@tangle-network/agent-eval 0.103.1 → 0.104.0

package/dist/{chunk-3722PKPB.js → chunk-V7HNA47Z.js}

@@ -1,6 +1,6 @@
 import {
   analyzeTraces
-} from "./chunk-2MLIEQSN.js";
+} from "./chunk-RPDDVKI7.js";
 import {
   LLM_CACHED_TOKEN_ATTR_KEYS,
   LLM_COST_ATTR_KEYS,
@@ -19,21 +19,21 @@ import {
   projectOtlpFlatLine,
   spanEpochMillis,
   traceSpanKindToOpenInferenceKind
-} from "./chunk-YBIGNSCZ.js";
+} from "./chunk-LNQEP766.js";
 import {
   defaultProviderRedactor,
   providerFromBaseUrl
 } from "./chunk-PC4UYEBM.js";
 import {
   validateRunRecord
-} from "./chunk-J5MUFXEY.js";
+} from "./chunk-VK6HBGAE.js";
 import {
   canonicalize,
   hashJson
 } from "./chunk-VSMTAMNK.js";
 import {
   ReplayError
-} from "./chunk-3BFEG2F6.js";
+} from "./chunk-ONWEPEDO.js";
 
 // src/trace-analyst/hook.ts
 var DEFAULT_QUESTION = "Summarise what happened in this run. Surface any failure modes, surprising findings, or evidence that the run's verdict is wrong.";
@@ -1954,4 +1954,4 @@ export {
   createReplayFetch,
   iterateRawCalls
 };
-//# sourceMappingURL=chunk-3722PKPB.js.map
+//# sourceMappingURL=chunk-V7HNA47Z.js.map

package/dist/{chunk-J5MUFXEY.js → chunk-VK6HBGAE.js}

@@ -1,9 +1,9 @@
 import {
   validateAgentProfileCell
-} from "./chunk-AGYDMORK.js";
+} from "./chunk-XJYR7XFV.js";
 import {
   ValidationError
-} from "./chunk-3BFEG2F6.js";
+} from "./chunk-ONWEPEDO.js";
 
 // src/run-record.ts
 var MANDATORY_TOP_LEVEL = [
@@ -246,4 +246,4 @@ export {
   roundTripRunRecord,
   modelHasSnapshot
 };
-//# sourceMappingURL=chunk-J5MUFXEY.js.map
+//# sourceMappingURL=chunk-VK6HBGAE.js.map

package/dist/{chunk-QCBB6ZIU.js → chunk-X6NIXVOD.js}

@@ -1,7 +1,7 @@
 import {
   pearsonR,
   spearmanR
-} from "./chunk-HYGRFL7C.js";
+} from "./chunk-XMSYF4A7.js";
 
 // src/meta-eval/rubric-predictive-validity.ts
 async function rubricPredictiveValidity(input) {
@@ -133,4 +133,4 @@ function makeRng(seed) {
 export {
   rubricPredictiveValidity
 };
-//# sourceMappingURL=chunk-QCBB6ZIU.js.map
+//# sourceMappingURL=chunk-X6NIXVOD.js.map

package/dist/{chunk-AGYDMORK.js → chunk-XJYR7XFV.js}

@@ -3,7 +3,7 @@ import {
 } from "./chunk-VSMTAMNK.js";
 import {
   ValidationError
-} from "./chunk-3BFEG2F6.js";
+} from "./chunk-ONWEPEDO.js";
 
 // src/agent-profile-cell.ts
 var AgentProfileCellValidationError = class extends ValidationError {
@@ -314,4 +314,4 @@ export {
   toAgentProfileJson,
   buildAgentInterfaceProfileCell
 };
-//# sourceMappingURL=chunk-AGYDMORK.js.map
+//# sourceMappingURL=chunk-XJYR7XFV.js.map

package/dist/{chunk-HYGRFL7C.js → chunk-XMSYF4A7.js}

@@ -1,6 +1,6 @@
 import {
   ValidationError
-} from "./chunk-3BFEG2F6.js";
+} from "./chunk-ONWEPEDO.js";
 
 // src/statistics.ts
 var normalizeScores = (scores) => scores;
@@ -1103,4 +1103,4 @@ export {
   eProcess,
   mulberry32
 };
-//# sourceMappingURL=chunk-HYGRFL7C.js.map
+//# sourceMappingURL=chunk-XMSYF4A7.js.map

package/dist/{chunk-66T25VWE.js → chunk-Z7FK73R5.js}

@@ -1,16 +1,16 @@
 import {
   runCanaries,
   scoreRedTeamOutput
-} from "./chunk-2B4HPFXN.js";
+} from "./chunk-ZKIL3VOD.js";
 import {
   runCampaign
-} from "./chunk-VNM52AGA.js";
+} from "./chunk-3PFZBGMR.js";
 import {
   detectRewardHacking
-} from "./chunk-LOZOZYHU.js";
+} from "./chunk-N22ZO7FV.js";
 import {
   pairedBootstrap
-} from "./chunk-HYGRFL7C.js";
+} from "./chunk-XMSYF4A7.js";
 
 // src/campaign/gates/compose.ts
 function composeGate(...gates) {
@@ -432,4 +432,4 @@ export {
   runEval,
   evolutionaryProposer
 };
-//# sourceMappingURL=chunk-66T25VWE.js.map
+//# sourceMappingURL=chunk-Z7FK73R5.js.map

package/dist/{chunk-2B4HPFXN.js → chunk-ZKIL3VOD.js}

@@ -1,16 +1,16 @@
 import {
   runCampaign,
   summarizeBackendIntegrity
-} from "./chunk-VNM52AGA.js";
+} from "./chunk-3PFZBGMR.js";
 import {
   callLlm
-} from "./chunk-CWNP4DV4.js";
+} from "./chunk-FUCQVFMU.js";
 import {
   DEFAULT_REDACTION_RULES
 } from "./chunk-GGE4NNQT.js";
 import {
   ValidationError
-} from "./chunk-3BFEG2F6.js";
+} from "./chunk-ONWEPEDO.js";
 
 // src/pareto.ts
 function dominates(a, b, objectives) {
@@ -1460,7 +1460,7 @@ async function runImprovementLoop(opts) {
   const dispatchTimeoutMs = opts.dispatchTimeoutMs ?? DEFAULT_DISPATCH_TIMEOUT_MS;
   const optimization = await runOptimization({ ...opts, dispatchTimeoutMs });
   const winnerIsBaseline = optimization.winnerSurfaceHash === surfaceHash(opts.baselineSurface);
-  const { runCampaign: runCampaign2 } = await import("./run-campaign-DAHKO5CT.js");
+  const { runCampaign: runCampaign2 } = await import("./run-campaign-PNHAJIT2.js");
   const baselineOnHoldout = await runCampaign2({
     ...opts,
     dispatchTimeoutMs,
@@ -2029,4 +2029,4 @@ export {
   provenanceSpansPath,
   emitLoopProvenance
 };
-//# sourceMappingURL=chunk-2B4HPFXN.js.map
+//# sourceMappingURL=chunk-ZKIL3VOD.js.map

package/dist/cli.js

@@ -5,10 +5,10 @@ import {
   runRpcBatch,
   runRpcOnce,
   startServer
-} from "./chunk-D5KBVULY.js";
-import "./chunk-CWNP4DV4.js";
+} from "./chunk-DRPIZQIT.js";
+import "./chunk-FUCQVFMU.js";
 import "./chunk-PC4UYEBM.js";
-import "./chunk-3BFEG2F6.js";
+import "./chunk-ONWEPEDO.js";
 import "./chunk-PZ5AY32C.js";
 
 // src/cli.ts

package/dist/{code-agent-session-Cen-qD2y.d.ts → code-agent-session-rnJKlqmT.d.ts}

@@ -1,4 +1,4 @@
-import { b as RunSplitTag, R as RunRecord } from './run-record-CPfd1ARZ.js';
+import { b as RunSplitTag, R as RunRecord } from './run-record-I-Z3JNvO.js';
 
 type CodeAgentSessionSource = 'codex' | 'claude-code' | 'opencode' | 'kimi-code' | 'pi';
 interface ParsedCodeAgentJsonl {

package/dist/contract/index.d.ts

@@ -1,22 +1,22 @@
-import { S as Scenario, M as MutableSurface, b as DispatchContext, a as JudgeConfig, f as SurfaceProposer, g as Gate, L as LabeledScenarioStore, C as CampaignResult, j as GateDecision } from '../types-Bihq6-a3.js';
-export { k as CampaignAggregates, l as CampaignArtifactWriter, m as CampaignCellResult, n as CampaignCostMeter, d as CampaignTraceWriter, o as CodeSurface, D as Dispatch, h as GateContext, G as GateResult, p as GenerationCandidate, e as GenerationRecord, c as JudgeDimension, J as JudgeScore, i as Mutator, O as OptimizationProposer, q as OptimizerConfig, r as SessionScript } from '../types-Bihq6-a3.js';
-import { L as LoopProvenanceRecord, R as RunEvalOptions } from '../provenance-B2VsP0jP.js';
-export { A as AxisEvidence, a as AxisVerdict, B as BuildEvidenceVectorOptions, D as DefaultProductionGateOptions, E as EvidenceVector, b as EvolutionaryProposerOptions, H as HeldOutGateOptions, O as ObjectiveSource, P as ParetoSignificanceGateOptions, c as PromotionObjective, d as PromotionPolicy, e as buildEvidenceVector, f as composeGate, g as defaultProductionGate, h as evolutionaryProposer, i as heldOutGate, p as paretoPolicy, j as paretoSignificanceGate, r as runEval } from '../provenance-B2VsP0jP.js';
-import { C as CampaignStorage, a as RunOptimizationOptions, b as RunImprovementLoopResult } from '../gepa-COlCAkHN.js';
-export { G as GepaProposerOptions, R as RunCampaignOptions, c as RunImprovementLoopOptions, f as fsCampaignStorage, g as gepaProposer, i as inMemoryCampaignStorage, r as runCampaign, d as runImprovementLoop } from '../gepa-COlCAkHN.js';
+import { S as Scenario, M as MutableSurface, b as DispatchContext, a as JudgeConfig, f as SurfaceProposer, g as Gate, L as LabeledScenarioStore, C as CampaignResult, j as GateDecision } from '../types-Mh6I44ub.js';
+export { k as CampaignAggregates, l as CampaignArtifactWriter, m as CampaignCellResult, n as CampaignCostMeter, d as CampaignTraceWriter, o as CodeSurface, D as Dispatch, h as GateContext, G as GateResult, p as GenerationCandidate, e as GenerationRecord, c as JudgeDimension, J as JudgeScore, i as Mutator, O as OptimizationProposer, q as OptimizerConfig, r as SessionScript } from '../types-Mh6I44ub.js';
+import { L as LoopProvenanceRecord, R as RunEvalOptions } from '../provenance-phkRBl4h.js';
+export { A as AxisEvidence, a as AxisVerdict, B as BuildEvidenceVectorOptions, D as DefaultProductionGateOptions, E as EvidenceVector, b as EvolutionaryProposerOptions, H as HeldOutGateOptions, O as ObjectiveSource, P as ParetoSignificanceGateOptions, c as PromotionObjective, d as PromotionPolicy, e as buildEvidenceVector, f as composeGate, g as defaultProductionGate, h as evolutionaryProposer, i as heldOutGate, p as paretoPolicy, j as paretoSignificanceGate, r as runEval } from '../provenance-phkRBl4h.js';
+import { C as CampaignStorage, a as RunOptimizationOptions, b as RunImprovementLoopResult } from '../gepa-Dn0gtj9n.js';
+export { G as GepaProposerOptions, R as RunCampaignOptions, c as RunImprovementLoopOptions, f as fsCampaignStorage, g as gepaProposer, i as inMemoryCampaignStorage, r as runCampaign, d as runImprovementLoop } from '../gepa-Dn0gtj9n.js';
 export { D as DeploymentOutcome, F as FileSystemOutcomeStore, a as FileSystemOutcomeStoreOptions, I as InMemoryOutcomeStore, b as OutcomeStore } from '../outcome-store-rnXLEqSn.js';
 import { HostedTenant, EvalRunCellScore, EvalRunGenerationSnapshot, EvalRunEvent, TraceSpanEvent } from '../hosted/index.js';
-import { R as RunRecord, b as RunSplitTag } from '../run-record-CPfd1ARZ.js';
-import { I as InsightReport } from '../insight-report-DumCfEur.js';
-export { F as FailureClusterInsight, a as InterRaterInsight, J as JudgeInsight, L as LiftInsight, O as OutcomeCorrelationInsight, R as Recommendation, b as ReleaseSummary, S as ScalarDistribution } from '../insight-report-DumCfEur.js';
-export { D as DefaultAnalystRegistryOptions, c as buildDefaultAnalystRegistry } from '../default-registry-Ez4cxuZ4.js';
-export { A as AnalystFinding } from '../types-DA9yj-Jd.js';
-import { A as AnalyzeRunsOptions } from '../analyze-runs-rF2eGxZ_.js';
-export { a as analyzeRuns } from '../analyze-runs-rF2eGxZ_.js';
-export { C as CodeAgentSessionDiagnostic, a as CodeAgentSessionIntakeOptions, b as CodeAgentSessionIntakeResult, c as CodeAgentSessionMetrics, d as CodeAgentSessionSource, P as ParsedCodeAgentJsonl, f as fromClaudeCodeSession, e as fromCodexSession, g as fromKimiCodeSession, h as fromOpenCodeSession, i as fromPiSession, j as fromPigraphSession, p as parseCodeAgentJsonl } from '../code-agent-session-Cen-qD2y.js';
-import '../red-team-BWdoyleI.js';
-import '../dataset-BbGkaN2I.js';
-import '../errors-CzMUYo7b.js';
+import { R as RunRecord, b as RunSplitTag } from '../run-record-I-Z3JNvO.js';
+import { I as InsightReport } from '../insight-report-B4xrdwEK.js';
+export { F as FailureClusterInsight, a as InterRaterInsight, J as JudgeInsight, L as LiftInsight, O as OutcomeCorrelationInsight, R as Recommendation, b as ReleaseSummary, S as ScalarDistribution } from '../insight-report-B4xrdwEK.js';
+export { D as DefaultAnalystRegistryOptions, c as buildDefaultAnalystRegistry } from '../default-registry-BswHCXnU.js';
+export { A as AnalystFinding } from '../types-D1ytG0Yg.js';
+import { A as AnalyzeRunsOptions } from '../analyze-runs-DJYpep3L.js';
+export { a as analyzeRuns } from '../analyze-runs-DJYpep3L.js';
+export { C as CodeAgentSessionDiagnostic, a as CodeAgentSessionIntakeOptions, b as CodeAgentSessionIntakeResult, c as CodeAgentSessionMetrics, d as CodeAgentSessionSource, P as ParsedCodeAgentJsonl, f as fromClaudeCodeSession, e as fromCodexSession, g as fromKimiCodeSession, h as fromOpenCodeSession, i as fromPiSession, j as fromPigraphSession, p as parseCodeAgentJsonl } from '../code-agent-session-rnJKlqmT.js';
+import '../red-team-KmmiqBlY.js';
+import '../dataset-DS7ytHZU.js';
+import '../errors-oeQrLqXC.js';
 import '../store-BcFXE6LG.js';
 import '../schema-m0gsnbt3.js';
 import '../pareto-E-pembql.js';
@@ -24,13 +24,13 @@ import '../statistics-D88peojY.js';
 import '../judge-calibration-7C-IDmKr.js';
 import '../types-C7DGg5ex.js';
 import '@tangle-network/tcloud';
-import '../llm-client-Bj7g0rqu.js';
+import '../llm-client-DyqEH4jH.js';
 import '../raw-provider-sink-C46HDghv.js';
 import '@tangle-network/agent-interface';
-import '../summary-report-Fc_YFJat.js';
+import '../summary-report-QMZVe3P-.js';
 import '../failure-cluster-DH9Flgcf.js';
 import '@ax-llm/ax';
-import '../kind-factory-BLHxwX71.js';
+import '../kind-factory-DvIGo_cP.js';
 import '../store-C1YxJDEK.js';
 import 'zod';
 

package/dist/contract/index.js

@@ -18,36 +18,36 @@ import {
   paretoPolicy,
   paretoSignificanceGate,
   runEval
-} from "../chunk-66T25VWE.js";
+} from "../chunk-Z7FK73R5.js";
 import {
   analyzeRuns
-} from "../chunk-TCPP4Z5Q.js";
+} from "../chunk-GDZAWO2I.js";
 import {
   emitLoopProvenance,
   gepaProposer,
   heldOutGate,
   runImprovementLoop,
   surfaceContentHash
-} from "../chunk-2B4HPFXN.js";
+} from "../chunk-ZKIL3VOD.js";
 import {
   fsCampaignStorage,
   inMemoryCampaignStorage,
   runCampaign
-} from "../chunk-VNM52AGA.js";
+} from "../chunk-3PFZBGMR.js";
 import "../chunk-VI2UW6B6.js";
 import {
   buildDefaultAnalystRegistry
-} from "../chunk-L5TVEZFT.js";
-import "../chunk-FRI6RG3P.js";
-import "../chunk-LOZOZYHU.js";
+} from "../chunk-QRVS7MX4.js";
+import "../chunk-2OGPXHOB.js";
+import "../chunk-N22ZO7FV.js";
 import {
   FileSystemOutcomeStore,
   InMemoryOutcomeStore
 } from "../chunk-3RF76KTD.js";
-import "../chunk-CWNP4DV4.js";
-import "../chunk-CLS3374R.js";
+import "../chunk-FUCQVFMU.js";
+import "../chunk-6MFFKPZ4.js";
 import "../chunk-45EEMHTC.js";
-import "../chunk-HYGRFL7C.js";
+import "../chunk-XMSYF4A7.js";
 import "../chunk-GGE4NNQT.js";
 import "../chunk-JZXGWLK5.js";
 import "../chunk-5BKGXME7.js";
@@ -55,14 +55,14 @@ import {
   LLM_INPUT_TOKEN_ATTR_KEYS,
   LLM_MODEL_ATTR_KEYS,
   LLM_OUTPUT_TOKEN_ATTR_KEYS
-} from "../chunk-YBIGNSCZ.js";
+} from "../chunk-LNQEP766.js";
 import "../chunk-PC4UYEBM.js";
 import {
   parseRunRecordSafe
-} from "../chunk-J5MUFXEY.js";
-import "../chunk-AGYDMORK.js";
+} from "../chunk-VK6HBGAE.js";
+import "../chunk-XJYR7XFV.js";
 import "../chunk-VSMTAMNK.js";
-import "../chunk-3BFEG2F6.js";
+import "../chunk-ONWEPEDO.js";
 import "../chunk-PZ5AY32C.js";
 
 // src/contract/self-improve.ts

package/dist/{control-KofK3gfG.d.ts → control-B8UthSBL.d.ts}

@@ -1,9 +1,9 @@
-import { b as FeedbackLabel, p as ProposedSideEffect } from './feedback-trajectory-BxY0cKfs.js';
+import { b as FeedbackLabel, p as ProposedSideEffect } from './feedback-trajectory-C9KCo8ag.js';
 import { C as ControlEvalResult, a as ControlRunResult, h as ControlRuntimeConfig } from './control-runtime-Acf9CGhw.js';
 import { T as TraceEmitter } from './emitter-C2rqGH_l.js';
 import { F as FailureClass } from './schema-m0gsnbt3.js';
 import { T as TraceStore } from './store-BcFXE6LG.js';
-import { b as RunSplitTag, c as RunTokenUsage, R as RunRecord } from './run-record-CPfd1ARZ.js';
+import { b as RunSplitTag, c as RunTokenUsage, R as RunRecord } from './run-record-I-Z3JNvO.js';
 
 interface ActionExecutionPolicy {
     allowedTypes?: string[];

package/dist/control.d.ts

@@ -1,10 +1,10 @@
-export { A as ActionExecutionPolicy, a as ActionPolicyDecision, C as ControlRunToRunRecordOptions, e as ProposeReviewConfig, f as ProposeReviewControlAction, g as ProposeReviewControlConfig, h as ProposeReviewControlResult, i as ProposeReviewControlState, j as ProposeReviewReport, p as RunEvidenceMetadata, s as controlRunToRunRecord, u as evaluateActionPolicy, x as runProposeReview, y as runProposeReviewAsControlLoop, z as scoreFromEvals } from './control-KofK3gfG.js';
+export { A as ActionExecutionPolicy, a as ActionPolicyDecision, C as ControlRunToRunRecordOptions, e as ProposeReviewConfig, f as ProposeReviewControlAction, g as ProposeReviewControlConfig, h as ProposeReviewControlResult, i as ProposeReviewControlState, j as ProposeReviewReport, p as RunEvidenceMetadata, s as controlRunToRunRecord, u as evaluateActionPolicy, x as runProposeReview, y as runProposeReviewAsControlLoop, z as scoreFromEvals } from './control-B8UthSBL.js';
 export { c as ControlActionFailureMode, d as ControlActionOutcome, e as ControlBudget, f as ControlContext, g as ControlDecision, C as ControlEvalResult, a as ControlRunResult, h as ControlRuntimeConfig, i as ControlRuntimeError, j as ControlSeverity, b as ControlStep, k as ControlStopPolicies, S as StopDecision, l as allCriticalPassed, o as objectiveEval, r as runAgentControlLoop, s as stopOnNoProgress, m as stopOnRepeatedAction, n as subjectiveEval } from './control-runtime-Acf9CGhw.js';
-import './feedback-trajectory-BxY0cKfs.js';
-import './dataset-BbGkaN2I.js';
-import './errors-CzMUYo7b.js';
+import './feedback-trajectory-C9KCo8ag.js';
+import './dataset-DS7ytHZU.js';
+import './errors-oeQrLqXC.js';
 import './emitter-C2rqGH_l.js';
 import './schema-m0gsnbt3.js';
 import './store-BcFXE6LG.js';
-import './run-record-CPfd1ARZ.js';
+import './run-record-I-Z3JNvO.js';
 import '@tangle-network/agent-interface';

package/dist/control.js

@@ -4,7 +4,7 @@ import {
   runProposeReview,
   runProposeReviewAsControlLoop,
   scoreFromEvals
-} from "./chunk-BHCFJGL4.js";
+} from "./chunk-K7QEIHHJ.js";
 import {
   allCriticalPassed,
   objectiveEval,
@@ -13,11 +13,11 @@ import {
   stopOnRepeatedAction,
   subjectiveEval
 } from "./chunk-YEHAEDUD.js";
-import "./chunk-J5MUFXEY.js";
+import "./chunk-VK6HBGAE.js";
 import "./chunk-TVVP3ZZQ.js";
-import "./chunk-AGYDMORK.js";
+import "./chunk-XJYR7XFV.js";
 import "./chunk-VSMTAMNK.js";
-import "./chunk-3BFEG2F6.js";
+import "./chunk-ONWEPEDO.js";
 import "./chunk-PZ5AY32C.js";
 export {
   allCriticalPassed,

package/dist/{corpus-CysLCiK4.d.ts → corpus-eBVwhCp1.d.ts}

@@ -1,4 +1,4 @@
-import { R as RunRecord, b as RunSplitTag } from './run-record-CPfd1ARZ.js';
+import { R as RunRecord, b as RunSplitTag } from './run-record-I-Z3JNvO.js';
 import { S as Span } from './schema-m0gsnbt3.js';
 import { T as TraceStore } from './store-BcFXE6LG.js';
 

package/dist/{dataset-BbGkaN2I.d.ts → dataset-DS7ytHZU.d.ts}

@@ -1,4 +1,4 @@
-import { V as ValidationError } from './errors-CzMUYo7b.js';
+import { V as ValidationError } from './errors-oeQrLqXC.js';
 
 /**
  * Dataset — versioned, sliceable, content-hashed scenario collection.

package/dist/{default-registry-Ez4cxuZ4.d.ts → default-registry-BswHCXnU.d.ts}

@@ -1,6 +1,6 @@
 import { AxAIService } from '@ax-llm/ax';
-import { T as TraceAnalystKindSpec } from './kind-factory-BLHxwX71.js';
-import { a as Analyst, b as AnalystContext, c as AnalystRunSummary, A as AnalystFinding, d as AnalystRunResult, C as ChatClient, e as AnalystRunInputs, f as AnalystRunEvent } from './types-DA9yj-Jd.js';
+import { T as TraceAnalystKindSpec } from './kind-factory-DvIGo_cP.js';
+import { a as Analyst, b as AnalystContext, c as AnalystRunSummary, A as AnalystFinding, d as AnalystRunResult, C as ChatClient, e as AnalystRunInputs, f as AnalystRunEvent } from './types-D1ytG0Yg.js';
 
 /**
  * AnalystRegistry — orchestrate N analysts against one run.

package/dist/diagnose.d.ts

@@ -3,15 +3,15 @@ export { c as CounterfactualContext, d as CounterfactualResult } from './counter
 import { S as Span } from './schema-m0gsnbt3.js';
 import { T as TraceStore } from './store-BcFXE6LG.js';
 import { a as TrajectoryStep, T as Trajectory } from './trajectory-2TkpSEVh.js';
-import { h as AnalystSeverity, A as AnalystFinding } from './types-DA9yj-Jd.js';
-import { C as CorpusRecord } from './corpus-CysLCiK4.js';
-import { R as RunRecord } from './run-record-CPfd1ARZ.js';
+import { h as AnalystSeverity, A as AnalystFinding } from './types-D1ytG0Yg.js';
+import { C as CorpusRecord } from './corpus-eBVwhCp1.js';
+import { R as RunRecord } from './run-record-I-Z3JNvO.js';
 import './emitter-C2rqGH_l.js';
 import './store-C1YxJDEK.js';
 import './types-C7DGg5ex.js';
 import '@tangle-network/tcloud';
-import './llm-client-Bj7g0rqu.js';
-import './errors-CzMUYo7b.js';
+import './llm-client-DyqEH4jH.js';
+import './errors-oeQrLqXC.js';
 import './raw-provider-sink-C46HDghv.js';
 import '@tangle-network/agent-interface';
 

package/dist/diagnose.js

@@ -1,7 +1,7 @@
 import {
   attributeCounterfactuals,
   runCounterfactual
-} from "./chunk-REVYNR6C.js";
+} from "./chunk-6SK5VFYK.js";
 import {
   buildTrajectory
 } from "./chunk-RZTMDUO7.js";
@@ -10,16 +10,16 @@ import {
 } from "./chunk-45EEMHTC.js";
 import {
   confidenceInterval
-} from "./chunk-HYGRFL7C.js";
+} from "./chunk-XMSYF4A7.js";
 import {
   validateRunRecord
-} from "./chunk-J5MUFXEY.js";
+} from "./chunk-VK6HBGAE.js";
 import "./chunk-TVVP3ZZQ.js";
-import "./chunk-AGYDMORK.js";
+import "./chunk-XJYR7XFV.js";
 import "./chunk-VSMTAMNK.js";
 import {
   ValidationError
-} from "./chunk-3BFEG2F6.js";
+} from "./chunk-ONWEPEDO.js";
 import "./chunk-PZ5AY32C.js";
 
 // src/diagnose/causal-sweep.ts

package/dist/{errors-CzMUYo7b.d.ts → errors-oeQrLqXC.d.ts}

@@ -13,6 +13,10 @@
  * not consumer-catchable contract failures.
  */
 type AgentEvalErrorCode = 'validation' | 'not_found' | 'config' | 'capture_integrity' | 'judge' | 'verification' | 'replay' | 'backend_integrity' | 'profile_matrix';
+/**
+ * Base class for every contract error this package throws — carries the stable
+ * string `code` taxonomy so consumers can `instanceof`-match or switch on `code`.
+ */
 declare class AgentEvalError extends Error {
     /** Stable string code. Survives minification; safe to switch on. */
     readonly code: AgentEvalErrorCode;

package/dist/{feedback-trajectory-BxY0cKfs.d.ts → feedback-trajectory-C9KCo8ag.d.ts}

@@ -1,5 +1,5 @@
 import { C as ControlEvalResult, a as ControlRunResult, b as ControlStep } from './control-runtime-Acf9CGhw.js';
-import { D as DatasetSplit, a as DatasetScenario } from './dataset-BbGkaN2I.js';
+import { D as DatasetSplit, a as DatasetScenario } from './dataset-DS7ytHZU.js';
 
 type FeedbackArtifactType = 'text' | 'code' | 'plan' | 'research' | 'action' | 'ui' | 'decision' | 'data' | 'other';
 type FeedbackLabelSource = 'user' | 'judge' | 'environment' | 'metric' | 'policy' | 'system';

package/dist/fuzz.js

@@ -3,7 +3,7 @@ import {
 } from "./chunk-VZSRQ272.js";
 import {
   ValidationError
-} from "./chunk-3BFEG2F6.js";
+} from "./chunk-ONWEPEDO.js";
 import "./chunk-PZ5AY32C.js";
 
 // src/fuzz/cube.ts

package/dist/{gepa-COlCAkHN.d.ts → gepa-Dn0gtj9n.d.ts}

@@ -1,5 +1,5 @@
-import { S as Scenario, C as CampaignResult, G as GateResult, D as DispatchFn, a as JudgeConfig, L as LabeledScenarioStore, d as CampaignTraceWriter, e as GenerationRecord, M as MutableSurface, P as ParetoParent, f as SurfaceProposer, g as Gate } from './types-Bihq6-a3.js';
-import { L as LlmClientOptions } from './llm-client-Bj7g0rqu.js';
+import { S as Scenario, C as CampaignResult, G as GateResult, D as DispatchFn, a as JudgeConfig, L as LabeledScenarioStore, d as CampaignTraceWriter, e as GenerationRecord, M as MutableSurface, P as ParetoParent, f as SurfaceProposer, g as Gate } from './types-Mh6I44ub.js';
+import { L as LlmClientOptions } from './llm-client-DyqEH4jH.js';
 
 /**
  * `openAutoPr` — thin shell-out helper for the `runImprovementLoop` preset's
@@ -219,6 +219,10 @@ interface PlanCampaignRunOptions<TScenario extends Scenario, TArtifact> {
     repo?: string;
     storage?: CampaignStorage;
 }
+/**
+ * Plan a campaign WITHOUT dispatching: computes the manifest hash and the per-cell
+ * run-vs-cached schedule so callers can preview cost and resumability before spending.
+ */
 declare function planCampaignRun<TScenario extends Scenario, TArtifact>(opts: PlanCampaignRunOptions<TScenario, TArtifact>): CampaignRunPlan;
 
 /**

package/dist/governance/index.d.ts

@@ -1,9 +1,9 @@
-import { c as DatasetManifest } from '../dataset-BbGkaN2I.js';
+import { c as DatasetManifest } from '../dataset-DS7ytHZU.js';
 import { a as CalibrationResult } from '../judge-calibration-7C-IDmKr.js';
 import { b as OutcomeStore } from '../outcome-store-rnXLEqSn.js';
-import { d as RedTeamReport } from '../red-team-BWdoyleI.js';
+import { d as RedTeamReport } from '../red-team-KmmiqBlY.js';
 import { T as TraceStore } from '../store-BcFXE6LG.js';
-import '../errors-CzMUYo7b.js';
+import '../errors-oeQrLqXC.js';
 import '../schema-m0gsnbt3.js';
 
 /**

package/dist/hosted/index.d.ts

@@ -1,10 +1,10 @@
-import { M as MutableSurface, j as GateDecision } from '../types-Bihq6-a3.js';
-import { I as InsightReport } from '../insight-report-DumCfEur.js';
-import '../run-record-CPfd1ARZ.js';
+import { M as MutableSurface, j as GateDecision } from '../types-Mh6I44ub.js';
+import { I as InsightReport } from '../insight-report-B4xrdwEK.js';
+import '../run-record-I-Z3JNvO.js';
 import '@tangle-network/agent-interface';
-import '../errors-CzMUYo7b.js';
+import '../errors-oeQrLqXC.js';
 import '../schema-m0gsnbt3.js';
-import '../summary-report-Fc_YFJat.js';
+import '../summary-report-QMZVe3P-.js';
 import '../failure-cluster-DH9Flgcf.js';
 import '../store-BcFXE6LG.js';
 import '../judge-calibration-7C-IDmKr.js';

package/dist/{index-unCSYRJJ.d.ts → index-C2CC_dry.d.ts}

@@ -1,4 +1,4 @@
-import { b as RunSplitTag } from './run-record-CPfd1ARZ.js';
+import { b as RunSplitTag } from './run-record-I-Z3JNvO.js';
 
 /**
  * Shared types for the reference benchmark wrappers under