@tangle-network/agent-eval 0.122.7 → 0.122.9

package/CHANGELOG.md

@@ -6,8 +6,14 @@ All notable changes to `@tangle-network/agent-eval` and its sibling `agent-eval-
 
 ## [Unreleased]
 
+### Added
+
+- `CostLedger.listPending()` exposes immutable pending paid calls and distinguishes calls that are active, late after cancellation, or interrupted by a prior process so durable workflows can reconcile exact reservations before resuming.
+- `traceAnalystProposer()` accepts an opt-in `resolvePriorFindings` callback that forwards canonical prior findings into the existing analyst registry.
+
 ### Changed
 
+- Trace-analysis actors are instructed to emit one executable JavaScript program per turn, report named turn-limit exhaustion, and preserve per-analyst failure details when a proposer produces no findings.
 - Keep one live tip per lineage track when another track branches or merges from it, and compare those track tips when building the frontier.
 - Pass track identity, operation, vision, ancestry, and proposer choice to candidate generation so independent tracks can pursue distinct strategies.
 - Route named tracks to caller-supplied proposers, with the default proposer as fallback, and make heuristic branches inherit their parent track's proposer.

package/dist/analyst/index.d.ts

@@ -819,6 +819,12 @@ interface CostCallBase {
     tags?: Record<string, string>;
     timestamp: number;
 }
+interface PendingCostCall extends CostCallBase {
+    status: 'pending';
+}
+interface PendingCostCallView extends PendingCostCall {
+    state: 'active' | 'late' | 'interrupted';
+}
 interface CostReceipt extends CostCallBase, CostUsage {
     status: 'settled';
     costUsd: number;
@@ -942,6 +948,8 @@ declare class CostLedger {
         error?: string;
     }): CostReceipt;
     list(filter?: CostLedgerFilter): CostReceipt[];
+    /** Read pending calls without exposing mutable ledger state. */
+    listPending(filter?: CostLedgerFilter): PendingCostCallView[];
     summary(filter?: CostLedgerFilter): CostLedgerSummary;
     markCompleted(count?: number): void;
     costPerCompletedTask(): number | null;
@@ -963,7 +971,7 @@ declare class CostLedger {
  * Keeping callback contracts structural lets those subpaths compose while the
  * concrete {@link CostLedger} retains its private durable state.
  */
-type CostLedgerHandle = Pick<CostLedger, Exclude<keyof CostLedger, 'waitForIdle'>> & Partial<Pick<CostLedger, 'waitForIdle'>>;
+type CostLedgerHandle = Pick<CostLedger, Exclude<keyof CostLedger, 'listPending' | 'waitForIdle'>> & Partial<Pick<CostLedger, 'listPending' | 'waitForIdle'>>;
 
 /**
  * LLM client with graceful degrade.

package/dist/analyst/index.js

@@ -9,13 +9,13 @@ import {
   diffFindings,
   emitSkillUsageFindings,
   runSemanticConceptJudge
-} from "../chunk-ULW7AATT.js";
+} from "../chunk-G2GPNLSX.js";
 import {
   behavioralAnalyst,
   buildDefaultAnalystRegistry,
   createChatClient,
   deriveEfficiencyFindings
-} from "../chunk-SFXDMFGV.js";
+} from "../chunk-VJ7T5WIO.js";
 import "../chunk-HHWE3POT.js";
 import {
   ANALYST_SEVERITIES,
@@ -71,13 +71,13 @@ import {
   validatePolicyEdit,
   validatePolicyEditCandidateRecord,
   validateUsageSettlementTimeout
-} from "../chunk-EAWKAVID.js";
+} from "../chunk-FC5NDO3E.js";
 import "../chunk-NJC7U437.js";
 import {
   CostLedger
-} from "../chunk-ZMXDQ4K7.js";
+} from "../chunk-BGVTIE2C.js";
 import "../chunk-VI2UW6B6.js";
-import "../chunk-PXD6ZFNY.js";
+import "../chunk-IR3KBHOY.js";
 import "../chunk-PC4UYEBM.js";
 import "../chunk-XJYR7XFV.js";
 import "../chunk-VSMTAMNK.js";

package/dist/benchmarks/index.js

@@ -16,19 +16,19 @@ import {
   routing_exports,
   runBenchmarkAdapter,
   summarizeBenchmarkCampaign
-} from "../chunk-WULRYLNC.js";
-import "../chunk-67H37Q6I.js";
-import "../chunk-YA6MJYZN.js";
-import "../chunk-FZ26ORA6.js";
-import "../chunk-EAWKAVID.js";
+} from "../chunk-FHFTYX2Q.js";
+import "../chunk-CWF6SCBK.js";
+import "../chunk-4EDDIK77.js";
+import "../chunk-3FCG7FBV.js";
+import "../chunk-FC5NDO3E.js";
 import "../chunk-ARU2PZFM.js";
 import "../chunk-NJC7U437.js";
 import "../chunk-PJQFMIOX.js";
-import "../chunk-ZMXDQ4K7.js";
+import "../chunk-BGVTIE2C.js";
 import "../chunk-VI2UW6B6.js";
-import "../chunk-WW2A73HW.js";
+import "../chunk-NUKSVU3W.js";
 import "../chunk-GGE4NNQT.js";
-import "../chunk-PXD6ZFNY.js";
+import "../chunk-IR3KBHOY.js";
 import "../chunk-PC4UYEBM.js";
 import "../chunk-S3UZOQ5Y.js";
 import "../chunk-MA6HLL3S.js";

package/dist/campaign/index.d.ts

@@ -57,6 +57,12 @@ interface CostCallBase {
     tags?: Record<string, string>;
     timestamp: number;
 }
+interface PendingCostCall extends CostCallBase {
+    status: 'pending';
+}
+interface PendingCostCallView extends PendingCostCall {
+    state: 'active' | 'late' | 'interrupted';
+}
 interface CostReceipt extends CostCallBase, CostUsage {
     status: 'settled';
     costUsd: number;
@@ -180,6 +186,8 @@ declare class CostLedger {
         error?: string;
     }): CostReceipt;
     list(filter?: CostLedgerFilter): CostReceipt[];
+    /** Read pending calls without exposing mutable ledger state. */
+    listPending(filter?: CostLedgerFilter): PendingCostCallView[];
     summary(filter?: CostLedgerFilter): CostLedgerSummary;
     markCompleted(count?: number): void;
     costPerCompletedTask(): number | null;
@@ -201,7 +209,7 @@ declare class CostLedger {
  * Keeping callback contracts structural lets those subpaths compose while the
  * concrete {@link CostLedger} retains its private durable state.
  */
-type CostLedgerHandle = Pick<CostLedger, Exclude<keyof CostLedger, 'waitForIdle'>> & Partial<Pick<CostLedger, 'waitForIdle'>>;
+type CostLedgerHandle = Pick<CostLedger, Exclude<keyof CostLedger, 'listPending' | 'waitForIdle'>> & Partial<Pick<CostLedger, 'listPending' | 'waitForIdle'>>;
 
 /**
  * `CampaignStorage` — the filesystem seam `runCampaign` writes through
@@ -1340,6 +1348,13 @@ interface EvidenceRef {
     uri: string;
     excerpt?: string;
 }
+/**
+ * The discriminator the registry uses to pass the right input.
+ * `custom` is the escape hatch — analysts that need something else
+ * (e.g. an embedding cache, a partner SDK handle) read it from
+ * `AnalystRunInputs.custom[<analyst id>]`.
+ */
+type AnalystInputKind = 'trace-store' | 'artifact-dir' | 'run-record' | 'judge-input' | 'custom';
 interface AnalystCost {
     /** `deterministic` analysts MUST NOT call the LLM. */
     kind: 'deterministic' | 'llm';
@@ -1350,6 +1365,12 @@ interface AnalystCost {
     /** Maximum post-cancellation wait for provider usage. Model analysts default to 5 seconds. */
     settlement_timeout_ms?: number;
 }
+interface AnalystRequirements {
+    /** Min number of shots / samples the analyst needs to produce signal. */
+    min_shots?: number;
+    /** Capabilities the runtime must supply (e.g. ['network', 'gpu']). */
+    capabilities?: string[];
+}
 interface AnalystContext {
     runId: string;
     /** Stable correlation id so logs from a single registry.run() share a tag. */
@@ -1398,6 +1419,24 @@ interface AnalystContext {
     /** Optional abort signal. Analysts SHOULD pass it through to LLM calls. */
     signal?: AbortSignal;
 }
+/**
+ * The minimal contract. Concrete analysts can refine `TInput` so
+ * implementations stay type-safe (e.g. a trace analyst's `TInput` is
+ * `TraceAnalysisStore`); the registry passes the right field from
+ * `AnalystRunInputs` based on `inputKind`.
+ */
+interface Analyst<TInput = unknown> {
+    /** Stable identifier — appears in finding_id, telemetry, and registry exclusion lists. */
+    readonly id: string;
+    /** Human-readable. One sentence. */
+    readonly description: string;
+    readonly inputKind: AnalystInputKind;
+    readonly cost: AnalystCost;
+    readonly requires?: AnalystRequirements;
+    /** Bump on breaking changes to claim wording or area so old finding_ids don't collide. */
+    readonly version: string;
+    analyze(input: TInput, ctx: AnalystContext): Promise<AnalystFinding[]>;
+}
 /** Metered work performed by one analyst call. */
 interface AnalystUsageReceipt {
     /** Number of model-usage records observed at the provider boundary. */
@@ -6053,6 +6092,76 @@ interface TraceAnalystGolden {
     expected: ReadonlyArray<Omit<CanonicalRawAnalystFinding, 'confidence'>>;
 }
 
+/**
+ * AnalystRegistry — orchestrate N analysts against one run.
+ *
+ * Owns three responsibilities and only three:
+ *   1. Registration — ids must be unique; bad registrations fail loudly
+ *      at register-time, not run-time.
+ *   2. Routing — each analyst declares its `inputKind`; the registry
+ *      picks the matching field from AnalystRunInputs and skips the
+ *      analyst with a logged reason if it's missing.
+ *   3. Isolation — one analyst's exception MUST NOT stop other analysts.
+ *      Failed analysts produce zero findings + a 'failed' summary row.
+ *
+ * Cross-cutting concerns (telemetry, error → finding conversion, cost
+ * ingestion, storage rotation) live in `AnalystHooks`. Budget shaping
+ * (equal split vs weighted vs custom) lives in `BudgetPolicy`. Both
+ * have sensible defaults; consumers override only what they need.
+ */
+
+interface BudgetPolicy {
+    /** Overall USD cap across the registry.run(). */
+    totalUsd?: number;
+    /** Per-analyst weight for the default allocator. Missing ids get weight 1. */
+    weights?: Record<string, number>;
+    /**
+     * Custom allocator — receives the analyst, remaining/total budget, and
+     * the count of analysts that will run. Returns the per-analyst budget
+     * (or undefined only when the run has no overall cap). Overrides weights
+     * when set.
+     */
+    allocate?: (args: {
+        analyst: Analyst;
+        totalUsd: number | undefined;
+        remainingUsd: number | undefined;
+        runningCount: number;
+    }) => number | undefined;
+}
+interface RegistryRunOpts {
+    /** Restrict to a subset of registered analysts by id. */
+    only?: string[];
+    /** Skip these analysts even if registered. Useful for cheap iteration. */
+    skip?: string[];
+    /** Budget policy — totalUsd + optional weights/allocator. Falls back to options.defaultBudget. */
+    budget?: BudgetPolicy;
+    /** Active-work cap for the complete registry run. Model receipt settlement may follow. */
+    timeoutMs?: number;
+    /** Abort signal — forwarded into every analyst's context. */
+    signal?: AbortSignal;
+    /** Shared paid-call account forwarded to every analyst. */
+    costLedger?: CostLedgerHandle;
+    /** Attribution phase for calls written to `costLedger`. */
+    costPhase?: string;
+    /** Tags echoed into AnalystContext.tags — useful for tracking environment/version in findings. */
+    tags?: Record<string, string>;
+    /**
+     * Prior-run findings made available as retrieval context to every
+     * analyst via `ctx.priorFindings`. The registry forwards the slice
+     * whose `analyst_id` matches each registered analyst so a kind sees
+     * only its own history. Pass `{ '*': findings }` to broadcast to
+     * every analyst (useful when several kinds share the same historical
+     * context). For findings from this run, use `chainFindings` instead.
+     */
+    priorFindings?: ReadonlyArray<AnalystFinding> | Record<string, ReadonlyArray<AnalystFinding>>;
+    /**
+     * Pass findings produced earlier in this registry run to each later analyst
+     * via `ctx.upstreamFindings`. Registration order is dependency order.
+     * Disabled by default because independent analyst suites must opt in.
+     */
+    chainFindings?: boolean;
+}
+
 /**
  * `traceAnalystProposer` — wraps agent-eval's OWN trace-analyst engine
  * (`AnalystRegistry` over the agentic OTLP reader) as a `SurfaceProposer`.
@@ -6070,7 +6179,8 @@ interface TraceAnalystGolden {
  * throw. Never fabricate a candidate.
  */
 
-interface TraceAnalystProposerOptions {
+type TraceAnalystPriorFindings = NonNullable<RegistryRunOpts['priorFindings']>;
+interface TraceAnalystProposerOptions<TFindings = unknown> {
     /** OpenAI-compatible base URL for BOTH the analyst's agentic reads and the
      *  apply step (e.g. `https://api.deepseek.com/v1` or the Tangle router). */
     baseUrl: string;
@@ -6093,15 +6203,19 @@ interface TraceAnalystProposerOptions {
     kinds?: readonly TraceAnalystKindSpec[];
     /** Resolve the OTLP traces (JSONL string) the analyst should read for THIS
      *  generation — identical contract to `haloProposer.resolveTraces`. */
-    resolveTraces: (ctx: ProposeContext) => string | Promise<string>;
+    resolveTraces: (ctx: ProposeContext<TFindings>) => string | Promise<string>;
+    /** Opt in to forwarding historical findings into every registered analyst.
+     *  Return an array to route by analyst id or `{ '*': findings }` to broadcast.
+     *  No findings are forwarded when this resolver is absent. */
+    resolvePriorFindings?: (ctx: ProposeContext<TFindings>) => TraceAnalystPriorFindings | undefined | Promise<TraceAnalystPriorFindings | undefined>;
     /** Override the findings producer. Default: the shipped `AnalystRegistry`
      *  over `kinds`. The unit suite injects canned findings here. */
-    analyze?: (tracePath: string, ctx: ProposeContext) => Promise<ReadonlyArray<AnalystFinding>>;
+    analyze?: (tracePath: string, ctx: ProposeContext<TFindings>) => Promise<ReadonlyArray<AnalystFinding>>;
     /** Test seam: inject a fetch for the apply-step `callLlm`. */
     fetchImpl?: LlmClientOptions['fetch'];
 }
 /** Wrap agent-eval's trace-analyst registry as a SurfaceProposer (prompt-tier). */
-declare function traceAnalystProposer(opts: TraceAnalystProposerOptions): SurfaceProposer;
+declare function traceAnalystProposer<TFindings = unknown>(opts: TraceAnalystProposerOptions<TFindings>): SurfaceProposer<TFindings>;
 
 /**
  * Reporting helpers — production summaries and paper-quality figures — sit alongside `reporter.ts` rather
@@ -7525,4 +7639,4 @@ 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 AxisEvidence, type AxisVerdict, type BuildAnalystSurfaceDispatchOptions, type BuildEvidenceVectorOptions, type BuildLoopProvenanceArgs, type CampaignAggregates, type CampaignArtifactWriter, type CampaignBreakdown, type CampaignCellResult, type CampaignCostMeter, type CampaignResult, type CampaignRunPlan, type CampaignRunPlanCell, type CampaignScenarioIdentity, type CampaignStorage, type CampaignTokenUsage, type CampaignTraceWriter, type CodeSurface, type CodeSurfaceVerification, type CompareProposersOptions, type CompositeProposerOptions, type CostLedgerHandle, 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 DefaultProductionGateOptions, type DimensionRegression, type DiscriminationScore, type DispatchContext, type DispatchFn, type EmitLoopProvenanceArgs, type EmitLoopProvenanceResult, type EvalFixture, type EvalFixtureFile, type EvalFixtureLoadOptions, type EvalFixtureRunPlan, type EvalFixtureScenario, type EvalFixtureValidationMode, type EvidenceVector, type EvolutionaryProposerOptions, type FailureModeRecallJudgeOptions, type FapoAttributionSignals, type FapoEntryConfig, type FapoFailureCluster, type FapoOptimizationLevel, type FapoProposerOptions, type FapoReviewInput, type FapoReviewIssue, type FapoReviewResult, type FapoScopeContract, FileSearchLedger, FsLabeledScenarioStore, type FsLabeledScenarioStoreOptions, type Gate, type GateContext, type GateDecision, type GateResult, type GenerationCandidate, type GenerationRecord, type GepaProposerConstraints, type GepaProposerOptions, type GitWorktreeAdapterOptions, type Governor, type GovernorContext, type GovernorOp, type HaloProposerOptions, type HeldOutGateOptions, type HeldoutSignificance, type HeldoutSignificanceOptions, type HeuristicGovernorOptions, type JsonPolicyEditTargetSurface, type JsonPrimitive, type JsonValue, type JudgeAggregate, type JudgeConfig, type JudgeDimension, type JudgeScore, type LabelTrust, type LabeledScenarioRecord, type LabeledScenarioSampleArgs, type LabeledScenarioSource, type LabeledScenarioStore, LabeledScenarioStoreError, type LabeledScenarioWrite, Lineage, type LineageEdge, type LineageGraph, type LineageNode, type LineageNodeInput, type LineageStore, LineageStoreConflictError, type LlmJudgeDimension, type LlmJudgeOptions, type LlmPolicyEditProposerOptions, type LoadEvalFixtureScenariosOptions, type LoopProvenanceArgsFromResult, type LoopProvenanceBackend, type LoopProvenanceCandidate, type LoopProvenanceEvidence, type LoopProvenanceRecord, type MemoryCurationProposerOptions, type MutableSurface, type Mutator, type NeutralizationGateOptions, type ObjectiveSource, type OpenAutoPrOptions, type OpenAutoPrResult, type OpenSearchLedgerOptions, type OptimizationProposer, type OptimizerConfig, type OptimizerEntryConfig, POLICY_EDIT_CANDIDATE_RECORD_SCHEMA, type PairedHoldout, type ParameterCandidate, type ParameterChange, type ParameterSweepProposerOptions, type ParetoParent, type ParetoSignificanceGateOptions, type PlanCampaignRunOptions, type PlanEvalFixtureRunOptions, type PlaybackContext, type PlaybackDriver, type PlaybackStep, type PolicyEditAuthorScenarioRow, type PolicyEditCandidateRecord, type PolicyEditCandidateSummary, type PolicyEditFindingInput, type PolicyEditFindingSource, type PolicyEditHistoryCandidateContext, type PolicyEditHistoryGenerationContext, type PolicyEditHistoryProjectionOptions, type PolicyEditObjective, type PolicyEditOutcomeContext, type PolicyEditProposerOptions, type PowerPreflight, type PowerPreflightOptions, type PremeasuredOptimizationBaseline, type ProfileDispatchFn, ProfileMatrixError, type ProfileSummary, type PromotionObjective, type PromotionPolicy, type ProposalTrackContext, type ProposeContext, type ProposePatchesArgs, type ProposedCandidate, type ProposerComparison, type ProposerEntry, type ProposerPairwise, type ProposerScore, type RedactionStatus, type ReferenceEquivalenceJudgeOptions, type ReferenceEquivalenceScenario, type RejectedEdit, type RolloutArgumentDiff, type RolloutArgumentDiffOptions, type RolloutCall, type RunCampaignOptions, type RunEvalOptions, type RunImprovementLoopOptions, type RunImprovementLoopResult, type RunLineageLoopOptions, type RunLineageLoopResult, type RunLineageLoopSeed, type RunLineageOptions, type RunLineageResult, type RunLineageSeed, type RunLineageStepResult, type RunOptimizationOptions, type RunOptimizationResult, type RunProfileMatrixOptions, type RunProfileMatrixResult, type RunSkillOptOptions, type RunSkillOptResult, SEARCH_LEDGER_SCHEMA, type Scenario, type ScenarioAggregate, type ScenarioRollup, type ScenarioSignal, type ScoreboardRenderOptions, type ScoreboardRow, type ScoreboardSummary, type ScoredRollout, type ScoredSurfaceOutcome, 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 SessionScript, type SingleRunLock, type SingleRunLockOptions, type SkillOptEpochRecord, type SkillOptEvidence, type SkillOptProposer, type SkillOptProposerOptions, type SkillPatch, type SkillPatchOp, SkillPatchParseError, type SkillPatchRejection, type SurfaceProposer, type SurfaceScore, type TraceAnalystProposerOptions, type TraceSpan, type TransientFailureOptions, type UngroundedLiteralReport, type UserStory, type UserStoryVerdict, type Worktree, type WorktreeAdapter, WorktreeAdapterError, aceProposer, acquireSingleRunLock, analyzeCrossSurfaceInteractions, applySkillPatch, assertCampaignDesign, assertCampaignSplitIdentity, assertCodeSurfaceIdentity, assertPolicyEditAuthorContextBudget, buildAnalystSurfaceDispatch, buildEvidenceVector, buildLoopProvenanceRecord, callbackGovernor, campaignBreakdown, campaignLineageStore, campaignMeanComposite, campaignMeasurementDigest, campaignScenarioIdentity, campaignSplitDigest, campaignSplitDigestFromIdentities, canonicalDigest, classifyUngroundedLiterals, codeSurfaceIdentityMaterial, compareProposers, composeGate, compositeProposer, countSentenceEdits, createReferenceEquivalenceJudge, createRunCostLedger, defaultProductionGate, detectScale, dimensionRegressions, discoverEvalFixtures, emitLoopProvenance, evolutionaryProposer, extractFapoAttributionSignals, extractH2Sections, failureModeRecallJudge, fapoEscalationEntry, fapoProposer, fsCampaignStorage, fsLineageStore, gepaParetoEntry, gepaProposer, gepaReflectionEntry, gitWorktreeAdapter, haloProposer, heldOutGate, heldoutSignificance, heuristicGovernor, inMemoryCampaignStorage, isProposedCandidate, isTransientTransportFailure, labelTrustRank, lineageNodeId, llmJudge, llmPolicyEditProposer, loadEvalFixture, loadEvalFixtureScenarios, loopProvenanceArgsFromResult, loopProvenanceSpans, makePlaybackDispatch, memLineageStore, memoryCurationProposer, neutralizationGate, neutralizeText, openAutoPr, openSearchLedger, pairHoldout, parameterSweepProposer, paretoPolicy, paretoSignificanceGate, parseSkillPatchResponse, patchEditCount, planCampaignRun, planEvalFixtureRun, policyEditProposer, powerPreflight, projectPolicyEditHistory, provenanceRecordPath, provenanceSpansPath, renderScoreboardMarkdown, renderSurfaceDiff, resolveRunDir, resolveWorktreePath, rolloutArgumentDiff, runCampaign, runEval, runImprovementLoop, runLineage, runLineageLoop, runOptimization, runProfileMatrix, runSkillOpt, scoreDiscrimination, scoreUserStory, scoreboardSummary, selectDiscriminative, selectPolicyEditAuthorRows, sequentialDecide, sequentialPairedGate, skillOptEntry, skillOptProposer, surfaceContentHash, surfaceHash, tangleTracesRoot, traceAnalystProposer, userStoryScoreboard, validatePolicyEditCandidateRecord, validateSearchLedgerEvent, verifyCodeSurface, verifyLoopProvenanceRecord };
+export { type AcceptedEdit, type AceProposerOptions, type AnalystArtifact, type AnalystScenario, type AnalyzeCrossSurfaceInteractionsInput, type ApplySkillPatchResult, type AxisEvidence, type AxisVerdict, type BuildAnalystSurfaceDispatchOptions, type BuildEvidenceVectorOptions, type BuildLoopProvenanceArgs, type CampaignAggregates, type CampaignArtifactWriter, type CampaignBreakdown, type CampaignCellResult, type CampaignCostMeter, type CampaignResult, type CampaignRunPlan, type CampaignRunPlanCell, type CampaignScenarioIdentity, type CampaignStorage, type CampaignTokenUsage, type CampaignTraceWriter, type CodeSurface, type CodeSurfaceVerification, type CompareProposersOptions, type CompositeProposerOptions, type CostLedgerHandle, 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 DefaultProductionGateOptions, type DimensionRegression, type DiscriminationScore, type DispatchContext, type DispatchFn, type EmitLoopProvenanceArgs, type EmitLoopProvenanceResult, type EvalFixture, type EvalFixtureFile, type EvalFixtureLoadOptions, type EvalFixtureRunPlan, type EvalFixtureScenario, type EvalFixtureValidationMode, type EvidenceVector, type EvolutionaryProposerOptions, type FailureModeRecallJudgeOptions, type FapoAttributionSignals, type FapoEntryConfig, type FapoFailureCluster, type FapoOptimizationLevel, type FapoProposerOptions, type FapoReviewInput, type FapoReviewIssue, type FapoReviewResult, type FapoScopeContract, FileSearchLedger, FsLabeledScenarioStore, type FsLabeledScenarioStoreOptions, type Gate, type GateContext, type GateDecision, type GateResult, type GenerationCandidate, type GenerationRecord, type GepaProposerConstraints, type GepaProposerOptions, type GitWorktreeAdapterOptions, type Governor, type GovernorContext, type GovernorOp, type HaloProposerOptions, type HeldOutGateOptions, type HeldoutSignificance, type HeldoutSignificanceOptions, type HeuristicGovernorOptions, type JsonPolicyEditTargetSurface, type JsonPrimitive, type JsonValue, type JudgeAggregate, type JudgeConfig, type JudgeDimension, type JudgeScore, type LabelTrust, type LabeledScenarioRecord, type LabeledScenarioSampleArgs, type LabeledScenarioSource, type LabeledScenarioStore, LabeledScenarioStoreError, type LabeledScenarioWrite, Lineage, type LineageEdge, type LineageGraph, type LineageNode, type LineageNodeInput, type LineageStore, LineageStoreConflictError, type LlmJudgeDimension, type LlmJudgeOptions, type LlmPolicyEditProposerOptions, type LoadEvalFixtureScenariosOptions, type LoopProvenanceArgsFromResult, type LoopProvenanceBackend, type LoopProvenanceCandidate, type LoopProvenanceEvidence, type LoopProvenanceRecord, type MemoryCurationProposerOptions, type MutableSurface, type Mutator, type NeutralizationGateOptions, type ObjectiveSource, type OpenAutoPrOptions, type OpenAutoPrResult, type OpenSearchLedgerOptions, type OptimizationProposer, type OptimizerConfig, type OptimizerEntryConfig, POLICY_EDIT_CANDIDATE_RECORD_SCHEMA, type PairedHoldout, type ParameterCandidate, type ParameterChange, type ParameterSweepProposerOptions, type ParetoParent, type ParetoSignificanceGateOptions, type PendingCostCallView, type PlanCampaignRunOptions, type PlanEvalFixtureRunOptions, type PlaybackContext, type PlaybackDriver, type PlaybackStep, type PolicyEditAuthorScenarioRow, type PolicyEditCandidateRecord, type PolicyEditCandidateSummary, type PolicyEditFindingInput, type PolicyEditFindingSource, type PolicyEditHistoryCandidateContext, type PolicyEditHistoryGenerationContext, type PolicyEditHistoryProjectionOptions, type PolicyEditObjective, type PolicyEditOutcomeContext, type PolicyEditProposerOptions, type PowerPreflight, type PowerPreflightOptions, type PremeasuredOptimizationBaseline, type ProfileDispatchFn, ProfileMatrixError, type ProfileSummary, type PromotionObjective, type PromotionPolicy, type ProposalTrackContext, type ProposeContext, type ProposePatchesArgs, type ProposedCandidate, type ProposerComparison, type ProposerEntry, type ProposerPairwise, type ProposerScore, type RedactionStatus, type ReferenceEquivalenceJudgeOptions, type ReferenceEquivalenceScenario, type RejectedEdit, type RolloutArgumentDiff, type RolloutArgumentDiffOptions, type RolloutCall, type RunCampaignOptions, type RunEvalOptions, type RunImprovementLoopOptions, type RunImprovementLoopResult, type RunLineageLoopOptions, type RunLineageLoopResult, type RunLineageLoopSeed, type RunLineageOptions, type RunLineageResult, type RunLineageSeed, type RunLineageStepResult, type RunOptimizationOptions, type RunOptimizationResult, type RunProfileMatrixOptions, type RunProfileMatrixResult, type RunSkillOptOptions, type RunSkillOptResult, SEARCH_LEDGER_SCHEMA, type Scenario, type ScenarioAggregate, type ScenarioRollup, type ScenarioSignal, type ScoreboardRenderOptions, type ScoreboardRow, type ScoreboardSummary, type ScoredRollout, type ScoredSurfaceOutcome, 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 SessionScript, type SingleRunLock, type SingleRunLockOptions, type SkillOptEpochRecord, type SkillOptEvidence, type SkillOptProposer, type SkillOptProposerOptions, type SkillPatch, type SkillPatchOp, SkillPatchParseError, type SkillPatchRejection, type SurfaceProposer, type SurfaceScore, type TraceAnalystPriorFindings, type TraceAnalystProposerOptions, type TraceSpan, type TransientFailureOptions, type UngroundedLiteralReport, type UserStory, type UserStoryVerdict, type Worktree, type WorktreeAdapter, WorktreeAdapterError, aceProposer, acquireSingleRunLock, analyzeCrossSurfaceInteractions, applySkillPatch, assertCampaignDesign, assertCampaignSplitIdentity, assertCodeSurfaceIdentity, assertPolicyEditAuthorContextBudget, buildAnalystSurfaceDispatch, buildEvidenceVector, buildLoopProvenanceRecord, callbackGovernor, campaignBreakdown, campaignLineageStore, campaignMeanComposite, campaignMeasurementDigest, campaignScenarioIdentity, campaignSplitDigest, campaignSplitDigestFromIdentities, canonicalDigest, classifyUngroundedLiterals, codeSurfaceIdentityMaterial, compareProposers, composeGate, compositeProposer, countSentenceEdits, createReferenceEquivalenceJudge, createRunCostLedger, defaultProductionGate, detectScale, dimensionRegressions, discoverEvalFixtures, emitLoopProvenance, evolutionaryProposer, extractFapoAttributionSignals, extractH2Sections, failureModeRecallJudge, fapoEscalationEntry, fapoProposer, fsCampaignStorage, fsLineageStore, gepaParetoEntry, gepaProposer, gepaReflectionEntry, gitWorktreeAdapter, haloProposer, heldOutGate, heldoutSignificance, heuristicGovernor, inMemoryCampaignStorage, isProposedCandidate, isTransientTransportFailure, labelTrustRank, lineageNodeId, llmJudge, llmPolicyEditProposer, loadEvalFixture, loadEvalFixtureScenarios, loopProvenanceArgsFromResult, loopProvenanceSpans, makePlaybackDispatch, memLineageStore, memoryCurationProposer, neutralizationGate, neutralizeText, openAutoPr, openSearchLedger, pairHoldout, parameterSweepProposer, paretoPolicy, paretoSignificanceGate, parseSkillPatchResponse, patchEditCount, planCampaignRun, planEvalFixtureRun, policyEditProposer, powerPreflight, projectPolicyEditHistory, provenanceRecordPath, provenanceSpansPath, renderScoreboardMarkdown, renderSurfaceDiff, resolveRunDir, resolveWorktreePath, rolloutArgumentDiff, runCampaign, runEval, runImprovementLoop, runLineage, runLineageLoop, runOptimization, runProfileMatrix, runSkillOpt, scoreDiscrimination, scoreUserStory, scoreboardSummary, selectDiscriminative, selectPolicyEditAuthorRows, sequentialDecide, sequentialPairedGate, skillOptEntry, skillOptProposer, surfaceContentHash, surfaceHash, tangleTracesRoot, traceAnalystProposer, userStoryScoreboard, validatePolicyEditCandidateRecord, validateSearchLedgerEvent, verifyCodeSurface, verifyLoopProvenanceRecord };

package/dist/campaign/index.js

@@ -68,7 +68,7 @@ import {
   userStoryScoreboard,
   validateSearchLedgerEvent,
   verifyCodeSurface
-} from "../chunk-67H37Q6I.js";
+} from "../chunk-CWF6SCBK.js";
 import {
   assertCodeSurfaceIdentity,
   buildEvidenceVector,
@@ -109,7 +109,7 @@ import {
   surfaceContentHash,
   surfaceHash,
   verifyLoopProvenanceRecord
-} from "../chunk-YA6MJYZN.js";
+} from "../chunk-4EDDIK77.js";
 import {
   SearchLedgerConflictError,
   SearchLedgerError,
@@ -126,19 +126,19 @@ import {
   resolveRunDir,
   runCampaign,
   tangleTracesRoot
-} from "../chunk-FZ26ORA6.js";
+} from "../chunk-3FCG7FBV.js";
 import {
   POLICY_EDIT_CANDIDATE_RECORD_SCHEMA,
   validatePolicyEditCandidateRecord
-} from "../chunk-EAWKAVID.js";
+} from "../chunk-FC5NDO3E.js";
 import "../chunk-ARU2PZFM.js";
 import "../chunk-NJC7U437.js";
 import "../chunk-PJQFMIOX.js";
-import "../chunk-ZMXDQ4K7.js";
+import "../chunk-BGVTIE2C.js";
 import "../chunk-VI2UW6B6.js";
-import "../chunk-WW2A73HW.js";
+import "../chunk-NUKSVU3W.js";
 import "../chunk-GGE4NNQT.js";
-import "../chunk-PXD6ZFNY.js";
+import "../chunk-IR3KBHOY.js";
 import "../chunk-PC4UYEBM.js";
 import "../chunk-S3UZOQ5Y.js";
 import "../chunk-MA6HLL3S.js";

package/dist/{chunk-FZ26ORA6.js → chunk-3FCG7FBV.js}

@@ -4,7 +4,7 @@ import {
 import {
   CostAccountingIncompleteError,
   CostLedger
-} from "./chunk-ZMXDQ4K7.js";
+} from "./chunk-BGVTIE2C.js";
 import {
   AgentEvalError,
   ValidationError
@@ -1378,4 +1378,4 @@ export {
   runCampaign,
   planCampaignRun
 };
-//# sourceMappingURL=chunk-FZ26ORA6.js.map
+//# sourceMappingURL=chunk-3FCG7FBV.js.map

package/dist/{chunk-YA6MJYZN.js → chunk-4EDDIK77.js}

@@ -10,12 +10,12 @@ import {
   resolveRunDir,
   runCampaign,
   summarizeAgentReceiptIntegrity
-} from "./chunk-FZ26ORA6.js";
+} from "./chunk-3FCG7FBV.js";
 import {
   clamp01,
   mapConcurrent,
   validatePolicyEditCandidateRecord
-} from "./chunk-EAWKAVID.js";
+} from "./chunk-FC5NDO3E.js";
 import {
   detectRewardHacking
 } from "./chunk-ARU2PZFM.js";
@@ -32,7 +32,7 @@ import {
 } from "./chunk-PJQFMIOX.js";
 import {
   CostLedger
-} from "./chunk-ZMXDQ4K7.js";
+} from "./chunk-BGVTIE2C.js";
 import {
   DEFAULT_REDACTION_RULES
 } from "./chunk-GGE4NNQT.js";
@@ -3005,7 +3005,7 @@ async function runImprovementLoop(opts) {
   const dispatchTimeoutMs = opts.dispatchTimeoutMs ?? DEFAULT_DISPATCH_TIMEOUT_MS;
   const optimization = await runOptimization({ ...opts, dispatchTimeoutMs, costLedger });
   const winnerIsBaseline = optimization.winnerSurfaceHash === surfaceHash(opts.baselineSurface);
-  const { runCampaign: runCampaign2 } = await import("./run-campaign-56EB2UN5.js");
+  const { runCampaign: runCampaign2 } = await import("./run-campaign-ZKR5MQMQ.js");
   const baselineOnHoldout = await runCampaign2({
     ...opts,
     costLedger,
@@ -3797,4 +3797,4 @@ export {
   provenanceSpansPath,
   emitLoopProvenance
 };
-//# sourceMappingURL=chunk-YA6MJYZN.js.map
+//# sourceMappingURL=chunk-4EDDIK77.js.map

package/dist/{chunk-JUOZI4L3.js → chunk-6WX7CBAR.js}

@@ -1,6 +1,6 @@
 import {
   summarizeBackendIntegrity
-} from "./chunk-FZ26ORA6.js";
+} from "./chunk-3FCG7FBV.js";
 import {
   paretoChart
 } from "./chunk-DPZAEKA6.js";
@@ -995,4 +995,4 @@ export {
   summarizeExecution,
   analyzeRuns
 };
-//# sourceMappingURL=chunk-JUOZI4L3.js.map
+//# sourceMappingURL=chunk-6WX7CBAR.js.map

package/dist/{chunk-ULOKLHIQ.js → chunk-A5S77LSE.js}

@@ -8,7 +8,7 @@ import {
 } from "./chunk-PXE2VKMX.js";
 import {
   analyzeTraces
-} from "./chunk-WW2A73HW.js";
+} from "./chunk-NUKSVU3W.js";
 import {
   applyToolSpanOtlpAttributes,
   compareSpanTime,
@@ -16,7 +16,7 @@ import {
   projectOtlpFlatLine,
   spanEpochMillis,
   traceSpanKindToOpenInferenceKind
-} from "./chunk-PXD6ZFNY.js";
+} from "./chunk-IR3KBHOY.js";
 import {
   defaultProviderRedactor,
   providerFromBaseUrl
@@ -1934,4 +1934,4 @@ export {
   createReplayFetch,
   iterateRawCalls
 };
-//# sourceMappingURL=chunk-ULOKLHIQ.js.map
+//# sourceMappingURL=chunk-A5S77LSE.js.map

package/dist/{chunk-ZMXDQ4K7.js → chunk-BGVTIE2C.js}

@@ -218,6 +218,13 @@ var CostLedger = class {
   list(filter) {
     return [...this.records.values()].filter((record) => record.status === "settled").filter((receipt) => matches(receipt, filter)).map(cloneReceipt);
   }
+  /** Read pending calls without exposing mutable ledger state. */
+  listPending(filter) {
+    return [...this.records.values()].filter((record) => record.status === "pending").filter((record) => matches(record, filter)).map((record) => ({
+      ...clonePendingCall(record),
+      state: this.lateCallIds.has(record.callId) ? "late" : this.activeCallIds.has(record.callId) ? "active" : "interrupted"
+    }));
+  }
   summary(filter) {
     const records = [...this.records.values()].filter((record) => matches(record, filter));
     const pending = records.filter(
@@ -748,6 +755,9 @@ function matches(record, filter) {
 }
 function cloneRecord(record) {
   if (record.status === "settled") return cloneReceipt(record);
+  return clonePendingCall(record);
+}
+function clonePendingCall(record) {
   const { tags, ...rest } = record;
   return { ...rest, ...tags ? { tags: { ...tags } } : {} };
 }
@@ -867,4 +877,4 @@ export {
   modelPriceKey,
   costForUsage
 };
-//# sourceMappingURL=chunk-ZMXDQ4K7.js.map
+//# sourceMappingURL=chunk-BGVTIE2C.js.map