@tangle-network/agent-eval 0.139.2 → 0.139.3

package/CHANGELOG.md

@@ -4,6 +4,14 @@ All notable changes to `@tangle-network/agent-eval` and its sibling `agent-eval-
 
 ---
 
+## [0.139.3] - 2026-07-31 - supervisor runs under `.agent`
+
+### Changed
+
+- Find supervisor runs under `<ws>/.agent/supervisor`, falling back to the pre-rename `<ws>/.loops/supervisor` so historical runs stay analyzable.
+  agent-runtime 0.111.0 moves the run contract to `.agent`, the one dot-dir for agent-owned state; `.agent` wins when both exist.
+  The "no supervisor run dir" gap messages name both locations.
+
 ## [0.139.2] - 2026-07-31 - interface 0.40 cohort alignment
 
 ### Fixed

package/dist/analyst/index.d.ts

@@ -2,7 +2,7 @@ import { a as MultiLayerVerifier, l as VerifyOptions, o as Severity } from "../m
 import { A as parseFindingSubject, C as FINDING_SUBJECT_KINDS, D as FindingSubjectStringSchema, E as FindingSubjectKind, F as DefineExactCustomAnalystOptions, G as SemanticConceptJudgeOptions, I as defineCustomAnalyst, L as defineTraceAnalyst, M as DspyRlmTraceEngineOptions, N as createDspyRlmTraceEngine, O as KIND_EXPECTED_SUBJECTS, P as DefineCustomAnalystOptions, S as FINDING_SUBJECT_GRAMMAR_PROMPT, T as FindingSubject, W as SemanticConceptJudgeInput, Y as RunCritic, Z as RunTrace, _ as FindingsDiff, a as SkillUsageScanConfig, b as defaultIsMaterial, c as DEFAULT_TRACE_ANALYST_KINDS, d as IMPROVEMENT_KIND_SPEC, f as FAILURE_MODE_KIND_SPEC, g as DiffPolicy, h as emitControlIntegrityFindings, i as SkillUsageReport, j as renderFindingSubject, k as findingSubjectGrammarPromptFor, l as KNOWLEDGE_POISONING_KIND_SPEC, m as ControlIntegrityAnalyst, n as SkillUsageAnalyst, o as buildSkillUsageReport, p as CONTROL_INTEGRITY_ANALYST, r as SkillUsageRecord, s as emitSkillUsageFindings, t as SKILL_USAGE_ANALYST, u as KNOWLEDGE_GAP_KIND_SPEC, v as FindingsStore, w as FINDING_SUBJECT_SYNTAX, x as diffFindings, y as PersistedFinding } from "../skill-usage-B-BFS8M2.js";
 import { b as CustomTokenPricing, c as CostLedgerHandle } from "../cost-ledger-DKgyIWRj.js";
 import { A as ChatClient, B as SandboxSdkTransportOpts, F as CreateChatClientOpts, I as CustomTransportOpts, L as DirectProviderTransportOpts, M as ChatResponse, N as ChatTransport, P as CliBridgeTransportOpts, R as MockTransportOpts, V as createChatClient, j as ChatRequest, k as ChatCallOpts, m as JudgeInput, p as JudgeFn, z as RouterTransportOpts } from "../types-uPrS6mD-.js";
-import { _ as makeFinding, a as AnalystInputKind, c as AnalystRunInputs, d as AnalystSeverity, f as AnalystUsageReceipt, g as computeFindingId, h as ProposalFindingOrigin, i as AnalystFinding, l as AnalystRunResult, m as ProposalFinding, n as AnalystContext, o as AnalystRequirements, p as EvidenceRef, r as AnalystCost, s as AnalystRunEvent, t as Analyst, u as AnalystRunSummary, v as makeProposalFinding } from "../types-DoEYskCd.js";
+import { _ as makeFinding, a as AnalystInputKind, c as AnalystRunInputs, d as AnalystSeverity, f as AnalystUsageReceipt, g as computeFindingId, h as ProposalFindingOrigin, i as AnalystFinding, l as AnalystRunResult, m as ProposalFinding, n as AnalystContext, o as AnalystRequirements, p as EvidenceRef, r as AnalystCost, s as AnalystRunEvent, t as Analyst, u as AnalystRunSummary, v as makeProposalFinding, x as TraceAnalysisStore } from "../types-DoEYskCd.js";
 import { a as createTraceAnalyst, c as runTraceAnalyst, d as deriveEfficiencyFindings, i as TraceAnalystDefinition, l as BehavioralAnalystOptions, n as buildDefaultAnalystRegistry, o as renderPriorFindings, r as CreateTraceAnalystOptions, s as renderUpstreamFindings, t as DefaultAnalystRegistryOptions, u as behavioralAnalyst } from "../default-registry-B8vf7Rmf.js";
 import { a as ExactAnalystRunPolicySnapshot, c as ExactAnalystSnapshot, d as ExactExecutionComponentSnapshot, i as ExactAnalystRunEvent, l as ExactCapableAnalyst, n as ExactAnalystExecutionPlanSnapshot, o as ExactAnalystRunResult, r as ExactAnalystRunCompletion, s as ExactAnalystRunSummary, t as ExactAnalystBudgetSnapshot, u as ExactExecutionComponentIdentity } from "../exact-types-MaaFcllV.js";
 import { A as ExactAnalystRunExecutionError, D as AnalystRegistryOptions, E as AnalystRegistry, M as RegistryRunOpts, N as assertExactRegistryRunOpts, O as BudgetPolicy, T as AnalystHooks, j as ExactRegistryRunOpts, k as ExactAnalystBudgetPolicy } from "../completion-verifier-IPoP4fQO.js";
@@ -331,8 +331,72 @@ interface PublicBenchmarkSelectionReport {
 }
 //#endregion
 //#region src/analyst/benchmark-public-adapters.d.ts
+/**
+ * One contiguous run of incorrect assistant steps, in the shape both public
+ * benchmark runners produce. The direct runner parses it from JSON fields; the
+ * recursive runner parses it from the finding subject. Expansion, evidence
+ * resolution and scoring are identical from here on.
+ */
+interface CodeTraceFailureBlock {
+  firstStep: number;
+  lastStep: number;
+  /** Later step whose action or observation shows the damage this block caused. */
+  consequenceStep: number;
+  escapeStatus: 'escaped' | 'unescaped';
+  severity: AnalystSeverity;
+  claim: string;
+  confidence: number;
+  rationale?: string;
+  recommendedAction?: string;
+  metadata?: Record<string, unknown>;
+}
+/**
+ * What the expansion did with the model's blocks.
+ *
+ * Escaped blocks are scored exactly like unescaped ones — the escape decision
+ * is recorded, never applied — so `escapedBlocks` measures the model's own
+ * judgement without moving precision, recall, or the trusted-negative rate.
+ */
+interface CodeTraceBlockDiagnostics {
+  reportedBlocks: number;
+  escapedBlocks: number;
+  /** Blocks dropped because their consequence step is not a real assistant step. */
+  blocksWithoutConsequenceEvidence: CodeTraceFailureBlock[];
+  /** Interior steps a block claimed that the trace does not carry as assistant steps. */
+  unresolvedBlockInteriorSteps: number[];
+  /** Steps claimed by more than one block; the first block keeps the step. */
+  overlappingBlockSteps: number[];
+}
 declare function emptyPublicBenchmarkRunner(): AnalystBenchmarkRunner<AnalystRunInputs>;
-declare function adaptPublicBenchmarkFindings(dataset: PublicAnalystBenchmarkDataset, trajectoryId: string, findings: readonly AnalystFinding[], analystId: string): AnalystFinding[];
+declare function adaptPublicBenchmarkFindings(options: {
+  dataset: PublicAnalystBenchmarkDataset;
+  trajectoryId: string;
+  findings: readonly AnalystFinding[];
+  analystId: string;
+  store: TraceAnalysisStore;
+  signal?: AbortSignal;
+}): Promise<{
+  findings: AnalystFinding[];
+  diagnostics: CodeTraceBlockDiagnostics | undefined;
+}>;
+/**
+ * Expand contiguous failure blocks into one scored finding per member step.
+ *
+ * The official scorer matches on area plus the exact step evidence URI, so
+ * blocks never reach it: every runner reports blocks, and this function turns
+ * them into the per-step findings the benchmark defines.
+ */
+declare function expandCodeTraceFailureBlocks(options: {
+  trajectoryId: string;
+  blocks: readonly CodeTraceFailureBlock[];
+  store: TraceAnalysisStore;
+  analystId: string;
+  producedAt?: string;
+  signal?: AbortSignal;
+}): Promise<{
+  findings: AnalystFinding[];
+  diagnostics: CodeTraceBlockDiagnostics;
+}>;
 //#endregion
 //#region src/analyst/benchmark-public-data.d.ts
 declare function loadPublicBenchmarkRows(path: string): Promise<Array<Record<string, unknown>>>;
@@ -352,11 +416,29 @@ declare function preparePublicAnalystBenchmark(options: {
   seed: number;
 }): Promise<PreparedPublicAnalystBenchmark>;
 //#endregion
+//#region src/analyst/benchmark-public-prompt.d.ts
+/** Widest contiguous failure block a model may report. The published corpus's
+ * widest labeled block is 8 steps and its widest stage span is 9, so this bound
+ * never binds honest enumeration; it caps how far one over-wide block can push
+ * unlabeled steps into the precision denominator. */
+declare const MAX_INCORRECT_BLOCK_STEPS = 12;
+/** Most blocks a model may report for one trajectory. The published corpus's
+ * densest case carries 4 disjoint labeled blocks. Together with the per-block
+ * cap this bounds one case at 192 predicted steps without a second ceiling. */
+declare const MAX_INCORRECT_BLOCKS = 16;
+declare const CODE_TRACE_BENCH_ANALYST_PROMPT = "Analyze exactly one coding-agent trajectory and its attached final verification.\nYour task is the CodeTraceBench incorrect-step task: identify every wrong state-changing action, bad hypothesis that drives an action, and regression.\nAn incorrect step remains incorrect when the agent later recovers or the final verification passes; a trajectory that ends in success still contains every mistake the agent made along the way.\nIncorrect steps occur in contiguous failure blocks: one mistake plus every consecutive following step that commits to, compounds, or acts on it.\nReport each failure block as exactly one finding whose first_step is the block's first incorrect step and whose last_step is its last, covering every consecutive step between them.\nSet first_step to the first step that commits the mistake, not the step that planned it and not a later step that repeats it.\nExtend last_step one step at a time, and only while the next step independently satisfies the incorrect-step definition on its own action and its own following observation.\nStop at the first step where the agent detects the problem, inspects it, or begins repairing it: a diagnostic probe, a test run that exposes the defect, or a repair action ends the block and is never inside it.\nA one-step block is a complete and correct answer.\nEvery step inside a block is scored on its own: naming a correct step costs exactly as much as missing an incorrect one, and naming only the first step of a longer block forfeits every unnamed step.\nReport blocks separated by at least one correct step as separate findings, and never let two blocks overlap.\nInspect the complete supplied trace data.\nUse the final-verification outcome as evidence about the final state, not as a rule for whether earlier steps were incorrect.\nFor each candidate block, inspect every assistant action in it and its following observation.\nAdmit a block only when you can point at the specific evidence it produced: name as consequence_step the step number whose action or observation shows the damage — a failing command, a wrong file state, a repeated failure, or rework the agent had to do because of this block. That step is the block's own last step when its observation already shows the damage, and a later step otherwise.\nWhen you cannot name that later step number from the trace you were given, drop the block; a plausible story about why a step looks wrong is not evidence that it was.\nJudge that consequence from the trajectory itself: a passing final verification is not evidence that a block caused nothing, and a failing final verification is not evidence that any particular block caused it.\nFor every block, decide whether the agent escaped the failure.\nMark escape_status \"escaped\" only when you can name the single later step that fully reversed the block, the agent needed no other step to recover, and nothing after that step revisits the same file, command, or hypothesis; write that step number in the rationale.\nMark escape_status \"unescaped\" in every other case, including whenever you are unsure.\nA passing final verification never makes a block escaped: the agent may have made the mistake and repaired it over several steps, and those steps are still incorrect.\nLabel a failed command when the assistant caused it through a wrong action or unsupported hypothesis.\nLabel the later corrective action only when that action is itself wrong.\nDo not label a diagnostic probe merely because it exposes an earlier defect.\nDo not label a redundant but correct read or search; CodeTraceBench scores unuseful steps separately, and this run scores incorrect steps only.\nDo not label a step solely because final verification failed.\nWhen final verification is unavailable, use only directly observed trajectory evidence.\nEvery step in a reported block MUST be the positive integer n from an existing assistant LLM span named step-<n>.\nNever select an EVALUATOR, TOOL, CHAIN, final-verification, benchmark-verification, or message-<n> span.\nBefore emitting a finding, inspect every covered span's attributes.content and describe only the actions shown there.\nReport at most 16 blocks and at most 12 steps in one block; when more candidates than that exist, report the ones you can support with the clearest downstream evidence.\nWhen the trajectory has no incorrect steps, return an empty findings array.";
+/** One-shot JSON transport prompt for the direct runner. */
+declare function publicBenchmarkSystemPrompt(dataset: PublicAnalystBenchmarkDataset): string;
+/** Tool-loop prompt for the recursive runner. Same task, subject-encoded block. */
+declare function publicBenchmarkRlmInstructions(dataset: PublicAnalystBenchmarkDataset): string;
+/** Digest of every prompt a runner can send plus the shared transport limits.
+ * Both runner contracts are hashed so an edit to either one changes the digest
+ * a run records, whichever runner executed. */
+declare function publicBenchmarkProtocolSha256(dataset: PublicAnalystBenchmarkDataset): string;
+//#endregion
 //#region src/analyst/benchmark-public-model.d.ts
 /** One-shot JSON baseline. This is not a recursive trace analyst. */
 declare function createPublicBenchmarkDirectRunner(dataset: PublicAnalystBenchmarkDataset, config: PublicAnalystBenchmarkModelConfig): AnalystBenchmarkRunner<AnalystRunInputs>;
-declare function publicBenchmarkProtocolSha256(dataset: PublicAnalystBenchmarkDataset): string;
-declare const CODE_TRACE_BENCH_ANALYST_PROMPT = "Analyze exactly one coding-agent trajectory and its attached final verification.\nYour task is the CodeTraceBench incorrect-step task: identify every wrong state-changing action, bad hypothesis that drives an action, and regression.\nAn incorrect step remains incorrect when the agent later recovers or the final verification passes.\nInspect the complete supplied trace data.\nUse the final-verification outcome as evidence about the final state, not as a rule for whether earlier steps were incorrect.\nFor each candidate, inspect the assistant action and its following observation.\nLabel a failed command when the assistant caused it through a wrong action or unsupported hypothesis.\nLabel the later corrective action only when that action is itself wrong.\nDo not label a diagnostic probe merely because it exposes an earlier defect.\nDo not label a redundant but correct read or search; CodeTraceBench scores unuseful steps separately, and this run scores incorrect steps only.\nDo not label a step solely because final verification failed.\nWhen final verification is unavailable, use only directly observed trajectory evidence.\nEmit one finding per incorrect assistant step.\nEach finding's step MUST be the positive integer n from an existing assistant LLM span named step-<n>.\nNever select an EVALUATOR, TOOL, CHAIN, final-verification, benchmark-verification, or message-<n> span.\nBefore emitting a finding, inspect its candidate span's attributes.content and describe only the action shown there.\nWhen the trajectory has no incorrect steps, return an empty findings array.";
 //#endregion
 //#region src/analyst/benchmark-public-rlm.d.ts
 /** Public benchmark candidate that runs the actual recursive trace analyst. */
@@ -505,7 +587,7 @@ interface AnalystBenchmarkRunIdentity {
     analystProtocolSha256: string;
     implementationSha256: string;
     dependencyLockSha256: string;
-    runnerIds: readonly ['empty', 'dspy-rlm'];
+    runnerIds: readonly ['empty', string];
   };
   inputs: {
     labelsSha256: string;
@@ -573,8 +655,18 @@ declare function readAnalystBenchmarkArtifact(path: string): Promise<AnalystBenc
 interface AnalystBenchmarkCommandDependencies {
   createAnalystRunner?: (dataset: PublicAnalystBenchmarkDataset, config: PublicAnalystBenchmarkModelConfig) => AnalystBenchmarkRunner<AnalystRunInputs>;
 }
+/**
+ * Which analyst produces the scored arm.
+ *
+ * `dspy-rlm` is the recursive engine. `direct` is the retired one-shot runner,
+ * kept reachable because the published evidence was produced by it: a
+ * comparison against those numbers is only sound when the same runner can be
+ * re-run over the same inputs.
+ */
+type AnalystBenchmarkRunnerKind = 'dspy-rlm' | 'direct';
 interface AnalystBenchmarkCommandConfig {
   dataset: PublicAnalystBenchmarkDataset;
+  analyst: AnalystBenchmarkRunnerKind;
   labelsPath: string;
   traceDir: string;
   artifactDir?: string;
@@ -593,13 +685,13 @@ interface AnalystBenchmarkCommandConfig {
   resume: boolean;
 }
 declare function runAnalystBenchmarkCommand(argv: readonly string[], env?: NodeJS.ProcessEnv, dependencies?: AnalystBenchmarkCommandDependencies): Promise<number>;
-declare const ANALYST_BENCHMARK_HELP = "agent-eval analyst-benchmark\n\nRun the recursive DSPy trace analyst against public AgentRx or CodeTraceBench labels.\n\nRequired:\n  --dataset agentrx|codetracebench\n  --labels <dataset.json|dataset.jsonl>\n  --trace-dir <one-trace-per-file OTLP JSONL directory>\n  --artifact-dir <extracted artifact root>  Required for CodeTraceBench\n  --out <new output directory>\n  --revision <full 40- or 64-character hex digest>\n  --split <dataset split>\n  --base-url <OpenAI-compatible /v1 URL>\n  --api-key-env <environment variable containing the bearer>\n  --model <provider model id>\n  --limit <positive case count>\n\nControls:\n  --resume                         Continue an interrupted run in --out\n  --seed <integer>                 Case-selection and comparison seed. Default: 0\n  --concurrency <positive integer> Parallel benchmark jobs. Default: 1\n  --repetitions <positive integer> Runs per case and runner. Default: 1\n  --max-output-tokens <positive>   Model output limit per call. Default: 4096\n  --python <executable>             Python with agent-eval-rpc[dspy]. Default: python\n  --timeout-ms <positive>          Model analyst deadline per case. Default: 300000\n  --max-cost-usd <positive>        Run-wide spend limit. Default: 5\n  --max-artifact-bytes <positive>  Final evidence bytes per case. Default: 8388608\n\nWrites result.json with every observation, metric, usage field, error, comparison,\ninput digest, artifact digest, case distribution, selected case id, and explicit\nunknown cost. Limited deterministic-hash subsets are marked non-representative.\nCompleted observations are fsynced to observations.jsonl. Shareable output is in\nresult.json and report.md. Machine-local paths, endpoint, and command are isolated\nin run.local.json.\nThe key is read from the named environment variable and is never written.";
+declare const ANALYST_BENCHMARK_HELP = "agent-eval analyst-benchmark\n\nRun the recursive DSPy trace analyst against public AgentRx or CodeTraceBench labels.\n\nRequired:\n  --dataset agentrx|codetracebench\n  --analyst dspy-rlm|direct        Scored analyst. Default: dspy-rlm.\n                                   'direct' is the retired one-shot runner that\n                                   produced the published evidence.\n  --labels <dataset.json|dataset.jsonl>\n  --trace-dir <one-trace-per-file OTLP JSONL directory>\n  --artifact-dir <extracted artifact root>  Required for CodeTraceBench\n  --out <new output directory>\n  --revision <full 40- or 64-character hex digest>\n  --split <dataset split>\n  --base-url <OpenAI-compatible /v1 URL>\n  --api-key-env <environment variable containing the bearer>\n  --model <provider model id>\n  --limit <positive case count>\n\nControls:\n  --resume                         Continue an interrupted run in --out\n  --seed <integer>                 Case-selection and comparison seed. Default: 0\n  --concurrency <positive integer> Parallel benchmark jobs. Default: 1\n  --repetitions <positive integer> Runs per case and runner. Default: 1\n  --max-output-tokens <positive>   Model output limit per call. Default: 4096\n  --python <executable>             Python with agent-eval-rpc[dspy]. Default: python\n  --timeout-ms <positive>          Model analyst deadline per case. Default: 300000\n  --max-cost-usd <positive>        Run-wide spend limit. Default: 5\n  --max-artifact-bytes <positive>  Final evidence bytes per case. Default: 8388608\n\nWrites result.json with every observation, metric, usage field, error, comparison,\ninput digest, artifact digest, case distribution, selected case id, and explicit\nunknown cost. Limited deterministic-hash subsets are marked non-representative.\nCompleted observations are fsynced to observations.jsonl. Shareable output is in\nresult.json and report.md. Machine-local paths, endpoint, and command are isolated\nin run.local.json.\nThe key is read from the named environment variable and is never written.";
 //#endregion
 //#region src/analyst/benchmark-implementation.d.ts
 declare const ANALYST_BENCHMARK_IMPLEMENTATION_DIGEST_ALGORITHM = "sha256-canonical-source-manifest";
 declare const ANALYST_BENCHMARK_DEPENDENCY_LOCK_DIGEST_ALGORITHM = "sha256-canonical-file-manifest";
 declare const ANALYST_BENCHMARK_DEPENDENCY_LOCK_FILES: readonly string[];
-declare const ANALYST_BENCHMARK_DEPENDENCY_LOCK_SHA256 = "1ca280998a9f3a416b40cadb9c14789ada490a727a4ec78491c92bb2022e6c95";
+declare const ANALYST_BENCHMARK_DEPENDENCY_LOCK_SHA256 = "11ea62baf2914dba71dd62e45ba4bb8e856b2cefdd5b1b73dc2f14b4b6e6d115";
 /** The published benchmark evidence was produced at this package version, by
  * the retired one-shot direct runner, before trace analysts moved to the
  * recursive DSPy RLM engine. Both evidence digests below are historical facts
@@ -611,7 +703,7 @@ declare const ANALYST_BENCHMARK_EVIDENCE_PACKAGE_VERSION = "0.137.0";
 declare const ANALYST_BENCHMARK_EVIDENCE_DEPENDENCY_LOCK_SHA256 = "1e03f2daed356d60316aabefb407ec1e437ac94d408d61eea4ae096e9c6fbb5b";
 declare const ANALYST_BENCHMARK_EVIDENCE_IMPLEMENTATION_SHA256 = "4dba263b6256a30d56c7fdb2d992d3a953c0035d731f359b704db806f68f75ac";
 declare const ANALYST_BENCHMARK_IMPLEMENTATION_FILES: readonly string[];
-declare const ANALYST_BENCHMARK_IMPLEMENTATION_SHA256 = "5e47fc9d9f49c468d3ef06c5d552925e6a026f9e22a75056a9c8c5a2879744c0";
+declare const ANALYST_BENCHMARK_IMPLEMENTATION_SHA256 = "26820de559ca069145eb9649283d2efba51a0a7b8d3662d4c378b9ecebfd4777";
 declare function analystBenchmarkImplementationDigest(): string;
 declare function analystBenchmarkDependencyLockDigest(): string;
 //#endregion
@@ -655,5 +747,5 @@ declare function isProposalFinding(finding: unknown): finding is ProposalFinding
  */
 declare function assertProposalFindings(findings: unknown, context?: string): ReadonlyArray<ProposalFinding>;
 //#endregion
-export { AGENT_RX_UPSTREAM_REVISION, ANALYST_BENCHMARK_COST_LEDGER_FILE, ANALYST_BENCHMARK_DEPENDENCY_LOCK_DIGEST_ALGORITHM, ANALYST_BENCHMARK_DEPENDENCY_LOCK_FILES, ANALYST_BENCHMARK_DEPENDENCY_LOCK_SHA256, ANALYST_BENCHMARK_EVIDENCE_DEPENDENCY_LOCK_SHA256, ANALYST_BENCHMARK_EVIDENCE_IMPLEMENTATION_SHA256, ANALYST_BENCHMARK_EVIDENCE_PACKAGE_VERSION, ANALYST_BENCHMARK_HELP, ANALYST_BENCHMARK_IMPLEMENTATION_DIGEST_ALGORITHM, ANALYST_BENCHMARK_IMPLEMENTATION_FILES, ANALYST_BENCHMARK_IMPLEMENTATION_SHA256, ANALYST_BENCHMARK_LOCAL_RECEIPT_FILE, ANALYST_BENCHMARK_MANIFEST_FILE, ANALYST_BENCHMARK_OBSERVATIONS_FILE, ANALYST_SEVERITIES, type AgentRxBenchmarkCaseOptions, type AgentRxCalibrationRunnerSummary, type AgentRxCalibrationSummary, type AgentRxFailure, type AgentRxPrediction, type AgentRxPredictionReport, type AgentRxRow, type Analyst, type AnalystBenchmarkArtifact, type AnalystBenchmarkCase, type AnalystBenchmarkCommandConfig, type AnalystBenchmarkCommandDependencies, type AnalystBenchmarkDatasetRef, type AnalystBenchmarkDescriptor, type AnalystBenchmarkError, type AnalystBenchmarkLabelState, type AnalystBenchmarkLocalRunReceipt, type AnalystBenchmarkObservation, type AnalystBenchmarkOutput, type AnalystBenchmarkProgressRow, type AnalystBenchmarkProvenance, type AnalystBenchmarkResult, type AnalystBenchmarkRunIdentity, type AnalystBenchmarkRunManifest, type AnalystBenchmarkRunner, type AnalystBenchmarkSummary, type AnalystComparisonMetric, type AnalystContext, type AnalystCost, type AnalystEvidenceExpectation, type AnalystEvidenceResolution, type AnalystEvidenceResolutionError, type AnalystEvidenceResolver, type AnalystFinding, type AnalystFindingScore, type AnalystHooks, type AnalystInputKind, type AnalystIssueExpectation, type AnalystLatencyDistribution, type AnalystMetricComparison, AnalystRegistry, type AnalystRegistryOptions, type AnalystRequirements, type AnalystRunEvent, type AnalystRunInputs, type AnalystRunResult, type AnalystRunSummary, type AnalystRunnerComparison, type AnalystSeverity, type AnalystUsageReceipt, type BehavioralAnalystOptions, type BudgetPolicy, CODE_TRACE_BENCH_ANALYST_PROMPT, CONTROL_INTEGRITY_ANALYST, type ChatCallOpts, type ChatClient, type ChatRequest, type ChatResponse, type ChatTransport, type CliBridgeTransportOpts, type CodeTraceBenchCaseOptions, type CodeTraceBenchLabelOptions, type CodeTraceBenchLabelSet, type CodeTraceBenchRow, type CodeTraceCalibrationRunnerSummary, type CodeTraceCalibrationSummary, type CodeTraceStageAnnotation, type CodeTracerLabelGroup, type CodeTracerPredictionAdapterOptions, type CodeTracerPredictions, type CodeTracerStepLabel, ControlIntegrityAnalyst, type CreateChatClientOpts, type CreateTraceAnalystOptions, type CustomTransportOpts, DEFAULT_MAX_VERIFICATION_ARTIFACT_BYTES, DEFAULT_TRACE_ANALYST_KINDS, DEFAULT_TRACE_ANALYST_LIMITS, type DefaultAnalystRegistryOptions, type DefineCustomAnalystOptions, type DefineExactCustomAnalystOptions, type DiffPolicy, type DirectProviderTransportOpts, type DspyRlmTraceEngineOptions, type EvidenceRef, type ExactAnalystBudgetPolicy, type ExactAnalystBudgetSnapshot, type ExactAnalystExecutionPlanSnapshot, type ExactAnalystRunCompletion, type ExactAnalystRunEvent, ExactAnalystRunExecutionError, type ExactAnalystRunPolicySnapshot, type ExactAnalystRunResult, type ExactAnalystRunSummary, type ExactAnalystSnapshot, type ExactCapableAnalyst, type ExactExecutionComponentIdentity, type ExactExecutionComponentSnapshot, type ExactRegistryRunOpts, FAILURE_MODE_KIND_SPEC, FINDING_SUBJECT_GRAMMAR_PROMPT, FINDING_SUBJECT_KINDS, FINDING_SUBJECT_SYNTAX, type FindingSubject, type FindingSubjectKind, FindingSubjectStringSchema, type FindingsDiff, FindingsStore, IMPROVEMENT_KIND_SPEC, type JudgeAdapterOpts, KIND_EXPECTED_SUBJECTS, KNOWLEDGE_GAP_KIND_SPEC, KNOWLEDGE_POISONING_KIND_SPEC, type LoadedVerificationArtifacts, type MockTransportOpts, type PersistedFinding, type PreparedPublicAnalystBenchmark, type ProposalFinding, type ProposalFindingOrigin, type PublicAnalystBenchmarkDataset, type PublicAnalystBenchmarkModelConfig, type PublicBenchmarkDistributions, type PublicBenchmarkSelectionReport, type PublicBenchmarkValueDistribution, RAW_FINDING_SCHEMA_PROMPT, type RawAnalystEvidence, RawAnalystEvidenceSchema, type RawAnalystFinding, RawAnalystFindingSchema, type RegistryRunOpts, type RouterTransportOpts, type RunAnalystBenchmarkOptions, type RunCriticAdapterOpts, SKILL_USAGE_ANALYST, type SandboxSdkTransportOpts, type SemanticConceptJudgeAdapterOpts, SkillUsageAnalyst, type SkillUsageRecord, type SkillUsageReport, type SkillUsageScanConfig, type StepLabelAdapterOptions, type TraceAnalysisEngine, type TraceAnalysisEngineRequest, type TraceAnalysisEngineResult, type TraceAnalystDefinition, type TraceAnalystLimits, type TraceToolGroupName, type UpstreamPredictionAdapterOptions, type VerificationArtifactFile, type VerificationArtifactManifest, type VerificationArtifactRole, type VerificationAvailabilitySummary, type VerificationOutcome, type VerificationOutcomeSource, type VerificationOutcomeStatus, type VerificationResultFile, type VerifierAdapterOpts, adaptPublicBenchmarkFindings, agentRxBenchmarkCase, agentRxPredictionsToFindings, analystBenchmarkDependencyLockDigest, analystBenchmarkImplementationDigest, appendVerificationArtifactsToOtlp, assertExactRegistryRunOpts, assertProposalFindings, behavioralAnalyst, buildDefaultAnalystRegistry, buildSkillUsageReport, buildTraceToolsForGroup, codeTraceBenchCase, codeTracerPredictionsToFindings, coerceJson, coerceToFindingRows, compareAnalystRunners, computeFindingId, createChatClient, createDspyRlmTraceEngine, createJudgeAdapter, createPublicBenchmarkDirectRunner, createPublicBenchmarkRlmRunner, createRunCriticAdapter, createSemanticConceptJudgeAdapter, createTraceAnalyst, createVerifierAdapter, defaultIsMaterial, defineCustomAnalyst, defineTraceAnalyst, deriveEfficiencyFindings, diffFindings, emitControlIntegrityFindings, emitSkillUsageFindings, emptyPublicBenchmarkRunner, evidenceRefsFromRawFinding, findingSubjectGrammarPromptFor, isProposalFinding, liftSeverity, loadCodeTraceVerificationArtifacts, loadPublicBenchmarkRows, makeFinding, makeProposalFinding, normalizeAgentRxCategory, normalizeBenchmarkLabel, parseFindingSubject, parseRawFinding, parseVerificationOutcome, preparePublicAnalystBenchmark, publicBenchmarkDistributions, publicBenchmarkProtocolSha256, publicBenchmarkSelectionReport, readAnalystBenchmarkArtifact, registryBenchmarkRunner, renderAgentRxCalibrationMarkdown, renderAnalystBenchmarkMarkdown, renderCodeTraceCalibrationMarkdown, renderFindingSubject, renderPriorFindings, renderUpstreamFindings, resolveTraceAnalystLimits, roundAgentRxStep, runAnalystBenchmark, runAnalystBenchmarkCommand, runTraceAnalyst, scoreAnalystFindings, selectPublicBenchmarkRows, stripCodeFences, summarizeAgentRxCalibration, summarizeCodeTraceCalibration, traceStoreEvidenceResolver };
+export { AGENT_RX_UPSTREAM_REVISION, ANALYST_BENCHMARK_COST_LEDGER_FILE, ANALYST_BENCHMARK_DEPENDENCY_LOCK_DIGEST_ALGORITHM, ANALYST_BENCHMARK_DEPENDENCY_LOCK_FILES, ANALYST_BENCHMARK_DEPENDENCY_LOCK_SHA256, ANALYST_BENCHMARK_EVIDENCE_DEPENDENCY_LOCK_SHA256, ANALYST_BENCHMARK_EVIDENCE_IMPLEMENTATION_SHA256, ANALYST_BENCHMARK_EVIDENCE_PACKAGE_VERSION, ANALYST_BENCHMARK_HELP, ANALYST_BENCHMARK_IMPLEMENTATION_DIGEST_ALGORITHM, ANALYST_BENCHMARK_IMPLEMENTATION_FILES, ANALYST_BENCHMARK_IMPLEMENTATION_SHA256, ANALYST_BENCHMARK_LOCAL_RECEIPT_FILE, ANALYST_BENCHMARK_MANIFEST_FILE, ANALYST_BENCHMARK_OBSERVATIONS_FILE, ANALYST_SEVERITIES, type AgentRxBenchmarkCaseOptions, type AgentRxCalibrationRunnerSummary, type AgentRxCalibrationSummary, type AgentRxFailure, type AgentRxPrediction, type AgentRxPredictionReport, type AgentRxRow, type Analyst, type AnalystBenchmarkArtifact, type AnalystBenchmarkCase, type AnalystBenchmarkCommandConfig, type AnalystBenchmarkCommandDependencies, type AnalystBenchmarkDatasetRef, type AnalystBenchmarkDescriptor, type AnalystBenchmarkError, type AnalystBenchmarkLabelState, type AnalystBenchmarkLocalRunReceipt, type AnalystBenchmarkObservation, type AnalystBenchmarkOutput, type AnalystBenchmarkProgressRow, type AnalystBenchmarkProvenance, type AnalystBenchmarkResult, type AnalystBenchmarkRunIdentity, type AnalystBenchmarkRunManifest, type AnalystBenchmarkRunner, type AnalystBenchmarkSummary, type AnalystComparisonMetric, type AnalystContext, type AnalystCost, type AnalystEvidenceExpectation, type AnalystEvidenceResolution, type AnalystEvidenceResolutionError, type AnalystEvidenceResolver, type AnalystFinding, type AnalystFindingScore, type AnalystHooks, type AnalystInputKind, type AnalystIssueExpectation, type AnalystLatencyDistribution, type AnalystMetricComparison, AnalystRegistry, type AnalystRegistryOptions, type AnalystRequirements, type AnalystRunEvent, type AnalystRunInputs, type AnalystRunResult, type AnalystRunSummary, type AnalystRunnerComparison, type AnalystSeverity, type AnalystUsageReceipt, type BehavioralAnalystOptions, type BudgetPolicy, CODE_TRACE_BENCH_ANALYST_PROMPT, CONTROL_INTEGRITY_ANALYST, type ChatCallOpts, type ChatClient, type ChatRequest, type ChatResponse, type ChatTransport, type CliBridgeTransportOpts, type CodeTraceBenchCaseOptions, type CodeTraceBenchLabelOptions, type CodeTraceBenchLabelSet, type CodeTraceBenchRow, type CodeTraceBlockDiagnostics, type CodeTraceCalibrationRunnerSummary, type CodeTraceCalibrationSummary, type CodeTraceFailureBlock, type CodeTraceStageAnnotation, type CodeTracerLabelGroup, type CodeTracerPredictionAdapterOptions, type CodeTracerPredictions, type CodeTracerStepLabel, ControlIntegrityAnalyst, type CreateChatClientOpts, type CreateTraceAnalystOptions, type CustomTransportOpts, DEFAULT_MAX_VERIFICATION_ARTIFACT_BYTES, DEFAULT_TRACE_ANALYST_KINDS, DEFAULT_TRACE_ANALYST_LIMITS, type DefaultAnalystRegistryOptions, type DefineCustomAnalystOptions, type DefineExactCustomAnalystOptions, type DiffPolicy, type DirectProviderTransportOpts, type DspyRlmTraceEngineOptions, type EvidenceRef, type ExactAnalystBudgetPolicy, type ExactAnalystBudgetSnapshot, type ExactAnalystExecutionPlanSnapshot, type ExactAnalystRunCompletion, type ExactAnalystRunEvent, ExactAnalystRunExecutionError, type ExactAnalystRunPolicySnapshot, type ExactAnalystRunResult, type ExactAnalystRunSummary, type ExactAnalystSnapshot, type ExactCapableAnalyst, type ExactExecutionComponentIdentity, type ExactExecutionComponentSnapshot, type ExactRegistryRunOpts, FAILURE_MODE_KIND_SPEC, FINDING_SUBJECT_GRAMMAR_PROMPT, FINDING_SUBJECT_KINDS, FINDING_SUBJECT_SYNTAX, type FindingSubject, type FindingSubjectKind, FindingSubjectStringSchema, type FindingsDiff, FindingsStore, IMPROVEMENT_KIND_SPEC, type JudgeAdapterOpts, KIND_EXPECTED_SUBJECTS, KNOWLEDGE_GAP_KIND_SPEC, KNOWLEDGE_POISONING_KIND_SPEC, type LoadedVerificationArtifacts, MAX_INCORRECT_BLOCKS, MAX_INCORRECT_BLOCK_STEPS, type MockTransportOpts, type PersistedFinding, type PreparedPublicAnalystBenchmark, type ProposalFinding, type ProposalFindingOrigin, type PublicAnalystBenchmarkDataset, type PublicAnalystBenchmarkModelConfig, type PublicBenchmarkDistributions, type PublicBenchmarkSelectionReport, type PublicBenchmarkValueDistribution, RAW_FINDING_SCHEMA_PROMPT, type RawAnalystEvidence, RawAnalystEvidenceSchema, type RawAnalystFinding, RawAnalystFindingSchema, type RegistryRunOpts, type RouterTransportOpts, type RunAnalystBenchmarkOptions, type RunCriticAdapterOpts, SKILL_USAGE_ANALYST, type SandboxSdkTransportOpts, type SemanticConceptJudgeAdapterOpts, SkillUsageAnalyst, type SkillUsageRecord, type SkillUsageReport, type SkillUsageScanConfig, type StepLabelAdapterOptions, type TraceAnalysisEngine, type TraceAnalysisEngineRequest, type TraceAnalysisEngineResult, type TraceAnalystDefinition, type TraceAnalystLimits, type TraceToolGroupName, type UpstreamPredictionAdapterOptions, type VerificationArtifactFile, type VerificationArtifactManifest, type VerificationArtifactRole, type VerificationAvailabilitySummary, type VerificationOutcome, type VerificationOutcomeSource, type VerificationOutcomeStatus, type VerificationResultFile, type VerifierAdapterOpts, adaptPublicBenchmarkFindings, agentRxBenchmarkCase, agentRxPredictionsToFindings, analystBenchmarkDependencyLockDigest, analystBenchmarkImplementationDigest, appendVerificationArtifactsToOtlp, assertExactRegistryRunOpts, assertProposalFindings, behavioralAnalyst, buildDefaultAnalystRegistry, buildSkillUsageReport, buildTraceToolsForGroup, codeTraceBenchCase, codeTracerPredictionsToFindings, coerceJson, coerceToFindingRows, compareAnalystRunners, computeFindingId, createChatClient, createDspyRlmTraceEngine, createJudgeAdapter, createPublicBenchmarkDirectRunner, createPublicBenchmarkRlmRunner, createRunCriticAdapter, createSemanticConceptJudgeAdapter, createTraceAnalyst, createVerifierAdapter, defaultIsMaterial, defineCustomAnalyst, defineTraceAnalyst, deriveEfficiencyFindings, diffFindings, emitControlIntegrityFindings, emitSkillUsageFindings, emptyPublicBenchmarkRunner, evidenceRefsFromRawFinding, expandCodeTraceFailureBlocks, findingSubjectGrammarPromptFor, isProposalFinding, liftSeverity, loadCodeTraceVerificationArtifacts, loadPublicBenchmarkRows, makeFinding, makeProposalFinding, normalizeAgentRxCategory, normalizeBenchmarkLabel, parseFindingSubject, parseRawFinding, parseVerificationOutcome, preparePublicAnalystBenchmark, publicBenchmarkDistributions, publicBenchmarkProtocolSha256, publicBenchmarkRlmInstructions, publicBenchmarkSelectionReport, publicBenchmarkSystemPrompt, readAnalystBenchmarkArtifact, registryBenchmarkRunner, renderAgentRxCalibrationMarkdown, renderAnalystBenchmarkMarkdown, renderCodeTraceCalibrationMarkdown, renderFindingSubject, renderPriorFindings, renderUpstreamFindings, resolveTraceAnalystLimits, roundAgentRxStep, runAnalystBenchmark, runAnalystBenchmarkCommand, runTraceAnalyst, scoreAnalystFindings, selectPublicBenchmarkRows, stripCodeFences, summarizeAgentRxCalibration, summarizeCodeTraceCalibration, traceStoreEvidenceResolver };
 //# sourceMappingURL=index.d.ts.map

package/dist/analyst/index.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"index.d.ts","names":[],"sources":["../../src/analyst/adapters.ts","../../src/analyst/benchmark-agentrx-calibration.ts","../../src/analyst/benchmark-dataset-types.ts","../../src/analyst/benchmark-dataset-agentrx.ts","../../src/analyst/benchmark-dataset-codetrace.ts","../../src/analyst/benchmark-dataset-utils.ts","../../src/analyst/benchmark-verification-outcome.ts","../../src/analyst/benchmark-verification-artifacts.ts","../../src/analyst/benchmark-public-types.ts","../../src/analyst/benchmark-public-adapters.ts","../../src/analyst/benchmark-public-data.ts","../../src/analyst/benchmark-public-model.ts","../../src/analyst/benchmark-public-rlm.ts","../../src/analyst/benchmark-comparison.ts","../../src/analyst/benchmark-public-calibration.ts","../../src/analyst/benchmark-command-artifact.ts","../../src/analyst/benchmark-command-result.ts","../../src/analyst/benchmark-command.ts","../../src/analyst/benchmark-implementation.ts","../../src/analyst/benchmark-report.ts","../../src/analyst/parse-tolerant.ts","../../src/analyst/proposal-findings.ts"],"mappings":";;;;;;;;;;;;iBA4CgB,aAAa,GAAG,WAAgB;UAe/B,oBAAoB;EACnC;EACA;EACA,UAAU,mBAAmB;;;;;EAK7B,UAAU,KAAK,cAAc;;iBAGf,sBAAsB,KAAK,MAAM,oBAAoB,OAAO,QAAQ;UAwEnE;EACf;EACA;EACA,SAAS;;EAET;;iBAGc,uBAAuB,OAAM,uBAA4B,QAAQ;UAiEhE;EACf;EACA;EACA,OAAO;;EAEP,MAAM;;EAEN,OAAO;;EAEP;;iBAGc,mBAAmB,MAAM,mBAAmB,QAAQ;UAiDnD;EACf;EACA;;EAEA,UAAU,KAAK;;EAEf;;iBAGc,kCACd,OAAM,kCACL,QAAQ;;;cC5RE;UAEI;EACf;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;;UAGe;EACf;EACA;EACA;EACA,SAAS;;iBAGK,4BACd,QAAQ,wBACR,2BACC;iBAkBa,iCAAiC,SAAS;;;KCtD9C;UAEK;EACf,YAAY;EACZ;EACA;EACA;EACA;EACA;;UAGe;EACf,eAAe;EACf,mBAAmB;EACnB;IAAe,YAAY;IAAY;;EACvC,wBAAwB;EACxB;EACA;EACA;;UAGe;EACf,UAAU;EACV;EACA;EACA;EACA;;UAGe;EACf,UAAU;EACV,mBAAmB;EACnB;EACA;EACA;EACA;EACA;;UAGe;EACf;EACA;EACA;EACA;;UAGe;EACf;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;;UAGe;EACf;EACA;EACA;EACA,iBAAiB;;KAGP,0CAEC,sCACA,iCACA;UAEI;EACf;EACA;EACA;EACA;;EAEA;EACA;EACA;EACA;EACA;EACA;EACA,oCAAoC;;UAGrB;EACf,eAAe;EACf,WAAW,sBAAsB;;KAGvB;UAEK;;;;;EAKf,WAAW;;UAGI,kCACP,yBACN;UAEa,oCAAoC;EACnD;;EAEA;;UAGe,yCAAyC;EACxD;EACA;EACA;EACA;;UAGe,2CACP,kCACN;;;iBCtGY,qBAAqB,QACnC,KAAK,YACL,OAAO,QACP,UAAS,8BACR,qBAAqB;;iBAoGR,6BACd,mBAAmB,YACnB,iBACA,UAAS,mCACR;iBA6Ea,yBAAyB;;iBAoNzB,iBAAiB;;;iBC7YjB,mBAAmB,QACjC,KAAK,mBACL,OAAO,QACP,UAAS,4BACR,qBAAqB;;iBAwFR,gCACd,2BACA,aAAa,uBACb,UAAS,qCACR;;;iBClHa,wBAAwB;;;KCA5B;UAEK;EACf;EACA;EACA,QAAQ;;UAGO;EACf,QAAQ;EACR;EAKA;IAAe;IAAe;;EAC9B,SAAS;EACT;EACA;EACA;EACA;;UAGe;EACf;EACA;;iBA8Ec,yBACd,gBAAgB,2BACf;;;cChGU;KAED;UAEK;EACf,MAAM;EACN;EACA;EACA;EACA;EACA;;UAGe;EACf;EACA;EACA,SAAS;EACT;EACA;EACA;EACA;EACA;EACA,OAAO;EACP,cAAc;EACd,UAAU,OAAO;;UAGF;EACf,UAAU;EACV,SAAS;EACT,OAAO,MAAM;IAA6B;;;iBAatB,mCAAmC;EACvD;EACA,KAAK;EACL;IACE,QAAQ;iBAwHI,kCACd,kBACA,iBACA,WAAW,6BACX;;;KChLU;UAEK;EACf;EACA;EACA;EACA;EACA;;EAEA,UAAU;;EAEV;EACA;IACE,SAAS;IACT;IACA;IACA;IACA;;EAEF,aAAa;EACb;IACE;IACA;;;EAGF,mBAAmB;;UAGJ;EACf,OAAO,qBAAqB;EAC5B;EACA;EACA;EACA,YAAY;IACV;IACA;IACA;;EAEF,uBAAuB;EACvB,WAAW;;UAGI;EACf;EACA;EACA,QAAQ;;UAGO;EACf,OAAO;EACP,OAAO;EACP,OAAO;EACP,YAAY;EACZ,QAAQ;;UAGO;EACf;EACA;EACA;EACA;EACA;EACA;EACA,QAAQ;EACR,UAAU;;;;iBC5DI,8BAA8B,uBAAuB;iBAiBrD,6BACd,SAAS,+BACT,sBACA,mBAAmB,kBACnB,oBACC;;;iBCgBmB,wBACpB,eACC,QAAQ,MAAM;iBA2BD,0BACd,SAAS,+BACT,eAAe,2BACf;EAAW;EAAe;IACzB,MAAM;iBAwBO,6BACd,SAAS,+BACT,eAAe,4BACd;iBAyCa,+BACd,SAAS,+BACT,iBAAiB,2BACjB,mBAAmB,2BACnB,eACC;iBAcmB,8BAA8B;EAClD,SAAS;EACT;EACA;EACA;EACA;EACA;EACA;IACE,QAAQ;;;;iBC7HI,kCACd,SAAS,+BACT,QAAQ,oCACP,uBAAuB;iBAofV,8BAA8B,SAAS;cAoE1C;;;;iBCljBG,+BACd,SAAS,+BACT,QAAQ,oCACP,uBAAuB;;;KC5Dd;UAoBK;EACf,QAAQ;EACR;;EAEA;;EAEA;;EAEA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;;UAGe;EACf;EACA;EACA,SAAS;;iBASK,sBACd,QAAQ,wBACR;EACE;EACA;EACA;EACA;EACA;IAED;;;UCnEc;EACf;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;;EAEA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;;UAGe;EACf;EACA;EACA,SAAS;;iBAGK,8BACd,QAAQ,yBACP;iBAca,mCAAmC,SAAS;;;UCrC3C;EACf;EACA;EACA;IACE,SAAS;IACT;IACA;IACA;IACA;IACA,YAAY;MAAQ;MAAiB;MAAsB;;IAC3D,uBAAuB;IACvB,0BAA0B;IAC1B;MACE;MACA;MACA;MACA,QAAQ;;IAEV;MACE;MACA;MACA;MACA;MACA;MACA;MACA;MACA;MACA;MACA;;;EAGJ,QAAQ;EACR,aAAa;EACb,uBAAuB;EACvB,qBAAqB;;UAGN;EACf;EACA;EACA;EACA;IACE;IACA;IACA;;;UAIa;EACf;IACE,SAAS;IACT;IACA;IACA;MACE;MACA;MACA;;IAEF;IACA;IACA;IACA;IACA;IACA;IACA;IACA;IACA;IACA;;EAEF;IACE;IACA;IACA;IACA,YAAY;MAAQ;MAAiB;MAAsB;;IAC3D;IACA;;;UAIa;EACf;EACA;EACA;EACA;EACA,UAAU;;UAGK;EACf;EACA;EACA;EACA;IACE;IACA;IACA;IACA;IACA;IACA;;EAEF;EACA;IACE;IACA;IACA;;EAEF;IACE;IACA;IACA;IACA;IACA;IACA;;;UAIa;EACf;EACA;EACA;EACA,aAAa;EACb;;cAGW;cACA;cACA;cACA;;;iBCxHS,6BACpB,eACC,QAAQ;;;UCyDM;EACf,uBACE,SAAS,+BACT,QAAQ,sCACL,uBAAuB;;UAGb;EACf,SAAS;EACT;EACA;EACA;EACA;EACA;EACA;EACA,OAAO;EACP;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;;iBAKoB,2BACpB,yBACA,MAAK,OAAO,YACZ,eAAc,sCACb;cA8QU;;;cC7XA;cAEA;cAEA;cAOA;;;;;;;;cAUA;cAEA;cAGA;cAGA;cAmFA;iBAGG;iBAIA;;;iBCpHA,+BACd,QAAQ,wBACR,uBAAsB;;;;;;;;;;;;;;;iBCQR,gBAAgB;;;;;iBAgBhB,WAAW;;;;;;;iBAeX,oBAAoB;;;;iBCXpB,kBAAkB,mBAAmB,WAAW;;;;;;iBAShD,uBACd,mBACA,mBACC,cAAc"}
+{"version":3,"file":"index.d.ts","names":[],"sources":["../../src/analyst/adapters.ts","../../src/analyst/benchmark-agentrx-calibration.ts","../../src/analyst/benchmark-dataset-types.ts","../../src/analyst/benchmark-dataset-agentrx.ts","../../src/analyst/benchmark-dataset-codetrace.ts","../../src/analyst/benchmark-dataset-utils.ts","../../src/analyst/benchmark-verification-outcome.ts","../../src/analyst/benchmark-verification-artifacts.ts","../../src/analyst/benchmark-public-types.ts","../../src/analyst/benchmark-public-adapters.ts","../../src/analyst/benchmark-public-data.ts","../../src/analyst/benchmark-public-prompt.ts","../../src/analyst/benchmark-public-model.ts","../../src/analyst/benchmark-public-rlm.ts","../../src/analyst/benchmark-comparison.ts","../../src/analyst/benchmark-public-calibration.ts","../../src/analyst/benchmark-command-artifact.ts","../../src/analyst/benchmark-command-result.ts","../../src/analyst/benchmark-command.ts","../../src/analyst/benchmark-implementation.ts","../../src/analyst/benchmark-report.ts","../../src/analyst/parse-tolerant.ts","../../src/analyst/proposal-findings.ts"],"mappings":";;;;;;;;;;;;iBA4CgB,aAAa,GAAG,WAAgB;UAe/B,oBAAoB;EACnC;EACA;EACA,UAAU,mBAAmB;;;;;EAK7B,UAAU,KAAK,cAAc;;iBAGf,sBAAsB,KAAK,MAAM,oBAAoB,OAAO,QAAQ;UAwEnE;EACf;EACA;EACA,SAAS;;EAET;;iBAGc,uBAAuB,OAAM,uBAA4B,QAAQ;UAiEhE;EACf;EACA;EACA,OAAO;;EAEP,MAAM;;EAEN,OAAO;;EAEP;;iBAGc,mBAAmB,MAAM,mBAAmB,QAAQ;UAiDnD;EACf;EACA;;EAEA,UAAU,KAAK;;EAEf;;iBAGc,kCACd,OAAM,kCACL,QAAQ;;;cC5RE;UAEI;EACf;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;;UAGe;EACf;EACA;EACA;EACA,SAAS;;iBAGK,4BACd,QAAQ,wBACR,2BACC;iBAkBa,iCAAiC,SAAS;;;KCtD9C;UAEK;EACf,YAAY;EACZ;EACA;EACA;EACA;EACA;;UAGe;EACf,eAAe;EACf,mBAAmB;EACnB;IAAe,YAAY;IAAY;;EACvC,wBAAwB;EACxB;EACA;EACA;;UAGe;EACf,UAAU;EACV;EACA;EACA;EACA;;UAGe;EACf,UAAU;EACV,mBAAmB;EACnB;EACA;EACA;EACA;EACA;;UAGe;EACf;EACA;EACA;EACA;;UAGe;EACf;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;;UAGe;EACf;EACA;EACA;EACA,iBAAiB;;KAGP,0CAEC,sCACA,iCACA;UAEI;EACf;EACA;EACA;EACA;;EAEA;EACA;EACA;EACA;EACA;EACA;EACA,oCAAoC;;UAGrB;EACf,eAAe;EACf,WAAW,sBAAsB;;KAGvB;UAEK;;;;;EAKf,WAAW;;UAGI,kCACP,yBACN;UAEa,oCAAoC;EACnD;;EAEA;;UAGe,yCAAyC;EACxD;EACA;EACA;EACA;;UAGe,2CACP,kCACN;;;iBCtGY,qBAAqB,QACnC,KAAK,YACL,OAAO,QACP,UAAS,8BACR,qBAAqB;;iBAoGR,6BACd,mBAAmB,YACnB,iBACA,UAAS,mCACR;iBA6Ea,yBAAyB;;iBAoNzB,iBAAiB;;;iBC7YjB,mBAAmB,QACjC,KAAK,mBACL,OAAO,QACP,UAAS,4BACR,qBAAqB;;iBAwFR,gCACd,2BACA,aAAa,uBACb,UAAS,qCACR;;;iBClHa,wBAAwB;;;KCA5B;UAEK;EACf;EACA;EACA,QAAQ;;UAGO;EACf,QAAQ;EACR;EAKA;IAAe;IAAe;;EAC9B,SAAS;EACT;EACA;EACA;EACA;;UAGe;EACf;EACA;;iBA8Ec,yBACd,gBAAgB,2BACf;;;cChGU;KAED;UAEK;EACf,MAAM;EACN;EACA;EACA;EACA;EACA;;UAGe;EACf;EACA;EACA,SAAS;EACT;EACA;EACA;EACA;EACA;EACA,OAAO;EACP,cAAc;EACd,UAAU,OAAO;;UAGF;EACf,UAAU;EACV,SAAS;EACT,OAAO,MAAM;IAA6B;;;iBAatB,mCAAmC;EACvD;EACA,KAAK;EACL;IACE,QAAQ;iBAwHI,kCACd,kBACA,iBACA,WAAW,6BACX;;;KChLU;UAEK;EACf;EACA;EACA;EACA;EACA;;EAEA,UAAU;;EAEV;EACA;IACE,SAAS;IACT;IACA;IACA;IACA;;EAEF,aAAa;EACb;IACE;IACA;;;EAGF,mBAAmB;;UAGJ;EACf,OAAO,qBAAqB;EAC5B;EACA;EACA;EACA,YAAY;IACV;IACA;IACA;;EAEF,uBAAuB;EACvB,WAAW;;UAGI;EACf;EACA;EACA,QAAQ;;UAGO;EACf,OAAO;EACP,OAAO;EACP,OAAO;EACP,YAAY;EACZ,QAAQ;;UAGO;EACf;EACA;EACA;EACA;EACA;EACA;EACA,QAAQ;EACR,UAAU;;;;;;;;;;UCnDK;EACf;EACA;;EAEA;EACA;EACA,UAAU;EACV;EACA;EACA;EACA;EACA,WAAW;;;;;;;;;UAUI;EACf;EACA;;EAEA,kCAAkC;;EAElC;;EAEA;;iBAGc,8BAA8B,uBAAuB;iBAiB/C,6BAA6B;EACjD,SAAS;EACT;EACA,mBAAmB;EACnB;EACA,OAAO;EACP,SAAS;IACP;EAAU,UAAU;EAAkB,aAAa;;;;;;;;;iBAqJjC,6BAA6B;EACjD;EACA,iBAAiB;EACjB,OAAO;EACP;EACA;EACA,SAAS;IACP;EAAU,UAAU;EAAkB,aAAa;;;;iBCvLjC,wBACpB,eACC,QAAQ,MAAM;iBA2BD,0BACd,SAAS,+BACT,eAAe,2BACf;EAAW;EAAe;IACzB,MAAM;iBAwBO,6BACd,SAAS,+BACT,eAAe,4BACd;iBAyCa,+BACd,SAAS,+BACT,iBAAiB,2BACjB,mBAAmB,2BACnB,eACC;iBAcmB,8BAA8B;EAClD,SAAS;EACT;EACA;EACA;EACA;EACA;EACA;IACE,QAAQ;;;;;;;cCxKC;;;;cAKA;cAMA;;iBAyFG,4BAA4B,SAAS;;iBAerC,+BAA+B,SAAS;;;;iBAaxC,8BAA8B,SAAS;;;;iBCxEvC,kCACd,SAAS,+BACT,QAAQ,oCACP,uBAAuB;;;;iBCtCV,+BACd,SAAS,+BACT,QAAQ,oCACP,uBAAuB;;;KC7Bd;UAoBK;EACf,QAAQ;EACR;;EAEA;;EAEA;;EAEA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;;UAGe;EACf;EACA;EACA,SAAS;;iBASK,sBACd,QAAQ,wBACR;EACE;EACA;EACA;EACA;EACA;IAED;;;UCnEc;EACf;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;;EAEA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;;UAGe;EACf;EACA;EACA,SAAS;;iBAGK,8BACd,QAAQ,yBACP;iBAca,mCAAmC,SAAS;;;UCrC3C;EACf;EACA;EACA;IACE,SAAS;IACT;IACA;IACA;IACA;IACA,YAAY;MAAQ;MAAiB;MAAsB;;IAC3D,uBAAuB;IACvB,0BAA0B;IAC1B;MACE;MACA;MACA;MACA,QAAQ;;IAEV;MACE;MACA;MACA;MACA;MACA;MACA;MACA;MACA;MACA;MACA;;;EAGJ,QAAQ;EACR,aAAa;EACb,uBAAuB;EACvB,qBAAqB;;UAGN;EACf;EACA;EACA;EACA;IACE;IACA;IACA;;;UAIa;EACf;IACE,SAAS;IACT;IACA;IACA;MACE;MACA;MACA;;IAEF;IACA;IACA;IACA;IACA;IACA;IACA;IACA;IACA;IACA;;EAEF;IACE;IACA;IACA;IACA,YAAY;MAAQ;MAAiB;MAAsB;;IAC3D;IACA;;;UAIa;EACf;EACA;EACA;EACA;EACA,UAAU;;UAGK;EACf;EACA;EACA;EACA;IACE;IACA;IACA;IACA;IACA;IACA;;EAEF;EACA;IACE;IACA;IACA;;EAEF;IACE;IACA;IACA;IACA;IACA;IACA;;;UAIa;EACf;EACA;EACA;EACA,aAAa;EACb;;cAGW;cACA;cACA;cACA;;;iBCxHS,6BACpB,eACC,QAAQ;;;UC0DM;EACf,uBACE,SAAS,+BACT,QAAQ,sCACL,uBAAuB;;;;;;;;;;KAWlB;UAEK;EACf,SAAS;EACT,SAAS;EACT;EACA;EACA;EACA;EACA;EACA;EACA,OAAO;EACP;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;;iBAKoB,2BACpB,yBACA,MAAK,OAAO,YACZ,eAAc,sCACb;cAmRU;;;cC9YA;cAEA;cAEA;cAOA;;;;;;;;cAUA;cAEA;cAGA;cAGA;cAoFA;iBAGG;iBAIA;;;iBCrHA,+BACd,QAAQ,wBACR,uBAAsB;;;;;;;;;;;;;;;iBCQR,gBAAgB;;;;;iBAgBhB,WAAW;;;;;;;iBAeX,oBAAoB;;;;iBCXpB,kBAAkB,mBAAmB,WAAW;;;;;;iBAShD,uBACd,mBACA,mBACC,cAAc"}

package/dist/analyst/index.js

@@ -6,7 +6,7 @@ import { n as isProposalFinding, t as assertProposalFindings } from "../proposal
 import { a as RunCritic, c as buildSkillUsageReport, d as defaultIsMaterial, f as diffFindings, h as defineTraceAnalyst, i as runSemanticConceptJudge, l as emitSkillUsageFindings, m as defineCustomAnalyst, n as SEMANTIC_CONCEPT_JUDGE_VERSION, o as SKILL_USAGE_ANALYST, s as SkillUsageAnalyst, u as FindingsStore } from "../semantic-concept-judge-BuIJ9IfB.js";
 import { t as createDspyRlmTraceEngine } from "../dspy-rlm-engine-DTkVyDX-.js";
 import { a as scoreAnalystFindings, n as runAnalystBenchmark, r as traceStoreEvidenceResolver, t as registryBenchmarkRunner } from "../benchmark-CYtcIF2V.js";
-import { A as ANALYST_BENCHMARK_IMPLEMENTATION_FILES, B as summarizeAgentRxCalibration, C as ANALYST_BENCHMARK_DEPENDENCY_LOCK_DIGEST_ALGORITHM, D as ANALYST_BENCHMARK_EVIDENCE_IMPLEMENTATION_SHA256, E as ANALYST_BENCHMARK_EVIDENCE_DEPENDENCY_LOCK_SHA256, F as ANALYST_BENCHMARK_LOCAL_RECEIPT_FILE, G as normalizeAgentRxCategory, H as codeTracerPredictionsToFindings, I as ANALYST_BENCHMARK_MANIFEST_FILE, K as roundAgentRxStep, L as ANALYST_BENCHMARK_OBSERVATIONS_FILE, M as analystBenchmarkDependencyLockDigest, N as analystBenchmarkImplementationDigest, O as ANALYST_BENCHMARK_EVIDENCE_PACKAGE_VERSION, P as ANALYST_BENCHMARK_COST_LEDGER_FILE, R as AGENT_RX_UPSTREAM_REVISION, S as emptyPublicBenchmarkRunner, T as ANALYST_BENCHMARK_DEPENDENCY_LOCK_SHA256, U as agentRxBenchmarkCase, V as codeTraceBenchCase, W as agentRxPredictionsToFindings, _ as DEFAULT_MAX_VERIFICATION_ARTIFACT_BYTES, a as renderCodeTraceCalibrationMarkdown, b as parseVerificationOutcome, c as createPublicBenchmarkRlmRunner, d as publicBenchmarkProtocolSha256, f as loadPublicBenchmarkRows, g as selectPublicBenchmarkRows, h as publicBenchmarkSelectionReport, i as readAnalystBenchmarkArtifact, j as ANALYST_BENCHMARK_IMPLEMENTATION_SHA256, k as ANALYST_BENCHMARK_IMPLEMENTATION_DIGEST_ALGORITHM, l as CODE_TRACE_BENCH_ANALYST_PROMPT, m as publicBenchmarkDistributions, n as runAnalystBenchmarkCommand, o as summarizeCodeTraceCalibration, p as preparePublicAnalystBenchmark, q as normalizeBenchmarkLabel, r as renderAnalystBenchmarkMarkdown, s as compareAnalystRunners, t as ANALYST_BENCHMARK_HELP, u as createPublicBenchmarkDirectRunner, v as appendVerificationArtifactsToOtlp, w as ANALYST_BENCHMARK_DEPENDENCY_LOCK_FILES, x as adaptPublicBenchmarkFindings, y as loadCodeTraceVerificationArtifacts, z as renderAgentRxCalibrationMarkdown } from "../benchmark-command-DzUZJl8M.js";
+import { A as ANALYST_BENCHMARK_DEPENDENCY_LOCK_SHA256, B as ANALYST_BENCHMARK_LOCAL_RECEIPT_FILE, C as publicBenchmarkSystemPrompt, D as parseVerificationOutcome, E as loadCodeTraceVerificationArtifacts, F as ANALYST_BENCHMARK_IMPLEMENTATION_FILES, G as summarizeAgentRxCalibration, H as ANALYST_BENCHMARK_OBSERVATIONS_FILE, I as ANALYST_BENCHMARK_IMPLEMENTATION_SHA256, J as agentRxBenchmarkCase, K as codeTraceBenchCase, L as analystBenchmarkDependencyLockDigest, M as ANALYST_BENCHMARK_EVIDENCE_IMPLEMENTATION_SHA256, N as ANALYST_BENCHMARK_EVIDENCE_PACKAGE_VERSION, O as ANALYST_BENCHMARK_DEPENDENCY_LOCK_DIGEST_ALGORITHM, P as ANALYST_BENCHMARK_IMPLEMENTATION_DIGEST_ALGORITHM, Q as normalizeBenchmarkLabel, R as analystBenchmarkImplementationDigest, S as publicBenchmarkRlmInstructions, T as appendVerificationArtifactsToOtlp, U as AGENT_RX_UPSTREAM_REVISION, V as ANALYST_BENCHMARK_MANIFEST_FILE, W as renderAgentRxCalibrationMarkdown, X as normalizeAgentRxCategory, Y as agentRxPredictionsToFindings, Z as roundAgentRxStep, _ as expandCodeTraceFailureBlocks, a as renderCodeTraceCalibrationMarkdown, b as MAX_INCORRECT_BLOCK_STEPS, c as createPublicBenchmarkRlmRunner, d as preparePublicAnalystBenchmark, f as publicBenchmarkDistributions, g as emptyPublicBenchmarkRunner, h as adaptPublicBenchmarkFindings, i as readAnalystBenchmarkArtifact, j as ANALYST_BENCHMARK_EVIDENCE_DEPENDENCY_LOCK_SHA256, k as ANALYST_BENCHMARK_DEPENDENCY_LOCK_FILES, l as createPublicBenchmarkDirectRunner, m as selectPublicBenchmarkRows, n as runAnalystBenchmarkCommand, o as summarizeCodeTraceCalibration, p as publicBenchmarkSelectionReport, q as codeTracerPredictionsToFindings, r as renderAnalystBenchmarkMarkdown, s as compareAnalystRunners, t as ANALYST_BENCHMARK_HELP, u as loadPublicBenchmarkRows, v as CODE_TRACE_BENCH_ANALYST_PROMPT, w as DEFAULT_MAX_VERIFICATION_ARTIFACT_BYTES, x as publicBenchmarkProtocolSha256, y as MAX_INCORRECT_BLOCKS, z as ANALYST_BENCHMARK_COST_LEDGER_FILE } from "../benchmark-command-xvi2liH7.js";
 //#region src/analyst/adapters.ts
 /**
 * Adapter factories — lift each existing agent-eval primitive into the
@@ -294,6 +294,6 @@ function createSemanticConceptJudgeAdapter(opts = {}) {
 	};
 }
 //#endregion
-export { AGENT_RX_UPSTREAM_REVISION, ANALYST_BENCHMARK_COST_LEDGER_FILE, ANALYST_BENCHMARK_DEPENDENCY_LOCK_DIGEST_ALGORITHM, ANALYST_BENCHMARK_DEPENDENCY_LOCK_FILES, ANALYST_BENCHMARK_DEPENDENCY_LOCK_SHA256, ANALYST_BENCHMARK_EVIDENCE_DEPENDENCY_LOCK_SHA256, ANALYST_BENCHMARK_EVIDENCE_IMPLEMENTATION_SHA256, ANALYST_BENCHMARK_EVIDENCE_PACKAGE_VERSION, ANALYST_BENCHMARK_HELP, ANALYST_BENCHMARK_IMPLEMENTATION_DIGEST_ALGORITHM, ANALYST_BENCHMARK_IMPLEMENTATION_FILES, ANALYST_BENCHMARK_IMPLEMENTATION_SHA256, ANALYST_BENCHMARK_LOCAL_RECEIPT_FILE, ANALYST_BENCHMARK_MANIFEST_FILE, ANALYST_BENCHMARK_OBSERVATIONS_FILE, ANALYST_SEVERITIES, AnalystRegistry, CODE_TRACE_BENCH_ANALYST_PROMPT, CONTROL_INTEGRITY_ANALYST, ControlIntegrityAnalyst, DEFAULT_MAX_VERIFICATION_ARTIFACT_BYTES, DEFAULT_TRACE_ANALYST_KINDS, DEFAULT_TRACE_ANALYST_LIMITS, ExactAnalystRunExecutionError, FAILURE_MODE_KIND_SPEC, FINDING_SUBJECT_GRAMMAR_PROMPT, FINDING_SUBJECT_KINDS, FINDING_SUBJECT_SYNTAX, FindingSubjectStringSchema, FindingsStore, IMPROVEMENT_KIND_SPEC, KIND_EXPECTED_SUBJECTS, KNOWLEDGE_GAP_KIND_SPEC, KNOWLEDGE_POISONING_KIND_SPEC, RAW_FINDING_SCHEMA_PROMPT, RawAnalystEvidenceSchema, RawAnalystFindingSchema, SKILL_USAGE_ANALYST, SkillUsageAnalyst, adaptPublicBenchmarkFindings, agentRxBenchmarkCase, agentRxPredictionsToFindings, analystBenchmarkDependencyLockDigest, analystBenchmarkImplementationDigest, appendVerificationArtifactsToOtlp, assertExactRegistryRunOpts, assertProposalFindings, behavioralAnalyst, buildDefaultAnalystRegistry, buildSkillUsageReport, buildTraceToolsForGroup, codeTraceBenchCase, codeTracerPredictionsToFindings, coerceJson, coerceToFindingRows, compareAnalystRunners, computeFindingId, createChatClient, createDspyRlmTraceEngine, createJudgeAdapter, createPublicBenchmarkDirectRunner, createPublicBenchmarkRlmRunner, createRunCriticAdapter, createSemanticConceptJudgeAdapter, createTraceAnalyst, createVerifierAdapter, defaultIsMaterial, defineCustomAnalyst, defineTraceAnalyst, deriveEfficiencyFindings, diffFindings, emitControlIntegrityFindings, emitSkillUsageFindings, emptyPublicBenchmarkRunner, evidenceRefsFromRawFinding, findingSubjectGrammarPromptFor, isProposalFinding, liftSeverity, loadCodeTraceVerificationArtifacts, loadPublicBenchmarkRows, makeFinding, makeProposalFinding, normalizeAgentRxCategory, normalizeBenchmarkLabel, parseFindingSubject, parseRawFinding, parseVerificationOutcome, preparePublicAnalystBenchmark, publicBenchmarkDistributions, publicBenchmarkProtocolSha256, publicBenchmarkSelectionReport, readAnalystBenchmarkArtifact, registryBenchmarkRunner, renderAgentRxCalibrationMarkdown, renderAnalystBenchmarkMarkdown, renderCodeTraceCalibrationMarkdown, renderFindingSubject, renderPriorFindings, renderUpstreamFindings, resolveTraceAnalystLimits, roundAgentRxStep, runAnalystBenchmark, runAnalystBenchmarkCommand, runTraceAnalyst, scoreAnalystFindings, selectPublicBenchmarkRows, stripCodeFences, summarizeAgentRxCalibration, summarizeCodeTraceCalibration, traceStoreEvidenceResolver };
+export { AGENT_RX_UPSTREAM_REVISION, ANALYST_BENCHMARK_COST_LEDGER_FILE, ANALYST_BENCHMARK_DEPENDENCY_LOCK_DIGEST_ALGORITHM, ANALYST_BENCHMARK_DEPENDENCY_LOCK_FILES, ANALYST_BENCHMARK_DEPENDENCY_LOCK_SHA256, ANALYST_BENCHMARK_EVIDENCE_DEPENDENCY_LOCK_SHA256, ANALYST_BENCHMARK_EVIDENCE_IMPLEMENTATION_SHA256, ANALYST_BENCHMARK_EVIDENCE_PACKAGE_VERSION, ANALYST_BENCHMARK_HELP, ANALYST_BENCHMARK_IMPLEMENTATION_DIGEST_ALGORITHM, ANALYST_BENCHMARK_IMPLEMENTATION_FILES, ANALYST_BENCHMARK_IMPLEMENTATION_SHA256, ANALYST_BENCHMARK_LOCAL_RECEIPT_FILE, ANALYST_BENCHMARK_MANIFEST_FILE, ANALYST_BENCHMARK_OBSERVATIONS_FILE, ANALYST_SEVERITIES, AnalystRegistry, CODE_TRACE_BENCH_ANALYST_PROMPT, CONTROL_INTEGRITY_ANALYST, ControlIntegrityAnalyst, DEFAULT_MAX_VERIFICATION_ARTIFACT_BYTES, DEFAULT_TRACE_ANALYST_KINDS, DEFAULT_TRACE_ANALYST_LIMITS, ExactAnalystRunExecutionError, FAILURE_MODE_KIND_SPEC, FINDING_SUBJECT_GRAMMAR_PROMPT, FINDING_SUBJECT_KINDS, FINDING_SUBJECT_SYNTAX, FindingSubjectStringSchema, FindingsStore, IMPROVEMENT_KIND_SPEC, KIND_EXPECTED_SUBJECTS, KNOWLEDGE_GAP_KIND_SPEC, KNOWLEDGE_POISONING_KIND_SPEC, MAX_INCORRECT_BLOCKS, MAX_INCORRECT_BLOCK_STEPS, RAW_FINDING_SCHEMA_PROMPT, RawAnalystEvidenceSchema, RawAnalystFindingSchema, SKILL_USAGE_ANALYST, SkillUsageAnalyst, adaptPublicBenchmarkFindings, agentRxBenchmarkCase, agentRxPredictionsToFindings, analystBenchmarkDependencyLockDigest, analystBenchmarkImplementationDigest, appendVerificationArtifactsToOtlp, assertExactRegistryRunOpts, assertProposalFindings, behavioralAnalyst, buildDefaultAnalystRegistry, buildSkillUsageReport, buildTraceToolsForGroup, codeTraceBenchCase, codeTracerPredictionsToFindings, coerceJson, coerceToFindingRows, compareAnalystRunners, computeFindingId, createChatClient, createDspyRlmTraceEngine, createJudgeAdapter, createPublicBenchmarkDirectRunner, createPublicBenchmarkRlmRunner, createRunCriticAdapter, createSemanticConceptJudgeAdapter, createTraceAnalyst, createVerifierAdapter, defaultIsMaterial, defineCustomAnalyst, defineTraceAnalyst, deriveEfficiencyFindings, diffFindings, emitControlIntegrityFindings, emitSkillUsageFindings, emptyPublicBenchmarkRunner, evidenceRefsFromRawFinding, expandCodeTraceFailureBlocks, findingSubjectGrammarPromptFor, isProposalFinding, liftSeverity, loadCodeTraceVerificationArtifacts, loadPublicBenchmarkRows, makeFinding, makeProposalFinding, normalizeAgentRxCategory, normalizeBenchmarkLabel, parseFindingSubject, parseRawFinding, parseVerificationOutcome, preparePublicAnalystBenchmark, publicBenchmarkDistributions, publicBenchmarkProtocolSha256, publicBenchmarkRlmInstructions, publicBenchmarkSelectionReport, publicBenchmarkSystemPrompt, readAnalystBenchmarkArtifact, registryBenchmarkRunner, renderAgentRxCalibrationMarkdown, renderAnalystBenchmarkMarkdown, renderCodeTraceCalibrationMarkdown, renderFindingSubject, renderPriorFindings, renderUpstreamFindings, resolveTraceAnalystLimits, roundAgentRxStep, runAnalystBenchmark, runAnalystBenchmarkCommand, runTraceAnalyst, scoreAnalystFindings, selectPublicBenchmarkRows, stripCodeFences, summarizeAgentRxCalibration, summarizeCodeTraceCalibration, traceStoreEvidenceResolver };
 
 //# sourceMappingURL=index.js.map