@tangle-network/agent-eval 0.130.1 → 0.131.1

package/CHANGELOG.md

@@ -4,6 +4,24 @@ All notable changes to `@tangle-network/agent-eval` and its sibling `agent-eval-
 
 ---
 
+## [0.131.1] - 2026-07-27 - worker-safe profile cells
+
+### Added
+
+- `@tangle-network/agent-eval/profile-cell` exports the portable agent-profile identity API without importing local transcript readers or Node-only SQLite support.
+
+## [0.131.0] - 2026-07-27 - profile improvement measurements
+
+### Added
+
+- Profile improvement measurement contracts in `/contract` that seal an exact profile diff and held-out task set, execute paired host-owned cells, and recompute the published score, uncertainty, cost, latency, and decision from complete receipts.
+
+### Fixed
+
+- `maxConcurrency` now limits individual host executions across both experiment arms.
+- A failed or cancelled execution stops new work, cancels active siblings, and settles them before returning.
+- Final sufficiency uses the observed paired interval instead of a conservative baseline-only planning estimate.
+
 ## [0.130.1] - 2026-07-26 - safe DSPy disk caching
 
 ### Fixed

package/README.md

@@ -365,6 +365,7 @@ See [concepts](./docs/concepts.md), [customer paths](./docs/customer-journeys.md
 |---|---|
 | `@tangle-network/agent-eval/contract` | Define an evaluation, run it, improve with a custom candidate generator, and analyze runs. |
 | `@tangle-network/agent-eval/campaign` | Control campaigns, official optimization methods, comparisons, storage, and release rules. |
+| `@tangle-network/agent-eval/profile-cell` | Create and validate portable agent-profile identities. |
 | `@tangle-network/agent-eval/ledger-core` | Generic hash-chained append-only journal: idempotent append, chain verification, replay-to-projection, cross-process locking. |
 | `@tangle-network/agent-eval/reporting` | Statistical comparisons and report rendering. |
 | `@tangle-network/agent-eval/analyst` | Model-assisted failure analysis. |

package/dist/agent-profile-cell-CCm3l2v2.d.ts

@@ -0,0 +1,108 @@
+import { c as ValidationError } from "./errors-CEk209JS.js";
+import { AgentProfile } from "@tangle-network/agent-interface";
+//#region src/agent-profile-cell.d.ts
+type AgentProfileCellSchemaVersion = 'agent-profile-cell/v1';
+type AgentProfileJsonObject = {
+  [key: string]: AgentProfileJson;
+};
+type AgentProfileJson = string | number | boolean | null | AgentProfileJson[] | AgentProfileJsonObject;
+type AgentProfileDimensionValue = string | number | boolean | null;
+interface AgentProfileSource {
+  /** Runtime/profile contract being fingerprinted, e.g. `agent-interface-profile`. */
+  kind: string;
+  /** sha256 over the canonical source profile object. */
+  hash: string;
+}
+interface AgentProfileSourceInput {
+  kind: string;
+  /** Precomputed sha256 for callers that already sign their profile artifact. */
+  hash?: string;
+  /** Full canonical runtime profile; hashed and then discarded from the cell. */
+  profile?: AgentProfileJson;
+}
+interface AgentProfileHarness {
+  id: string;
+  version?: string;
+  hash?: string;
+}
+interface AgentProfileCellInput {
+  profileId: string;
+  sourceProfile: AgentProfileSourceInput;
+  harness?: AgentProfileHarness;
+  model?: string;
+  promptHash?: string;
+  dimensions?: Record<string, AgentProfileDimensionValue>;
+}
+interface AgentProfileCell {
+  schemaVersion: AgentProfileCellSchemaVersion;
+  cellId: string;
+  profileId: string;
+  sourceProfile: AgentProfileSource;
+  harness?: AgentProfileHarness;
+  model?: string;
+  promptHash?: string;
+  dimensions?: Record<string, AgentProfileDimensionValue>;
+}
+declare class AgentProfileCellValidationError extends ValidationError {
+  readonly path: string;
+  constructor(message: string, path?: string);
+}
+declare function buildAgentProfileCell(input: AgentProfileCellInput): Promise<AgentProfileCell>;
+declare function agentProfileCellHashMaterial(cell: AgentProfileCell): Omit<AgentProfileCell, 'cellId'>;
+/**
+ * Verify an `AgentProfileCell`'s `cellId` matches the sha256 of its hash-material fields, confirming the record has not been tampered with.
+ */
+declare function verifyAgentProfileCell(cell: AgentProfileCell): Promise<boolean>;
+declare function validateAgentProfileCell(input: unknown): AgentProfileCell;
+declare function requireAgentProfileCell(record: {
+  runId: string;
+  agentProfile?: AgentProfileCell;
+}): AgentProfileCell;
+declare function agentProfileCellKey(record: {
+  runId: string;
+  agentProfile?: AgentProfileCell;
+}): string;
+declare function assertRunAgentProfileCell(record: {
+  runId: string;
+  model: string;
+  promptHash: string;
+  agentProfile?: AgentProfileCell;
+}): Promise<AgentProfileCell>;
+declare function groupRunsByAgentProfileCell<T extends {
+  runId: string;
+  agentProfile?: AgentProfileCell;
+}>(records: readonly T[]): Map<string, T[]>;
+/** Canonical `sourceProfile.kind` values. Two products fingerprinting the
+ *  same canonical profile MUST use the same kind for their cells to share
+ *  `sourceProfile.hash`. Extend rather than create new strings — adding a
+ *  new kind is a deliberate cross-product schema change. */
+declare const AGENT_PROFILE_KINDS: {
+  /** A profile declared via `defineAgentProfile(...)` from
+   *  `@tangle-network/agent-interface`. The default kind for router-backed
+   *  and sandbox-backed products. */
+  readonly AGENT_INTERFACE_PROFILE: 'agent-interface-profile';
+};
+type AgentProfileKind = (typeof AGENT_PROFILE_KINDS)[keyof typeof AGENT_PROFILE_KINDS];
+/** Canonicalize an arbitrary value into `AgentProfileJson` by JSON
+ *  round-trip. Throws when the value contains anything not representable
+ *  as JSON (functions, BigInt, cycles) — non-portable profiles fail loud
+ *  rather than silently dropping fields. */
+declare function toAgentProfileJson(value: unknown): AgentProfileJson;
+/** Canonical AgentProfile shape required when deriving a stable cell id. */
+type AgentInterfaceProfileLike = AgentProfile & {
+  name: string;
+  version: string;
+};
+/** Higher-level helper that hard-codes the canonical
+ *  `agent-interface-profile` kind plus the JSON canonicalization. Equivalent
+ *  to calling `buildAgentProfileCell` with `profileId = \`${name}@${version}\``
+ *  and `sourceProfile = { kind: AGENT_INTERFACE_PROFILE, profile: <round-tripped> }`.
+ *
+ *  Use this from any product consuming an agent-interface `AgentProfile`; the
+ *  manual `buildAgentProfileCell` call is reserved for advanced cases
+ *  (custom kinds, pre-computed source hashes, alternate profileId
+ *  conventions). */
+declare function buildAgentInterfaceProfileCell(profile: AgentInterfaceProfileLike, input: Omit<AgentProfileCellInput, 'profileId' | 'sourceProfile'>): Promise<AgentProfileCell>;
+//#endregion
+export { verifyAgentProfileCell as C, validateAgentProfileCell as S, buildAgentInterfaceProfileCell as _, AgentProfileCellSchemaVersion as a, requireAgentProfileCell as b, AgentProfileHarness as c, AgentProfileKind as d, AgentProfileSource as f, assertRunAgentProfileCell as g, agentProfileCellKey as h, AgentProfileCellInput as i, AgentProfileJson as l, agentProfileCellHashMaterial as m, AgentInterfaceProfileLike as n, AgentProfileCellValidationError as o, AgentProfileSourceInput as p, AgentProfileCell as r, AgentProfileDimensionValue as s, AGENT_PROFILE_KINDS as t, AgentProfileJsonObject as u, buildAgentProfileCell as v, toAgentProfileJson as x, groupRunsByAgentProfileCell as y };
+//# sourceMappingURL=agent-profile-cell-CCm3l2v2.d.ts.map

package/dist/agent-profile-cell-CCm3l2v2.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"agent-profile-cell-CCm3l2v2.d.ts","names":[],"sources":["../src/agent-profile-cell.ts"],"mappings":";;;KAIY;KAEA;GAA4B,cAAc;;KAE1C,sDAKR,qBACA;KAEQ;UAEK;;EAEf;;EAEA;;UAGe;EACf;;EAEA;;EAEA,UAAU;;UAGK;EACf;EACA;EACA;;UAGe;EACf;EACA,eAAe;EACf,UAAU;EACV;EACA;EACA,aAAa,eAAe;;UAGb;EACf,eAAe;EACf;EACA;EACA,eAAe;EACf,UAAU;EACV;EACA;EACA,aAAa,eAAe;;cAGjB,wCAAwC;WAC1C;EACT,YAAY,iBAAiB;;iBAST,sBACpB,OAAO,wBACN,QAAQ;iBAMK,6BACd,MAAM,mBACL,KAAK;;;;iBASc,uBAAuB,MAAM,mBAAmB;iBAQtD,yBAAyB,iBAAiB;iBAqB1C,wBAAwB;EACtC;EACA,eAAe;IACb;iBAUY,oBAAoB;EAClC;EACA,eAAe;;iBAKK,0BAA0B;EAC9C;EACA;EACA;EACA,eAAe;IACb,QAAQ;iBAuBI,4BACd;EAAY;EAAe,eAAe;GAC1C,kBAAkB,MAAM,YAAY;;;;;cA0NzB;;;;WAIX;;KAGU,2BAA2B,kCAAkC;;;;;iBAMzD,mBAAmB,iBAAiB;;KAoBxC,4BAA4B;EAAiB;EAAc;;;;;;;;;;;iBAWjD,+BACpB,SAAS,2BACT,OAAO,KAAK,wDACX,QAAQ"}

package/dist/analyst/index.d.ts

@@ -1,10 +1,10 @@
 import { a as MultiLayerVerifier, l as VerifyOptions, o as Severity } from "../multi-layer-verifier-BHY1gWAc.js";
-import { C as FindingSubjectKind, D as parseFindingSubject, E as findingSubjectGrammarPromptFor, F as createAnalystAi, H as SemanticConceptJudgeInput, O as renderFindingSubject, P as CreateAnalystAiConfig, S as FindingSubject, T as KIND_EXPECTED_SUBJECTS, U as SemanticConceptJudgeOptions, Y as RunTrace, _ as defaultIsMaterial, a as SkillUsageScanConfig, b as FINDING_SUBJECT_KINDS, c as DEFAULT_TRACE_ANALYST_KINDS, d as IMPROVEMENT_KIND_SPEC, f as FAILURE_MODE_KIND_SPEC, g as PersistedFinding, h as FindingsStore, i as SkillUsageReport, k as BehavioralMetrics, l as KNOWLEDGE_POISONING_KIND_SPEC, m as FindingsDiff, n as SkillUsageAnalyst, o as buildSkillUsageReport, p as DiffPolicy, q as RunCritic, r as SkillUsageRecord, s as emitSkillUsageFindings, t as SKILL_USAGE_ANALYST, u as KNOWLEDGE_GAP_KIND_SPEC, v as diffFindings, w as FindingSubjectStringSchema, x as FINDING_SUBJECT_SYNTAX, y as FINDING_SUBJECT_GRAMMAR_PROMPT } from "../skill-usage-C_pXm7lP.js";
+import { C as FindingSubjectKind, D as parseFindingSubject, E as findingSubjectGrammarPromptFor, F as createAnalystAi, H as SemanticConceptJudgeInput, O as renderFindingSubject, P as CreateAnalystAiConfig, S as FindingSubject, T as KIND_EXPECTED_SUBJECTS, U as SemanticConceptJudgeOptions, Y as RunTrace, _ as defaultIsMaterial, a as SkillUsageScanConfig, b as FINDING_SUBJECT_KINDS, c as DEFAULT_TRACE_ANALYST_KINDS, d as IMPROVEMENT_KIND_SPEC, f as FAILURE_MODE_KIND_SPEC, g as PersistedFinding, h as FindingsStore, i as SkillUsageReport, k as BehavioralMetrics, l as KNOWLEDGE_POISONING_KIND_SPEC, m as FindingsDiff, n as SkillUsageAnalyst, o as buildSkillUsageReport, p as DiffPolicy, q as RunCritic, r as SkillUsageRecord, s as emitSkillUsageFindings, t as SKILL_USAGE_ANALYST, u as KNOWLEDGE_GAP_KIND_SPEC, v as diffFindings, w as FindingSubjectStringSchema, x as FINDING_SUBJECT_SYNTAX, y as FINDING_SUBJECT_GRAMMAR_PROMPT } from "../skill-usage-D5mlWdAJ.js";
 import { c as CostLedgerHandle } from "../cost-ledger-Dye6jCgg.js";
 import { o as LlmClientOptions } from "../llm-client-B_nIBlYo.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-DGsxbAEd.js";
 import { t as TraceAnalysisStore } from "../store-CxJry_cs.js";
-import { A as AnalystRunResult, C as AnalystContext, D as AnalystRequirements, E as AnalystInputKind, F as computeFindingId, I as makeFinding, M as AnalystSeverity, N as AnalystUsageReceipt, O as AnalystRunEvent, P as EvidenceRef, S as Analyst, T as AnalystFinding, _ as RawAnalystEvidenceSchema, a as AnalystRegistryOptions, b as evidenceRefsFromRawFinding, c as CreateTraceAnalystKindOpts, d as createTraceAnalystKind, f as renderPriorFindings, g as RawAnalystEvidence, h as RAW_FINDING_SCHEMA_PROMPT, i as AnalystRegistry, j as AnalystRunSummary, k as AnalystRunInputs, l as TraceAnalystGolden, m as ANALYST_SEVERITIES, n as buildDefaultAnalystRegistry, o as BudgetPolicy, p as renderUpstreamFindings, r as AnalystHooks, s as RegistryRunOpts, t as DefaultAnalystRegistryOptions, u as TraceAnalystKindSpec, v as RawAnalystFinding, w as AnalystCost, x as parseRawFinding, y as RawAnalystFindingSchema } from "../default-registry-CNPo-Vsb.js";
+import { A as AnalystRunResult, C as AnalystContext, D as AnalystRequirements, E as AnalystInputKind, F as computeFindingId, I as makeFinding, M as AnalystSeverity, N as AnalystUsageReceipt, O as AnalystRunEvent, P as EvidenceRef, S as Analyst, T as AnalystFinding, _ as RawAnalystEvidenceSchema, a as AnalystRegistryOptions, b as evidenceRefsFromRawFinding, c as CreateTraceAnalystKindOpts, d as createTraceAnalystKind, f as renderPriorFindings, g as RawAnalystEvidence, h as RAW_FINDING_SCHEMA_PROMPT, i as AnalystRegistry, j as AnalystRunSummary, k as AnalystRunInputs, l as TraceAnalystGolden, m as ANALYST_SEVERITIES, n as buildDefaultAnalystRegistry, o as BudgetPolicy, p as renderUpstreamFindings, r as AnalystHooks, s as RegistryRunOpts, t as DefaultAnalystRegistryOptions, u as TraceAnalystKindSpec, v as RawAnalystFinding, w as AnalystCost, x as parseRawFinding, y as RawAnalystFindingSchema } from "../default-registry-Cj1oUpLN.js";
 import { AxFunction } from "@ax-llm/ax";
 //#region src/analyst/adapters.d.ts
 declare function liftSeverity(s: Severity): AnalystSeverity;

package/dist/analyst/index.js

@@ -1,6 +1,6 @@
 import { A as parseFindingSubject, C as stripCodeFences, D as FindingSubjectStringSchema, E as FINDING_SUBJECT_SYNTAX, F as makeFinding, L as createChatClient, M as behavioralAnalyst, N as deriveEfficiencyFindings, O as KIND_EXPECTED_SUBJECTS, P as computeFindingId, R as createAnalystAi, S as coerceToFindingRows, T as FINDING_SUBJECT_KINDS, _ as RawAnalystEvidenceSchema, a as KNOWLEDGE_GAP_KIND_SPEC, b as parseRawFinding, c as buildTraceToolsForGroup, d as renderUpstreamFindings, f as settleUsageReceiptFromCostLedger, g as RAW_FINDING_SCHEMA_PROMPT, h as ANALYST_SEVERITIES, i as KNOWLEDGE_POISONING_KIND_SPEC, j as renderFindingSubject, k as findingSubjectGrammarPromptFor, l as createTraceAnalystKind, m as structureFindings, n as AnalystRegistry, o as IMPROVEMENT_KIND_SPEC, p as validateUsageSettlementTimeout, r as DEFAULT_TRACE_ANALYST_KINDS, s as FAILURE_MODE_KIND_SPEC, t as buildDefaultAnalystRegistry, u as renderPriorFindings, v as RawAnalystFindingSchema, w as FINDING_SUBJECT_GRAMMAR_PROMPT, x as coerceJson, y as evidenceRefsFromRawFinding } from "../default-registry-C-vFCSEc.js";
 import { i as CostLedger } from "../cost-ledger-DIgQUFZZ.js";
-import { a as RunCritic, c as buildSkillUsageReport, d as defaultIsMaterial, f as diffFindings, i as runSemanticConceptJudge, l as emitSkillUsageFindings, n as SEMANTIC_CONCEPT_JUDGE_VERSION, o as SKILL_USAGE_ANALYST, s as SkillUsageAnalyst, u as FindingsStore } from "../semantic-concept-judge-B6cWNJ2K.js";
+import { a as RunCritic, c as buildSkillUsageReport, d as defaultIsMaterial, f as diffFindings, i as runSemanticConceptJudge, l as emitSkillUsageFindings, n as SEMANTIC_CONCEPT_JUDGE_VERSION, o as SKILL_USAGE_ANALYST, s as SkillUsageAnalyst, u as FindingsStore } from "../semantic-concept-judge-b5m3irbR.js";
 //#region src/analyst/adapters.ts
 /**
 * Adapter factories — lift each existing agent-eval primitive into the

package/dist/{analyze-runs-C1CavBMk.js → analyze-runs-BjPn_fOS.js}

@@ -1,7 +1,7 @@
-import { s as validateRunRecord } from "./run-record-BuoE80Dq.js";
 import { A as spearmanR, C as pairedTTest, E as pearsonR, L as continuousAgreement, O as requiredPairedSampleSize, b as pairedMde, v as pairedBootstrap, y as pairedCohensDz } from "./statistics-CnnxdpOg.js";
 import { r as pairRunRecords } from "./paired-arms-D9D0wXj2.js";
 import { r as observedSplitScore } from "./reward-nw2xZGZG.js";
+import { s as validateRunRecord } from "./run-record-CN8Zd21B.js";
 import { o as llmSpans } from "./query-Di7eEQ79.js";
 import { r as paretoChart } from "./summary-report-BNs5nmXI.js";
 //#region src/contamination-guard.ts
@@ -1064,4 +1064,4 @@ function buildRecommendations(ctx) {
 //#endregion
 export { checkBehavioralCanary as a, canaryLeakView as i, summarizeExecution as n, checkCanaries as o, HoldoutAuditor as r, runBehavioralCanaries as s, analyzeRuns as t };
 
-//# sourceMappingURL=analyze-runs-C1CavBMk.js.map
+//# sourceMappingURL=analyze-runs-BjPn_fOS.js.map