@tangle-network/agent-eval 0.105.0 → 0.106.0

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
@@ -1,4 +1,4 @@
1
- import { S as Scenario, D as DispatchFn, b as DispatchContext } from '../types-Mh6I44ub.js';
1
+ import { S as Scenario, D as DispatchFn, b as DispatchContext } from '../types-Ctf7XIAL.js';
2
2
  import '../run-record-I-Z3JNvO.js';
3
3
  import '@tangle-network/agent-interface';
4
4
  import '../errors-oeQrLqXC.js';
@@ -1,4 +1,4 @@
1
- import { S as Scenario, J as JudgeScore, D as DispatchFn, a as JudgeConfig } from '../types-Mh6I44ub.js';
1
+ import { S as Scenario, J as JudgeScore, D as DispatchFn, a as JudgeConfig } from '../types-Ctf7XIAL.js';
2
2
  import '../run-record-I-Z3JNvO.js';
3
3
  import '@tangle-network/agent-interface';
4
4
  import '../errors-oeQrLqXC.js';
@@ -1,5 +1,5 @@
1
1
  import { TraceSpanEvent, HostedClient } from '../hosted/index.js';
2
- import '../types-Mh6I44ub.js';
2
+ import '../types-Ctf7XIAL.js';
3
3
  import '../run-record-I-Z3JNvO.js';
4
4
  import '@tangle-network/agent-interface';
5
5
  import '../errors-oeQrLqXC.js';
@@ -1,11 +1,11 @@
1
- import { S as SignedManifest, B as BackendIntegrityReport, C as CompletionRequirement, R as RuntimeEventLike, a as CompletionVerdict, P as ProducedState, b as CorrectnessChecker } from '../pre-registration-D8Zr8-rr.js';
2
- export { L as LlmJudgeDimension, c as LlmJudgeOptions, l as llmJudge } from '../pre-registration-D8Zr8-rr.js';
1
+ import { S as SignedManifest, B as BackendIntegrityReport, C as CompletionRequirement, R as RuntimeEventLike, a as CompletionVerdict, P as ProducedState, b as CorrectnessChecker } from '../pre-registration-BkkTmtQG.js';
2
+ export { L as LlmJudgeDimension, c as LlmJudgeOptions, l as llmJudge } from '../pre-registration-BkkTmtQG.js';
3
3
  import { A as AnalyzeTracesOptions, a as AnalyzeTracesInput, b as AnalyzeTracesResult } from '../analyst-C8HHvfJp.js';
4
- import { S as Scenario, M as MutableSurface, b as DispatchContext, a as JudgeConfig, e as GenerationRecord, g as Gate, 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, y as LabeledScenarioSource, C as CampaignResult, o as CodeSurface } from '../types-Mh6I44ub.js';
5
- export { k as CampaignAggregates, l as CampaignArtifactWriter, m as CampaignCellResult, n as CampaignCostMeter, z as CampaignTokenUsage, d as CampaignTraceWriter, D as DispatchFn, h as GateContext, j as GateDecision, G as GateResult, p as GenerationCandidate, A as JudgeAggregate, c as JudgeDimension, i 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-Mh6I44ub.js';
6
- import { e as CampaignRunPlan, P as PlanCampaignRunOptions, C as CampaignStorage, R as RunCampaignOptions, c as RunImprovementLoopOptions } from '../gepa-Dn0gtj9n.js';
7
- export { h as CampaignRunPlanCell, j as GepaProposerConstraints, G as GepaProposerOptions, O as OpenAutoPrOptions, k as OpenAutoPrResult, b as RunImprovementLoopResult, a as RunOptimizationOptions, l as RunOptimizationResult, m as countSentenceEdits, n as defaultRenderDiff, o as extractH2Sections, f as fsCampaignStorage, g as gepaProposer, i as inMemoryCampaignStorage, p as openAutoPr, q as planCampaignRun, r as runCampaign, d as runImprovementLoop, s as runOptimization, t as surfaceHash } from '../gepa-Dn0gtj9n.js';
8
- 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 EvolutionaryProposerOptions, 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 evolutionaryProposer, 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-phkRBl4h.js';
4
+ import { S as Scenario, M as MutableSurface, b as DispatchContext, a as JudgeConfig, e as GenerationRecord, g as Gate, 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, y as LabeledScenarioSource, C as CampaignResult, m as CodeSurface } from '../types-Ctf7XIAL.js';
5
+ export { i as CampaignAggregates, j as CampaignArtifactWriter, k as CampaignCellResult, l as CampaignCostMeter, z as CampaignTokenUsage, d as CampaignTraceWriter, D as DispatchFn, n as GateContext, h as GateDecision, G as GateResult, o as GenerationCandidate, A as JudgeAggregate, c 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-Ctf7XIAL.js';
6
+ import { e as CampaignRunPlan, P as PlanCampaignRunOptions, C as CampaignStorage, b as RunCampaignOptions, c as RunImprovementLoopOptions } from '../gepa-DQ3ruj18.js';
7
+ export { h as CampaignRunPlanCell, j as GepaProposerConstraints, G as GepaProposerOptions, O as OpenAutoPrOptions, k as OpenAutoPrResult, a as RunImprovementLoopResult, R as RunOptimizationOptions, l as RunOptimizationResult, m as countSentenceEdits, n as defaultRenderDiff, o as extractH2Sections, f as fsCampaignStorage, g as gepaProposer, i as inMemoryCampaignStorage, p as openAutoPr, q as planCampaignRun, r as runCampaign, d as runImprovementLoop, s as runOptimization, t as surfaceHash } from '../gepa-DQ3ruj18.js';
8
+ 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, z as surfaceContentHash } from '../provenance-HDy5TtOY.js';
9
9
  import { E as EProcessState, a as PairedBootstrapResult } from '../statistics-D88peojY.js';
10
10
  import { L as LlmClientOptions } from '../llm-client-DyqEH4jH.js';
11
11
  import { AgentProfile } from '@tangle-network/agent-interface';
@@ -2,15 +2,12 @@ import {
2
2
  buildEvidenceVector,
3
3
  composeGate,
4
4
  defaultProductionGate,
5
- detectScale,
6
- dimensionRegressions,
7
5
  evolutionaryProposer,
8
- heldoutSignificance,
9
- pairHoldout,
10
6
  paretoPolicy,
11
7
  paretoSignificanceGate,
8
+ powerPreflight,
12
9
  runEval
13
- } from "../chunk-Z7FK73R5.js";
10
+ } from "../chunk-JTSA52OJ.js";
14
11
  import {
15
12
  HARNESS_NATIVE_MODEL,
16
13
  agentProfileHash,
@@ -20,21 +17,25 @@ import {
20
17
  harnessAxisOf,
21
18
  llmJudge,
22
19
  verifyCompletion
23
- } from "../chunk-BBQILVBH.js";
20
+ } from "../chunk-W7AJBSNO.js";
24
21
  import {
25
22
  buildLoopProvenanceRecord,
26
23
  campaignBreakdown,
27
24
  campaignMeanComposite,
28
25
  countSentenceEdits,
29
26
  defaultRenderDiff,
27
+ detectScale,
28
+ dimensionRegressions,
30
29
  emitLoopProvenance,
31
30
  extractH2Sections,
32
31
  gepaProposer,
33
32
  heldOutGate,
33
+ heldoutSignificance,
34
34
  isProposedCandidate,
35
35
  labelTrustRank,
36
36
  loopProvenanceSpans,
37
37
  openAutoPr,
38
+ pairHoldout,
38
39
  provenanceRecordPath,
39
40
  provenanceSpansPath,
40
41
  renderAnalystEvidence,
@@ -42,7 +43,7 @@ import {
42
43
  runOptimization,
43
44
  surfaceContentHash,
44
45
  surfaceHash
45
- } from "../chunk-ZKIL3VOD.js";
46
+ } from "../chunk-Y7RY553D.js";
46
47
  import {
47
48
  assertRealBackend,
48
49
  contentHash,
@@ -2778,6 +2779,7 @@ export {
2778
2779
  planCampaignRun,
2779
2780
  planEvalFixtureRun,
2780
2781
  policyEditProposer,
2782
+ powerPreflight,
2781
2783
  provenanceRecordPath,
2782
2784
  provenanceSpansPath,
2783
2785
  renderScoreboardMarkdown,