolicyEditAuthorRowsOptions): T[];
+/** Serialize once and fail before dispatch when author context exceeds its budget. */
+declare function assertPolicyEditAuthorContextBudget(value: unknown, maxChars: number): SerializedJsonBudget;
+
 /**
  * `traceAnalystProposer` — wraps agent-eval's OWN trace-analyst engine
  * (`AnalystRegistry` over the agentic OTLP reader) as a `SurfaceProposer`.
@@ -2399,9 +2586,11 @@ declare function selectDiscriminative(signals: ScenarioSignal[], k: number, opts
  * the optimizers cannot drift on how a surface's score is computed.
  */
 
-/** Mean composite across a campaign: per cell, the mean of its judges'
- *  composites; then the mean across cells. Cells with no judge scores are
- *  skipped. Empty ⇒ 0. */
+/** Mean composite across a campaign: per cell, the mean of its finite,
+ *  successful judge composites; then the mean across cells. Invalid scores
+ *  remain visible on raw cells and coverage receipts but never poison the
+ *  descriptive aggregate with NaN. Cells with no valid scores are skipped.
+ *  Empty ⇒ 0. */
 declare function campaignMeanComposite<TArtifact, TScenario extends Scenario>(campaign: CampaignResult<TArtifact, TScenario>): number;
 interface CampaignBreakdown {
     /** Mean score per judge dimension across all cells. */
@@ -2899,4 +3088,4 @@ declare function verifyCodeSurface(surface: CodeSurface, worktreeDir?: string):
  *  identity against the checkout at `worktreeRef`. */
 declare function resolveWorktreePath(surface: CodeSurface, worktreeDir?: string): string;
 
-export { type AcceptedEdit, type AceProposerOptions, type AnalystArtifact, type AnalystScenario, type AnalyzeCrossSurfaceInteractionsInput, type ApplySkillPatchResult, type BuildAnalystSurfaceDispatchOptions, type CampaignBreakdown, CampaignResult, CampaignRunPlan, CampaignStorage, CodeSurface, type CodeSurfaceVerification, type CompareProposersOptions, type CompositeProposerOptions, type CrossSurfaceAdditionDecision, type CrossSurfaceAdditionRejectionReason, type CrossSurfaceAttemptCompleteness, type CrossSurfaceBestSingleSelection, type CrossSurfaceBootstrapPolicy, type CrossSurfaceCandidate, type CrossSurfaceCandidateComparison, type CrossSurfaceCandidateEvidence, type CrossSurfaceCandidateOutcome, type CrossSurfaceCandidateSummary, type CrossSurfaceComponent, type CrossSurfaceComponentEvidence, type CrossSurfaceCompositionStep, type CrossSurfaceDistribution, type CrossSurfaceEligibility, type CrossSurfaceEvidenceBreakdown, type CrossSurfaceIneligibilityReason, type CrossSurfaceInteractionAwareSelection, type CrossSurfaceInteractionEffect, type CrossSurfaceInteractionPath, type CrossSurfaceInteractionReport, type CrossSurfaceInteractionTask, type CrossSurfaceNaiveStackSelection, type CrossSurfacePairCompatibility, type CrossSurfacePairEvidence, type CrossSurfacePairIncompatibilityReason, type CrossSurfacePairwiseEntry, type CrossSurfaceRankedSingle, type CrossSurfaceRelativeCost, type CrossSurfaceSelectionPolicy, type CrossSurfaceSelections, type CrossSurfaceTaskRow, type DimensionRegression, type DiscriminationScore, DispatchContext, type EvalFixture, type EvalFixtureFile, type EvalFixtureLoadOptions, type EvalFixtureRunPlan, type EvalFixtureScenario, type EvalFixtureValidationMode, type FailureModeRecallJudgeOptions, type FapoAttributionSignals, type FapoEntryConfig, type FapoFailureCluster, type FapoOptimizationLevel, type FapoProposerOptions, type FapoReviewInput, type FapoReviewIssue, type FapoReviewResult, type FapoScopeContract, FileSearchLedger, FsLabeledScenarioStore, type FsLabeledScenarioStoreOptions, Gate, GenerationRecord, type GitWorktreeAdapterOptions, type Governor, type GovernorContext, type GovernorOp, type HaloProposerOptions, type HeldoutSignificance, type HeldoutSignificanceOptions, type HeuristicGovernorOptions, type JsonPrimitive, type JsonValue, JudgeConfig, JudgeScore, LabelTrust, LabeledScenarioRecord, LabeledScenarioSampleArgs, LabeledScenarioSource, LabeledScenarioStore, LabeledScenarioStoreError, LabeledScenarioWrite, Lineage, type LineageEdge, type LineageGraph, type LineageNode, type LineageNodeInput, type LineageStore, type LoadEvalFixtureScenariosOptions, type MemoryCurationProposerOptions, MutableSurface, type NeutralizationGateOptions, type OpenSearchLedgerOptions, type OptimizerEntryConfig, type PairedHoldout, type ParameterCandidate, type ParameterChange, type ParameterSweepProposerOptions, PlanCampaignRunOptions, type PlanEvalFixtureRunOptions, type PlaybackContext, type PlaybackDriver, type PlaybackStep, type PolicyEditProposerOptions, type ProfileDispatchFn, ProfileMatrixError, type ProfileSummary, ProposeContext, type ProposePatchesArgs, ProposedCandidate, type ProposerComparison, type ProposerEntry, type ProposerPairwise, type ProposerScore, type RejectedEdit, type RolloutArgumentDiff, type RolloutArgumentDiffOptions, type RolloutCall, RunCampaignOptions, RunImprovementLoopOptions, type RunLineageLoopOptions, type RunLineageLoopResult, type RunLineageLoopSeed, type RunLineageOptions, type RunLineageResult, type RunLineageSeed, type RunLineageStepResult, type RunProfileMatrixOptions, type RunProfileMatrixResult, type RunSkillOptOptions, type RunSkillOptResult, SEARCH_LEDGER_SCHEMA, Scenario, type ScenarioRollup, type ScenarioSignal, type ScoreboardRenderOptions, type ScoreboardRow, type ScoreboardSummary, type ScoredRollout, type SearchAccountingAudit, type SearchArtifactRef, type SearchAttemptAccounting, type SearchCandidateDecidedEvent, type SearchCandidateLineage, type SearchCandidateRegisteredEvent, type SearchCandidateSlot, type SearchCandidateSlotClosedEvent, type SearchCandidateSurface, type SearchCompletedEvent, type SearchCostAccounting, type SearchFailureReason, type SearchLedger, type SearchLedgerAppendResult, SearchLedgerConflictError, type SearchLedgerEntry, SearchLedgerError, type SearchLedgerEvent, type SearchLedgerHash, SearchLedgerIntegrityError, type SearchLedgerReplay, type SearchModelIdentity, type SearchOperationKind, type SearchOperationRecordedEvent, type SearchPlan, type SearchPlannedEvent, type SearchPlannedOperation, type SearchPlannedTask, type SearchSourceRef, type SearchSurfaceEffect, type SearchSurfaceEvidence, type SearchSurfaceKind, type SearchTaskAttemptedEvent, type SearchTaskOutcome, type SearchTokenAccounting, type SequentialDecideFn, type SequentialDecideOptions, type SequentialDecision, type SequentialObservation, type SequentialPairedGate, type SequentialPairedGateOptions, type SingleRunLock, type SingleRunLockOptions, type SkillOptEpochRecord, type SkillOptEvidence, type SkillOptProposer, type SkillOptProposerOptions, type SkillPatch, type SkillPatchOp, SkillPatchParseError, type SkillPatchRejection, SurfaceProposer, type SurfaceScore, type TraceAnalystProposerOptions, type TransientFailureOptions, type UngroundedLiteralReport, type UserStory, type UserStoryVerdict, type Worktree, type WorktreeAdapter, WorktreeAdapterError, aceProposer, acquireSingleRunLock, analyzeCrossSurfaceInteractions, applySkillPatch, assertCodeSurfaceIdentity, buildAnalystSurfaceDispatch, callbackGovernor, campaignBreakdown, campaignMeanComposite, classifyUngroundedLiterals, codeSurfaceIdentityMaterial, compareProposers, compositeProposer, detectScale, dimensionRegressions, discoverEvalFixtures, extractFapoAttributionSignals, failureModeRecallJudge, fapoEscalationEntry, fapoProposer, fsLineageStore, gepaParetoEntry, gepaReflectionEntry, gitWorktreeAdapter, haloProposer, heldoutSignificance, heuristicGovernor, isTransientTransportFailure, lineageNodeId, loadEvalFixture, loadEvalFixtureScenarios, makePlaybackDispatch, memLineageStore, memoryCurationProposer, neutralizationGate, neutralizeText, openSearchLedger, pairHoldout, parameterSweepProposer, parseSkillPatchResponse, patchEditCount, planEvalFixtureRun, policyEditProposer, renderScoreboardMarkdown, resolveRunDir, resolveWorktreePath, rolloutArgumentDiff, runLineage, runLineageLoop, runProfileMatrix, runSkillOpt, scoreDiscrimination, scoreUserStory, scoreboardSummary, selectDiscriminative, sequentialDecide, sequentialPairedGate, skillOptEntry, skillOptProposer, surfaceContentHash, surfaceHash, tangleTracesRoot, traceAnalystProposer, userStoryScoreboard, validateSearchLedgerEvent, verifyCodeSurface };
+export { type AcceptedEdit, type AceProposerOptions, type AnalystArtifact, type AnalystScenario, type AnalyzeCrossSurfaceInteractionsInput, type ApplySkillPatchResult, type BuildAnalystSurfaceDispatchOptions, type CampaignBreakdown, CampaignResult, CampaignRunPlan, CampaignStorage, CodeSurface, type CodeSurfaceVerification, type CompareProposersOptions, type CompositeProposerOptions, type CrossSurfaceAdditionDecision, type CrossSurfaceAdditionRejectionReason, type CrossSurfaceAttemptCompleteness, type CrossSurfaceBestSingleSelection, type CrossSurfaceBootstrapPolicy, type CrossSurfaceCandidate, type CrossSurfaceCandidateComparison, type CrossSurfaceCandidateEvidence, type CrossSurfaceCandidateOutcome, type CrossSurfaceCandidateSummary, type CrossSurfaceComponent, type CrossSurfaceComponentEvidence, type CrossSurfaceCompositionStep, type CrossSurfaceDistribution, type CrossSurfaceEligibility, type CrossSurfaceEvidenceBreakdown, type CrossSurfaceIneligibilityReason, type CrossSurfaceInteractionAwareSelection, type CrossSurfaceInteractionEffect, type CrossSurfaceInteractionPath, type CrossSurfaceInteractionReport, type CrossSurfaceInteractionTask, type CrossSurfaceNaiveStackSelection, type CrossSurfacePairCompatibility, type CrossSurfacePairEvidence, type CrossSurfacePairIncompatibilityReason, type CrossSurfacePairwiseEntry, type CrossSurfaceRankedSingle, type CrossSurfaceRelativeCost, type CrossSurfaceSelectionPolicy, type CrossSurfaceSelections, type CrossSurfaceTaskRow, DEFAULT_POLICY_EDIT_HISTORY_LIMITS, type DimensionRegression, type DiscriminationScore, DispatchContext, type EvalFixture, type EvalFixtureFile, type EvalFixtureLoadOptions, type EvalFixtureRunPlan, type EvalFixtureScenario, type EvalFixtureValidationMode, type FailureModeRecallJudgeOptions, type FapoAttributionSignals, type FapoEntryConfig, type FapoFailureCluster, type FapoOptimizationLevel, type FapoProposerOptions, type FapoReviewInput, type FapoReviewIssue, type FapoReviewResult, type FapoScopeContract, FileSearchLedger, FsLabeledScenarioStore, type FsLabeledScenarioStoreOptions, Gate, GenerationRecord, type GitWorktreeAdapterOptions, type Governor, type GovernorContext, type GovernorOp, type HaloProposerOptions, type HeldoutSignificance, type HeldoutSignificanceOptions, type HeuristicGovernorOptions, type JsonPolicyEditTargetSurface, type JsonPrimitive, type JsonValue, JudgeConfig, JudgeScore, LabelTrust, LabeledScenarioRecord, LabeledScenarioSampleArgs, LabeledScenarioSource, LabeledScenarioStore, LabeledScenarioStoreError, LabeledScenarioWrite, Lineage, type LineageEdge, type LineageGraph, type LineageNode, type LineageNodeInput, type LineageStore, type LlmPolicyEditProposerOptions, type LoadEvalFixtureScenariosOptions, type MemoryCurationProposerOptions, MutableSurface, type NeutralizationGateOptions, type OpenSearchLedgerOptions, type OptimizerEntryConfig, type PairedHoldout, type ParameterCandidate, type ParameterChange, type ParameterSweepProposerOptions, PlanCampaignRunOptions, type PlanEvalFixtureRunOptions, type PlaybackContext, type PlaybackDriver, type PlaybackStep, type PolicyEditAuthorScenarioRow, type PolicyEditCandidateSummary, type PolicyEditFindingInput, type PolicyEditFindingSource, type PolicyEditHistoryCandidateContext, type PolicyEditHistoryGenerationContext, type PolicyEditHistoryProjectionOptions, type PolicyEditObjective, type PolicyEditOutcomeContext, type PolicyEditProposerOptions, type ProfileDispatchFn, ProfileMatrixError, type ProfileSummary, ProposeContext, type ProposePatchesArgs, ProposedCandidate, type ProposerComparison, type ProposerEntry, type ProposerPairwise, type ProposerScore, type RejectedEdit, type RolloutArgumentDiff, type RolloutArgumentDiffOptions, type RolloutCall, RunCampaignOptions, RunImprovementLoopOptions, type RunLineageLoopOptions, type RunLineageLoopResult, type RunLineageLoopSeed, type RunLineageOptions, type RunLineageResult, type RunLineageSeed, type RunLineageStepResult, type RunProfileMatrixOptions, type RunProfileMatrixResult, type RunSkillOptOptions, type RunSkillOptResult, SEARCH_LEDGER_SCHEMA, Scenario, type ScenarioRollup, type ScenarioSignal, type ScoreboardRenderOptions, type ScoreboardRow, type ScoreboardSummary, type ScoredRollout, type SearchAccountingAudit, type SearchArtifactRef, type SearchAttemptAccounting, type SearchCandidateDecidedEvent, type SearchCandidateLineage, type SearchCandidateRegisteredEvent, type SearchCandidateSlot, type SearchCandidateSlotClosedEvent, type SearchCandidateSurface, type SearchCompletedEvent, type SearchCostAccounting, type SearchFailureReason, type SearchLedger, type SearchLedgerAppendResult, SearchLedgerConflictError, type SearchLedgerEntry, SearchLedgerError, type SearchLedgerEvent, type SearchLedgerHash, SearchLedgerIntegrityError, type SearchLedgerReplay, type SearchModelIdentity, type SearchOperationKind, type SearchOperationRecordedEvent, type SearchPlan, type SearchPlannedEvent, type SearchPlannedOperation, type SearchPlannedTask, type SearchSourceRef, type SearchSurfaceEffect, type SearchSurfaceEvidence, type SearchSurfaceKind, type SearchTaskAttemptedEvent, type SearchTaskOutcome, type SearchTokenAccounting, type SelectPolicyEditAuthorRowsOptions, type SequentialDecideFn, type SequentialDecideOptions, type SequentialDecision, type SequentialObservation, type SequentialPairedGate, type SequentialPairedGateOptions, type SerializedJsonBudget, type SingleRunLock, type SingleRunLockOptions, type SkillOptEpochRecord, type SkillOptEvidence, type SkillOptProposer, type SkillOptProposerOptions, type SkillPatch, type SkillPatchOp, SkillPatchParseError, type SkillPatchRejection, SurfaceProposer, type SurfaceScore, type TraceAnalystProposerOptions, type TransientFailureOptions, type UngroundedLiteralReport, type UserStory, type UserStoryVerdict, type Worktree, type WorktreeAdapter, WorktreeAdapterError, aceProposer, acquireSingleRunLock, analyzeCrossSurfaceInteractions, applySkillPatch, assertCodeSurfaceIdentity, assertPolicyEditAuthorContextBudget, buildAnalystSurfaceDispatch, callbackGovernor, campaignBreakdown, campaignMeanComposite, classifyUngroundedLiterals, codeSurfaceIdentityMaterial, compareProposers, compositeProposer, detectScale, dimensionRegressions, discoverEvalFixtures, extractFapoAttributionSignals, failureModeRecallJudge, fapoEscalationEntry, fapoProposer, fsLineageStore, gepaParetoEntry, gepaReflectionEntry, gitWorktreeAdapter, haloProposer, heldoutSignificance, heuristicGovernor, isTransientTransportFailure, lineageNodeId, llmPolicyEditProposer, loadEvalFixture, loadEvalFixtureScenarios, makePlaybackDispatch, memLineageStore, memoryCurationProposer, neutralizationGate, neutralizeText, openSearchLedger, pairHoldout, parameterSweepProposer, parseSkillPatchResponse, patchEditCount, planEvalFixtureRun, policyEditProposer, projectPolicyEditHistory, renderScoreboardMarkdown, resolveRunDir, resolveWorktreePath, rolloutArgumentDiff, runLineage, runLineageLoop, runProfileMatrix, runSkillOpt, scoreDiscrimination, scoreUserStory, scoreboardSummary, selectDiscriminative, selectPolicyEditAuthorRows, sequentialDecide, sequentialPairedGate, skillOptEntry, skillOptProposer, surfaceContentHash, surfaceHash, tangleTracesRoot, traceAnalystProposer, userStoryScoreboard, validateSearchLedgerEvent, verifyCodeSurface };

package/dist/campaign/index.js

@@ -1,4 +1,5 @@
 import {
+  DEFAULT_POLICY_EDIT_HISTORY_LIMITS,
   FileSearchLedger,
   FsLabeledScenarioStore,
   LabeledScenarioStoreError,
@@ -14,6 +15,7 @@ import {
   acquireSingleRunLock,
   analyzeCrossSurfaceInteractions,
   applySkillPatch,
+  assertPolicyEditAuthorContextBudget,
   buildAnalystSurfaceDispatch,
   callbackGovernor,
   classifyUngroundedLiterals,
@@ -33,6 +35,7 @@ import {
   isTransientTransportFailure,
   lineageNodeId,
   llmJudge,
+  llmPolicyEditProposer,
   loadEvalFixture,
   loadEvalFixtureScenarios,
   makePlaybackDispatch,
@@ -46,6 +49,7 @@ import {
   patchEditCount,
   planEvalFixtureRun,
   policyEditProposer,
+  projectPolicyEditHistory,
   renderScoreboardMarkdown,
   resolveWorktreePath,
   rolloutArgumentDiff,
@@ -57,6 +61,7 @@ import {
   scoreUserStory,
   scoreboardSummary,
   selectDiscriminative,
+  selectPolicyEditAuthorRows,
   sequentialDecide,
   sequentialPairedGate,
   skillOptEntry,
@@ -65,7 +70,7 @@ import {
   userStoryScoreboard,
   validateSearchLedgerEvent,
   verifyCodeSurface
-} from "../chunk-5NVBGKPH.js";
+} from "../chunk-GSW3OBHK.js";
 import {
   assertCodeSurfaceIdentity,
   buildEvidenceVector,
@@ -100,7 +105,7 @@ import {
   runOptimization,
   surfaceContentHash,
   surfaceHash
-} from "../chunk-N6MTC3GK.js";
+} from "../chunk-3274WNK7.js";
 import "../chunk-VI2UW6B6.js";
 import {
   fsCampaignStorage,
@@ -110,8 +115,11 @@ import {
   runCampaign,
   tangleTracesRoot
 } from "../chunk-FAOEFFRT.js";
-import "../chunk-AN5UYSVD.js";
-import "../chunk-DWLIGZBX.js";
+import "../chunk-MPHTT5HE.js";
+import {
+  POLICY_EDIT_CANDIDATE_RECORD_SCHEMA,
+  validatePolicyEditCandidateRecord
+} from "../chunk-CIUOICJT.js";
 import "../chunk-ARU2PZFM.js";
 import "../chunk-PJQFMIOX.js";
 import "../chunk-RPDDVKI7.js";
@@ -120,15 +128,17 @@ import "../chunk-LNQEP766.js";
 import "../chunk-5UF54T55.js";
 import "../chunk-XJYR7XFV.js";
 import "../chunk-VSMTAMNK.js";
-import "../chunk-FUCQVFMU.js";
+import "../chunk-GY4SYVPJ.js";
 import "../chunk-PC4UYEBM.js";
 import "../chunk-ONWEPEDO.js";
 import "../chunk-PZ5AY32C.js";
 export {
+  DEFAULT_POLICY_EDIT_HISTORY_LIMITS,
   FileSearchLedger,
   FsLabeledScenarioStore,
   LabeledScenarioStoreError,
   Lineage,
+  POLICY_EDIT_CANDIDATE_RECORD_SCHEMA,
   ProfileMatrixError,
   SEARCH_LEDGER_SCHEMA,
   SearchLedgerConflictError,
@@ -141,6 +151,7 @@ export {
   analyzeCrossSurfaceInteractions,
   applySkillPatch,
   assertCodeSurfaceIdentity,
+  assertPolicyEditAuthorContextBudget,
   buildAnalystSurfaceDispatch,
   buildEvidenceVector,
   buildLoopProvenanceRecord,
@@ -181,6 +192,7 @@ export {
   labelTrustRank,
   lineageNodeId,
   llmJudge,
+  llmPolicyEditProposer,
   loadEvalFixture,
   loadEvalFixtureScenarios,
   loopProvenanceSpans,
@@ -201,6 +213,7 @@ export {
   planEvalFixtureRun,
   policyEditProposer,
   powerPreflight,
+  projectPolicyEditHistory,
   provenanceRecordPath,
   provenanceSpansPath,
   renderScoreboardMarkdown,
@@ -219,6 +232,7 @@ export {
   scoreUserStory,
   scoreboardSummary,
   selectDiscriminative,
+  selectPolicyEditAuthorRows,
   sequentialDecide,
   sequentialPairedGate,
   skillOptEntry,
@@ -228,6 +242,7 @@ export {
   tangleTracesRoot,
   traceAnalystProposer,
   userStoryScoreboard,
+  validatePolicyEditCandidateRecord,
   validateSearchLedgerEvent,
   verifyCodeSurface
 };