@tangle-network/agent-eval 0.93.0 → 0.94.0

package/dist/adapters/otel.d.ts

@@ -3,10 +3,10 @@ import '../types-BU-7W85F.js';
 import '../run-record-e7vj1uZQ.js';
 import '../errors-CzMUYo7b.js';
 import '../schema-m0gsnbt3.js';
-import '../insight-report-BBwvOh6x.js';
-import '../summary-report-DGmUucwQ.js';
-import '../failure-cluster-CL7IVgkJ.js';
-import '../store-CKUAgsJz.js';
+import '../insight-report-DWl3z9tl.js';
+import '../summary-report-BDOFevaT.js';
+import '../failure-cluster-DH9Flgcf.js';
+import '../store-BcFXE6LG.js';
 import '../judge-calibration-0p2QcWNE.js';
 
 /**

package/dist/analyst/index.d.ts

@@ -1,20 +1,20 @@
 import { AxAIService, AxFunction } from '@ax-llm/ax';
 import { M as MultiLayerVerifier, V as VerifyOptions, S as Severity } from '../multi-layer-verifier-DUZXrPDA.js';
-import { c as RunCritic, a as RunTrace } from '../run-critic-BAIjX99r.js';
-import { S as SemanticConceptJudgeOptions, a as SemanticConceptJudgeInput, B as BehavioralMetrics } from '../semantic-concept-judge-Dn8Z6KEG.js';
-export { C as CreateAnalystAiConfig, D as DEFAULT_TRACE_ANALYST_KINDS, b as DiffPolicy, F as FAILURE_MODE_KIND_SPEC, c as FINDING_SUBJECT_GRAMMAR_PROMPT, d as FINDING_SUBJECT_KINDS, e as FindingSubject, f as FindingSubjectKind, g as FindingSubjectStringSchema, h as FindingsDiff, i as FindingsStore, I as IMPROVEMENT_KIND_SPEC, K as KIND_EXPECTED_SUBJECTS, j as KNOWLEDGE_GAP_KIND_SPEC, k as KNOWLEDGE_POISONING_KIND_SPEC, P as PersistedFinding, l as SKILL_USAGE_ANALYST, m as SkillUsageAnalyst, n as SkillUsageRecord, o as SkillUsageReport, p as SkillUsageScanConfig, q as buildSkillUsageReport, r as createAnalystAi, s as defaultIsMaterial, t as diffFindings, u as emitSkillUsageFindings, v as parseFindingSubject, w as renderFindingSubject } from '../semantic-concept-judge-Dn8Z6KEG.js';
+import { c as RunCritic, a as RunTrace } from '../run-critic-CmMf05uV.js';
+import { S as SemanticConceptJudgeOptions, a as SemanticConceptJudgeInput, B as BehavioralMetrics } from '../semantic-concept-judge-B9MgmBnM.js';
+export { C as CreateAnalystAiConfig, D as DEFAULT_TRACE_ANALYST_KINDS, b as DiffPolicy, F as FAILURE_MODE_KIND_SPEC, c as FINDING_SUBJECT_GRAMMAR_PROMPT, d as FINDING_SUBJECT_KINDS, e as FindingSubject, f as FindingSubjectKind, g as FindingSubjectStringSchema, h as FindingsDiff, i as FindingsStore, I as IMPROVEMENT_KIND_SPEC, K as KIND_EXPECTED_SUBJECTS, j as KNOWLEDGE_GAP_KIND_SPEC, k as KNOWLEDGE_POISONING_KIND_SPEC, P as PersistedFinding, l as SKILL_USAGE_ANALYST, m as SkillUsageAnalyst, n as SkillUsageRecord, o as SkillUsageReport, p as SkillUsageScanConfig, q as buildSkillUsageReport, r as createAnalystAi, s as defaultIsMaterial, t as diffFindings, u as emitSkillUsageFindings, v as parseFindingSubject, w as renderFindingSubject } from '../semantic-concept-judge-B9MgmBnM.js';
 import { A as AnalyzeTracesOptions } from '../analyst-C8HHvfJp.js';
 import { T as TraceAnalysisStore } from '../store-C1YxJDEK.js';
-import { b as JudgeFn, a as JudgeInput } from '../types-Croy5h7V.js';
-import { A as Analyst, h as AnalystSeverity, c as AnalystFinding } from '../types-2VVIL04s.js';
-export { a as AnalystContext, g as AnalystCost, i as AnalystInputKind, j as AnalystRequirements, f as AnalystRunEvent, e as AnalystRunInputs, d as AnalystRunResult, b as AnalystRunSummary, k as ChatCallOpts, C as ChatClient, l as ChatRequest, m as ChatResponse, n as ChatTransport, o as CliBridgeTransportOpts, p as CreateChatClientOpts, D as DirectProviderTransportOpts, E as EvidenceRef, M as MockTransportOpts, R as RouterTransportOpts, S as SandboxSdkTransportOpts, q as computeFindingId, r as createChatClient, s as makeFinding } from '../types-2VVIL04s.js';
+import { b as JudgeFn, a as JudgeInput } from '../types-C7DGg5ex.js';
+import { A as Analyst, h as AnalystSeverity, c as AnalystFinding } from '../types-Ce17tDlG.js';
+export { a as AnalystContext, g as AnalystCost, i as AnalystInputKind, j as AnalystRequirements, f as AnalystRunEvent, e as AnalystRunInputs, d as AnalystRunResult, b as AnalystRunSummary, k as ChatCallOpts, C as ChatClient, l as ChatRequest, m as ChatResponse, n as ChatTransport, o as CliBridgeTransportOpts, p as CreateChatClientOpts, D as DirectProviderTransportOpts, E as EvidenceRef, M as MockTransportOpts, R as RouterTransportOpts, S as SandboxSdkTransportOpts, q as computeFindingId, r as createChatClient, s as makeFinding } from '../types-Ce17tDlG.js';
 import { TCloud } from '@tangle-network/tcloud';
-export { a as AnalystHooks, A as AnalystRegistry, c as AnalystRegistryOptions, B as BudgetPolicy, D as DefaultAnalystRegistryOptions, R as RegistryRunOpts, b as buildDefaultAnalystRegistry } from '../default-registry-zoGHUQEH.js';
-export { A as ANALYST_SEVERITIES, C as CreateTraceAnalystKindOpts, R as RAW_FINDING_SCHEMA_PROMPT, a as RawAnalystFinding, b as RawAnalystFindingSchema, c as TraceAnalystGolden, T as TraceAnalystKindSpec, d as createTraceAnalystKind, p as parseRawFinding, r as renderPriorFindings } from '../kind-factory-5b7xXXOr.js';
-import { L as LlmClientOptions } from '../llm-client-BeEcAokY.js';
+export { a as AnalystHooks, A as AnalystRegistry, c as AnalystRegistryOptions, B as BudgetPolicy, D as DefaultAnalystRegistryOptions, R as RegistryRunOpts, b as buildDefaultAnalystRegistry } from '../default-registry-6dhErQbs.js';
+export { A as ANALYST_SEVERITIES, C as CreateTraceAnalystKindOpts, R as RAW_FINDING_SCHEMA_PROMPT, a as RawAnalystFinding, b as RawAnalystFindingSchema, c as TraceAnalystGolden, T as TraceAnalystKindSpec, d as createTraceAnalystKind, p as parseRawFinding, r as renderPriorFindings } from '../kind-factory-0BhLSI27.js';
+import { L as LlmClientOptions } from '../llm-client-Bj7g0rqu.js';
 import '../verdict-C9MlYujm.js';
 import '../schema-m0gsnbt3.js';
-import '../store-CKUAgsJz.js';
+import '../store-BcFXE6LG.js';
 import 'zod';
 import '../run-record-e7vj1uZQ.js';
 import '../errors-CzMUYo7b.js';

package/dist/analyst/index.js

@@ -11,12 +11,12 @@ import {
   diffFindings,
   emitSkillUsageFindings,
   runSemanticConceptJudge
-} from "../chunk-XY4DDNEG.js";
+} from "../chunk-MPQWFX6Y.js";
 import {
   behavioralAnalyst,
   buildDefaultAnalystRegistry,
   deriveEfficiencyFindings
-} from "../chunk-UMMZHCPB.js";
+} from "../chunk-MIFZUPEK.js";
 import {
   ANALYST_SEVERITIES,
   AnalystRegistry,
@@ -41,16 +41,16 @@ import {
   renderPriorFindings,
   stripCodeFences,
   structureFindings
-} from "../chunk-D3V5B42D.js";
-import "../chunk-CVVHBFGN.js";
+} from "../chunk-Q5LIB7BC.js";
+import "../chunk-CWNP4DV4.js";
 import {
   computeFindingId,
   makeFinding
 } from "../chunk-45EEMHTC.js";
 import {
   analyzeTraces
-} from "../chunk-UHMJT4T7.js";
-import "../chunk-QAY5UIJO.js";
+} from "../chunk-TBDR6PAI.js";
+import "../chunk-QMUEXQJS.js";
 import "../chunk-PC4UYEBM.js";
 import "../chunk-3BFEG2F6.js";
 import "../chunk-PZ5AY32C.js";

package/dist/{analyze-runs-DwCEkpO_.d.ts → analyze-runs-B6Ljo_dI.d.ts}

@@ -1,7 +1,7 @@
-import { A as AnalystRegistry } from './default-registry-zoGHUQEH.js';
+import { A as AnalystRegistry } from './default-registry-6dhErQbs.js';
 import { a as DatasetScenario } from './dataset-BbGkaN2I.js';
 import { R as RunRecord } from './run-record-e7vj1uZQ.js';
-import { I as InsightReport } from './insight-report-BBwvOh6x.js';
+import { I as InsightReport } from './insight-report-DWl3z9tl.js';
 
 /**
  * # `analyzeRuns()` — turn a set of agent runs into an actionable decision packet.

package/dist/{baseline-DE36-Np7.d.ts → baseline-Bbid3WoO.d.ts}

@@ -1,4 +1,4 @@
-import { T as TraceStore } from './store-CKUAgsJz.js';
+import { T as TraceStore } from './store-BcFXE6LG.js';
 
 /**
  * Tool-use metrics — derived purely from trace data.

package/dist/belief-state/index.d.ts

@@ -1,8 +1,8 @@
-import { c as CalibrationReport } from '../calibration-Cpr3WaX3.js';
+import { c as CalibrationReport } from '../calibration-BPmzuVPk.js';
 import { O as OffPolicyEstimate, a as OffPolicyOptions, b as OffPolicyTrajectory } from '../off-policy-DiwuKKg7.js';
 import { d as CodeAgentSessionSource, a as CodeAgentSessionIntakeOptions, c as CodeAgentSessionMetrics, C as CodeAgentSessionDiagnostic } from '../code-agent-session-BO8nCnv3.js';
 import { R as RunRecord, a as RunSplitTag } from '../run-record-e7vj1uZQ.js';
-import { T as TraceStore } from '../store-CKUAgsJz.js';
+import { T as TraceStore } from '../store-BcFXE6LG.js';
 import { R as RuntimeTrajectoryRecord, P as ProjectRuntimeTrajectoryEvidenceOptions, a as RuntimeTrajectoryEvidenceProjection } from '../runtime-trajectory-BDgfGZSr.js';
 import '../schema-m0gsnbt3.js';
 import '../outcome-store-rnXLEqSn.js';

package/dist/belief-state/index.js

@@ -17,7 +17,7 @@ import {
 } from "../chunk-4DIJWVUT.js";
 import {
   confidenceInterval
-} from "../chunk-LMZQ2Z4U.js";
+} from "../chunk-56GC6VYO.js";
 import {
   ValidationError
 } from "../chunk-3BFEG2F6.js";

package/dist/builder-eval/index.d.ts

@@ -1,7 +1,7 @@
-import { S as SandboxDriver, H as HarnessConfig, a as SandboxHarnessResult, T as TestGradedScenario, b as TestGradedRunResult } from '../test-graded-scenario-BdVaPyHT.js';
-import { T as TraceEmitter } from '../emitter-DEZwY14K.js';
+import { S as SandboxDriver, H as HarnessConfig, a as SandboxHarnessResult, T as TestGradedScenario, b as TestGradedRunResult } from '../test-graded-scenario-DeODGLra.js';
+import { T as TraceEmitter } from '../emitter-C2rqGH_l.js';
 import { R as Run } from '../schema-m0gsnbt3.js';
-import { T as TraceStore } from '../store-CKUAgsJz.js';
+import { T as TraceStore } from '../store-BcFXE6LG.js';
 
 /**
  * BuilderSession — ties a builder-of-builders workflow together.

package/dist/builder-eval/index.js

@@ -1,10 +1,10 @@
 import {
   SandboxHarness,
   runTestGradedScenario
-} from "../chunk-T375SUOZ.js";
+} from "../chunk-ZOPLCJSY.js";
 import {
   judgeSpans
-} from "../chunk-47X6LRCE.js";
+} from "../chunk-UFSG7ACU.js";
 import "../chunk-5BKGXME7.js";
 import {
   TraceEmitter

package/dist/{calibration-Cpr3WaX3.d.ts → calibration-BPmzuVPk.d.ts}

@@ -1,4 +1,4 @@
-import { T as TraceStore } from './store-CKUAgsJz.js';
+import { T as TraceStore } from './store-BcFXE6LG.js';
 import { R as Run } from './schema-m0gsnbt3.js';
 import { O as OutcomeFilter, b as OutcomeStore } from './outcome-store-rnXLEqSn.js';
 

package/dist/campaign/index.d.ts

@@ -1,31 +1,31 @@
 import { A as AnalyzeTracesOptions, a as AnalyzeTracesInput, b as AnalyzeTracesResult } from '../analyst-C8HHvfJp.js';
 import { S as Scenario, M as MutableSurface, b as DispatchContext, a as JudgeConfig, I as ImprovementDriver, q as ProposeContext, m as GenerationRecord, G as Gate, J as JudgeScore, L as LabeledScenarioStore, r as LabeledScenarioWrite, s as LabeledScenarioSampleArgs, t as LabeledScenarioRecord, u as LabelTrust, v as LabeledScenarioSource, g as CampaignResult, i as CodeSurface } from '../types-BU-7W85F.js';
 export { C as CampaignAggregates, d as CampaignArtifactWriter, e as CampaignCellResult, f as CampaignCostMeter, w as CampaignTokenUsage, h as CampaignTraceWriter, D as DispatchFn, j as GateContext, c as GateDecision, k as GateResult, l as GenerationCandidate, x as JudgeAggregate, n as JudgeDimension, o as Mutator, O as OptimizerConfig, P as ParetoParent, y as ProposedCandidate, R as RedactionStatus, z as ScenarioAggregate, p as SessionScript, T as TraceSpan, A as isProposedCandidate, B as labelTrustRank } from '../types-BU-7W85F.js';
-import { b as RunCampaignOptions, c as RunImprovementLoopOptions, C as CampaignStorage } from '../run-improvement-loop-5z_l5zDz.js';
-export { e as GepaDriverConstraints, G as GepaDriverOptions, O as OpenAutoPrOptions, h as OpenAutoPrResult, a as RunImprovementLoopResult, R as RunOptimizationOptions, j as RunOptimizationResult, k as countSentenceEdits, l as defaultRenderDiff, m as extractH2Sections, f as fsCampaignStorage, g as gepaDriver, i as inMemoryCampaignStorage, o as openAutoPr, r as runCampaign, d as runImprovementLoop, n as runOptimization, s as surfaceHash } from '../run-improvement-loop-5z_l5zDz.js';
-export { A as AxisEvidence, a as AxisVerdict, B as BuildEvidenceVectorOptions, k as BuildLoopProvenanceArgs, D as DefaultProductionGateOptions, l as EmitLoopProvenanceArgs, m as EmitLoopProvenanceResult, E as EvidenceVector, b as EvolutionaryDriverOptions, H as HeldOutGateOptions, n as LoopProvenanceBackend, o as LoopProvenanceCandidate, L as LoopProvenanceRecord, O as ObjectiveSource, P as ParetoSignificanceGateOptions, c as PromotionObjective, d as PromotionPolicy, R as RunEvalOptions, e as buildEvidenceVector, q as buildLoopProvenanceRecord, f as composeGate, g as defaultProductionGate, s as emitLoopProvenance, h as evolutionaryDriver, i as heldOutGate, t as loopProvenanceSpans, p as paretoPolicy, j as paretoSignificanceGate, u as provenanceRecordPath, v as provenanceSpansPath, r as runEval, w as surfaceContentHash } from '../provenance-LnqRT0sS.js';
-import { L as LlmClientOptions } from '../llm-client-BeEcAokY.js';
-import { T as TraceAnalystKindSpec } from '../kind-factory-5b7xXXOr.js';
-import { c as AnalystFinding } from '../types-2VVIL04s.js';
+import { b as RunCampaignOptions, c as RunImprovementLoopOptions, C as CampaignStorage } from '../run-improvement-loop-DBahB8Ax.js';
+export { e as GepaDriverConstraints, G as GepaDriverOptions, O as OpenAutoPrOptions, h as OpenAutoPrResult, a as RunImprovementLoopResult, R as RunOptimizationOptions, j as RunOptimizationResult, k as countSentenceEdits, l as defaultRenderDiff, m as extractH2Sections, f as fsCampaignStorage, g as gepaDriver, i as inMemoryCampaignStorage, o as openAutoPr, r as runCampaign, d as runImprovementLoop, n as runOptimization, s as surfaceHash } from '../run-improvement-loop-DBahB8Ax.js';
+export { A as AxisEvidence, a as AxisVerdict, B as BuildEvidenceVectorOptions, k as BuildLoopProvenanceArgs, D as DefaultProductionGateOptions, l as EmitLoopProvenanceArgs, m as EmitLoopProvenanceResult, E as EvidenceVector, b as EvolutionaryDriverOptions, H as HeldOutGateOptions, n as LoopProvenanceBackend, o as LoopProvenanceCandidate, L as LoopProvenanceRecord, O as ObjectiveSource, P as ParetoSignificanceGateOptions, c as PromotionObjective, d as PromotionPolicy, R as RunEvalOptions, e as buildEvidenceVector, q as buildLoopProvenanceRecord, f as composeGate, g as defaultProductionGate, s as emitLoopProvenance, h as evolutionaryDriver, i as heldOutGate, t as loopProvenanceSpans, p as paretoPolicy, j as paretoSignificanceGate, u as provenanceRecordPath, v as provenanceSpansPath, r as runEval, w as surfaceContentHash } from '../provenance-P-bCL2Fo.js';
+import { L as LlmClientOptions } from '../llm-client-Bj7g0rqu.js';
+import { T as TraceAnalystKindSpec } from '../kind-factory-0BhLSI27.js';
+import { c as AnalystFinding } from '../types-Ce17tDlG.js';
 import { S as SignedManifest, A as AgentProfile, B as BackendIntegrityReport, C as CompletionRequirement, R as RuntimeEventLike, a as CompletionVerdict, P as ProducedState, b as CorrectnessChecker } from '../pre-registration-mAnCugl9.js';
-import { E as EProcessState, a as PairedBootstrapResult } from '../statistics-C7PozGrZ.js';
+import { E as EProcessState, a as PairedBootstrapResult } from '../statistics-CCJpTGOS.js';
 import { A as AgentEvalError } from '../errors-CzMUYo7b.js';
 import { a as RunSplitTag, R as RunRecord } from '../run-record-e7vj1uZQ.js';
 import '@ax-llm/ax';
 import '../store-C1YxJDEK.js';
-import '../red-team-BXHil6c8.js';
+import '../red-team-BWdoyleI.js';
 import '../dataset-BbGkaN2I.js';
-import '../store-CKUAgsJz.js';
+import '../store-BcFXE6LG.js';
 import '../schema-m0gsnbt3.js';
 import '../pareto-E-pembql.js';
 import '../hosted/index.js';
-import '../insight-report-BBwvOh6x.js';
-import '../summary-report-DGmUucwQ.js';
-import '../failure-cluster-CL7IVgkJ.js';
+import '../insight-report-DWl3z9tl.js';
+import '../summary-report-BDOFevaT.js';
+import '../failure-cluster-DH9Flgcf.js';
 import '../judge-calibration-0p2QcWNE.js';
 import '../raw-provider-sink-C46HDghv.js';
 import 'zod';
-import '../types-Croy5h7V.js';
+import '../types-C7DGg5ex.js';
 import '@tangle-network/tcloud';
 import '../verdict-C9MlYujm.js';
 

package/dist/campaign/index.js

@@ -10,7 +10,7 @@ import {
   paretoPolicy,
   paretoSignificanceGate,
   runEval
-} from "../chunk-6SOJM3VR.js";
+} from "../chunk-SD2YFWQQ.js";
 import {
   agentProfileHash,
   extractProducedState,
@@ -37,14 +37,14 @@ import {
   runOptimization,
   surfaceContentHash,
   surfaceHash
-} from "../chunk-ZFIBGEOL.js";
+} from "../chunk-EGPMSBEZ.js";
 import {
   assertRealBackend,
   fsCampaignStorage,
   inMemoryCampaignStorage,
   runCampaign,
   summarizeBackendIntegrity
-} from "../chunk-TWS7AZEY.js";
+} from "../chunk-2K6UUZ7P.js";
 import {
   estimateCost,
   isModelPriced
@@ -53,24 +53,24 @@ import {
   AnalystRegistry,
   DEFAULT_TRACE_ANALYST_KINDS,
   createTraceAnalystKind
-} from "../chunk-D3V5B42D.js";
+} from "../chunk-Q5LIB7BC.js";
 import "../chunk-5HRORJQY.js";
 import {
   callLlm
-} from "../chunk-CVVHBFGN.js";
+} from "../chunk-CWNP4DV4.js";
 import "../chunk-45EEMHTC.js";
 import {
   eProcess,
   mulberry32,
   pairedBootstrap
-} from "../chunk-LMZQ2Z4U.js";
+} from "../chunk-56GC6VYO.js";
 import "../chunk-GGE4NNQT.js";
 import {
   analyzeTraces
-} from "../chunk-UHMJT4T7.js";
+} from "../chunk-TBDR6PAI.js";
 import {
   OtlpFileTraceStore
-} from "../chunk-QAY5UIJO.js";
+} from "../chunk-QMUEXQJS.js";
 import "../chunk-PC4UYEBM.js";
 import {
   validateRunRecord

package/dist/{chunk-TWS7AZEY.js → chunk-2K6UUZ7P.js}

@@ -1,6 +1,6 @@
 import {
   confidenceInterval
-} from "./chunk-LMZQ2Z4U.js";
+} from "./chunk-56GC6VYO.js";
 import {
   AgentEvalError
 } from "./chunk-3BFEG2F6.js";
@@ -538,4 +538,4 @@ export {
   inMemoryCampaignStorage,
   runCampaign
 };
-//# sourceMappingURL=chunk-TWS7AZEY.js.map
+//# sourceMappingURL=chunk-2K6UUZ7P.js.map

package/dist/{chunk-Y47J2LJ3.js → chunk-2KNZHH3P.js}

@@ -1,19 +1,19 @@
 import {
   summarizeBackendIntegrity
-} from "./chunk-TWS7AZEY.js";
+} from "./chunk-2K6UUZ7P.js";
 import {
   paretoChart
-} from "./chunk-CY6U5S3X.js";
+} from "./chunk-URWVKRCS.js";
 import {
   cohensD,
   pairedBootstrap,
   pairedMde,
   pairedTTest,
   requiredSampleSize
-} from "./chunk-LMZQ2Z4U.js";
+} from "./chunk-56GC6VYO.js";
 import {
   llmSpans
-} from "./chunk-47X6LRCE.js";
+} from "./chunk-UFSG7ACU.js";
 
 // src/contamination-guard.ts
 function checkCanaries(output, scenarios) {
@@ -856,4 +856,4 @@ export {
   HoldoutAuditor,
   analyzeRuns
 };
-//# sourceMappingURL=chunk-Y47J2LJ3.js.map
+//# sourceMappingURL=chunk-2KNZHH3P.js.map

package/dist/{chunk-LMZQ2Z4U.js → chunk-56GC6VYO.js}

@@ -756,6 +756,37 @@ function pairedMde(opts) {
   const zBeta = zQuantile(power);
   return (zAlpha + zBeta) / Math.sqrt(opts.nPaired);
 }
+function mcnemarRequiredN(opts) {
+  const { p10, p01 } = opts;
+  if (p10 < 0 || p01 < 0 || p10 + p01 > 1) {
+    throw new Error(`mcnemarRequiredN: require p10,p01 \u2265 0 and p10+p01 \u2264 1 (got ${p10}, ${p01})`);
+  }
+  const delta = p10 - p01;
+  if (delta === 0) return Infinity;
+  const alpha = opts.alpha ?? 0.05;
+  const power = opts.power ?? 0.8;
+  const twoSided = opts.twoSided ?? true;
+  const pDisc = p10 + p01;
+  const zAlpha = zQuantile(twoSided ? 1 - alpha / 2 : 1 - alpha);
+  const zBeta = zQuantile(power);
+  const n = (zAlpha * Math.sqrt(pDisc) + zBeta * Math.sqrt(Math.max(0, pDisc - delta * delta))) ** 2 / (delta * delta);
+  return Math.ceil(n);
+}
+function mcnemarPower(opts) {
+  const { p10, p01, nPairs } = opts;
+  if (p10 < 0 || p01 < 0 || p10 + p01 > 1) {
+    throw new Error(`mcnemarPower: require p10,p01 \u2265 0 and p10+p01 \u2264 1 (got ${p10}, ${p01})`);
+  }
+  const alpha = opts.alpha ?? 0.05;
+  const twoSided = opts.twoSided ?? true;
+  const delta = p10 - p01;
+  if (delta === 0 || nPairs <= 0) return alpha;
+  const pDisc = p10 + p01;
+  const zAlpha = zQuantile(twoSided ? 1 - alpha / 2 : 1 - alpha);
+  const denom = Math.sqrt(Math.max(1e-12, pDisc - delta * delta));
+  const zBeta = (Math.sqrt(nPairs) * Math.abs(delta) - zAlpha * Math.sqrt(pDisc)) / denom;
+  return Math.min(1, Math.max(0, normalCdf(zBeta)));
+}
 function bonferroni(pValues, alpha = 0.05) {
   const k = pValues.length;
   const adjusted = pValues.map((p) => Math.min(1, p * k));
@@ -829,6 +860,94 @@ function pairedBootstrap(before, after, opts = {}) {
     resamples
   };
 }
+function wilson(successes, n, confidence = 0.95) {
+  if (n <= 0) return { estimate: 0, lower: 0, upper: 0 };
+  if (successes < 0 || successes > n) {
+    throw new Error(`wilson: successes (${successes}) must be in [0, ${n}]`);
+  }
+  const z = zQuantile(1 - (1 - confidence) / 2);
+  const p = successes / n;
+  const z2 = z * z;
+  const denom = 1 + z2 / n;
+  const center = (p + z2 / (2 * n)) / denom;
+  const half = z * Math.sqrt((p * (1 - p) + z2 / (4 * n)) / n) / denom;
+  return {
+    estimate: p,
+    lower: Math.max(0, center - half),
+    upper: Math.min(1, center + half)
+  };
+}
+function mcnemar(control, treatment) {
+  if (control.length !== treatment.length) {
+    throw new Error(`mcnemar: unequal sample sizes (${control.length} vs ${treatment.length})`);
+  }
+  const n = control.length;
+  let b = 0;
+  let c = 0;
+  for (let i = 0; i < n; i++) {
+    const ctrl = control[i] ? 1 : 0;
+    const treat = treatment[i] ? 1 : 0;
+    if (treat === 1 && ctrl === 0) b++;
+    else if (treat === 0 && ctrl === 1) c++;
+  }
+  const nDiscordant = b + c;
+  const statistic = nDiscordant === 0 ? 0 : (Math.abs(b - c) - 1) ** 2 / nDiscordant;
+  return { n, nDiscordant, b, c, statistic, pValue: binomialSignTwoSided(b, c) };
+}
+function pairedRiskDifference(control, treatment, confidence = 0.95) {
+  if (control.length !== treatment.length) {
+    throw new Error(
+      `pairedRiskDifference: unequal sample sizes (${control.length} vs ${treatment.length})`
+    );
+  }
+  const n = control.length;
+  if (n === 0) return { n: 0, b: 0, c: 0, riskDifference: 0, lower: 0, upper: 0, confidence };
+  let b = 0;
+  let c = 0;
+  for (let i = 0; i < n; i++) {
+    const ctrl = control[i] ? 1 : 0;
+    const treat = treatment[i] ? 1 : 0;
+    if (treat === 1 && ctrl === 0) b++;
+    else if (treat === 0 && ctrl === 1) c++;
+  }
+  const rd = (b - c) / n;
+  const variance = (b + c - (b - c) ** 2 / n) / (n * n);
+  const z = zQuantile(1 - (1 - confidence) / 2);
+  const half = z * Math.sqrt(Math.max(0, variance));
+  return {
+    n,
+    b,
+    c,
+    riskDifference: rd,
+    lower: Math.max(-1, rd - half),
+    upper: Math.min(1, rd + half),
+    confidence
+  };
+}
+function passAtK(n, c, k) {
+  if (!Number.isInteger(n) || !Number.isInteger(c) || !Number.isInteger(k)) {
+    throw new Error(`passAtK: n, c, k must be integers (got n=${n}, c=${c}, k=${k})`);
+  }
+  if (k < 1 || k > n || c < 0 || c > n) {
+    throw new Error(`passAtK: require 1 \u2264 k \u2264 n and 0 \u2264 c \u2264 n (got n=${n}, c=${c}, k=${k})`);
+  }
+  if (n - c < k) return 1;
+  let prob = 1;
+  for (let i = n - c + 1; i <= n; i++) prob *= 1 - k / i;
+  return 1 - prob;
+}
+function binomialSignTwoSided(b, c) {
+  const nd = b + c;
+  if (nd === 0) return 1;
+  const k = Math.min(b, c);
+  const logHalfN = nd * Math.log(0.5);
+  let tail = 0;
+  for (let i = 0; i <= k; i++) {
+    const logChoose = lnGamma(nd + 1) - lnGamma(i + 1) - lnGamma(nd - i + 1);
+    tail += Math.exp(logChoose + logHalfN);
+  }
+  return Math.min(1, 2 * tail);
+}
 function eProcess(opts = {}) {
   const alpha = opts.alpha ?? 0.05;
   const maxBet = opts.maxBet ?? 0.5;
@@ -968,10 +1087,16 @@ export {
   corpusInterRaterAgreementFromJudgeScores,
   requiredSampleSize,
   pairedMde,
+  mcnemarRequiredN,
+  mcnemarPower,
   bonferroni,
   benjaminiHochberg,
   pairedBootstrap,
+  wilson,
+  mcnemar,
+  pairedRiskDifference,
+  passAtK,
   eProcess,
   mulberry322 as mulberry32
 };
-//# sourceMappingURL=chunk-LMZQ2Z4U.js.map
+//# sourceMappingURL=chunk-56GC6VYO.js.map