@tangle-network/agent-eval 0.20.12 → 0.22.0

package/dist/chunk-6M774GY6.js

@@ -0,0 +1,53 @@
+// src/pre-registration.ts
+function canonicalize(v) {
+  if (v === null || typeof v !== "object") return v;
+  if (Array.isArray(v)) return v.map(canonicalize);
+  const keys = Object.keys(v).sort();
+  const out = {};
+  for (const k of keys) out[k] = canonicalize(v[k]);
+  return out;
+}
+async function hashJson(obj) {
+  const canonical = canonicalize(obj);
+  const bytes = new TextEncoder().encode(JSON.stringify(canonical));
+  const digest = await globalThis.crypto.subtle.digest("SHA-256", bytes);
+  return Array.from(new Uint8Array(digest)).map((b) => b.toString(16).padStart(2, "0")).join("");
+}
+async function signManifest(m) {
+  const hash = await hashJson(m);
+  return { ...m, contentHash: hash, algo: "sha256-content" };
+}
+async function verifyManifest(m) {
+  const { contentHash, algo: _algo, ...rest } = m;
+  void _algo;
+  const resigned = await signManifest(rest);
+  return resigned.contentHash === contentHash;
+}
+async function evaluateHypothesis(manifest, observed) {
+  if (!await verifyManifest(manifest)) {
+    throw new Error("evaluateHypothesis: manifest content hash mismatch (tampered)");
+  }
+  const reasons = [];
+  const directionOk = manifest.direction === "increase" ? observed.effect > 0 : observed.effect < 0;
+  if (!directionOk) reasons.push("wrong_direction");
+  if (Math.abs(observed.effect) < manifest.minEffect) reasons.push("effect_too_small");
+  if (observed.pValue >= manifest.alpha) reasons.push("not_significant");
+  if (observed.n < manifest.preRegisteredN) reasons.push("undersampled");
+  return {
+    manifest,
+    observedN: observed.n,
+    observedEffect: observed.effect,
+    observedPValue: observed.pValue,
+    confirmed: reasons.length === 0,
+    rejectionReasons: reasons
+  };
+}
+
+export {
+  canonicalize,
+  hashJson,
+  signManifest,
+  verifyManifest,
+  evaluateHypothesis
+};
+//# sourceMappingURL=chunk-6M774GY6.js.map

package/dist/chunk-6M774GY6.js.map

@@ -0,0 +1 @@
+{"version":3,"sources":["../src/pre-registration.ts"],"sourcesContent":["/**\n * Pre-registered hypotheses — declare what you're testing BEFORE the\n * run, check it AFTER. Prevents p-hacking, optional stopping, and the\n * \"we ran until it looked good\" failure mode.\n *\n * Manifest is a plain JSON-friendly object. Sign it with a content hash\n * + timestamp; the registered record becomes immutable. Post-run,\n * evaluate the manifest against observed results — the library refuses\n * to let you re-interpret a different metric as the declared one.\n */\n\nexport interface HypothesisManifest {\n  id: string\n  /** Human prose — goes into the audit trail. */\n  hypothesis: string\n  /** Metric the hypothesis claims to move. */\n  metric: string\n  /** 'increase' = candidate should score higher than baseline; 'decrease' = lower. */\n  direction: 'increase' | 'decrease'\n  /** Minimum effect size to count (same units as the metric). */\n  minEffect: number\n  /** Alpha threshold. */\n  alpha: number\n  /** Target statistical power at which sample size was pre-computed. */\n  power: number\n  /** Declared N per arm before running. */\n  preRegisteredN: number\n  /** ISO8601 timestamp the manifest was registered. */\n  registeredAt: string\n  /** Optional identifiers to tie into the trace corpus. */\n  baselineLabel?: string\n  candidateLabel?: string\n}\n\n/**\n * Identifier for the hashing scheme used to produce `contentHash`.\n *\n * `'sha256-content'` — sha256 hex over the canonicalized manifest with\n * the `contentHash` and `algo` fields stripped. This is what\n * `signManifest` produces today.\n *\n * Held as a string union so future schemes can be added without\n * breaking parsers; legacy SignedManifest values written before this\n * field existed will deserialize cleanly because the field is optional.\n */\nexport type SignedManifestAlgo = 'sha256-content'\n\nexport interface SignedManifest extends HypothesisManifest {\n  /** sha256 hex of canonicalized manifest (everything except contentHash and algo). */\n  contentHash: string\n  /**\n   * Algorithm string describing how `contentHash` was produced.\n   *\n   * Optional on the type so legacy serialized manifests (pre-`algo`)\n   * still parse, but ALWAYS populated by {@link signManifest}.\n   * Consumers that want to enforce a known algorithm should reject\n   * manifests where this field is missing or unrecognized.\n   */\n  algo?: SignedManifestAlgo\n}\n\nexport interface HypothesisResult {\n  manifest: SignedManifest\n  observedN: number\n  observedEffect: number\n  observedPValue: number\n  /** True iff the observed effect hits the pre-declared direction with\n   *  magnitude ≥ minEffect AND p < alpha. */\n  confirmed: boolean\n  /** Enumerated reasons the hypothesis was rejected (each a machine-tag). */\n  rejectionReasons: Array<'wrong_direction' | 'effect_too_small' | 'not_significant' | 'undersampled'>\n  notes?: string\n}\n\n/**\n * Deterministic JSON canonicalization — sort object keys recursively.\n *\n * Two semantically-equal objects produce byte-identical canonicalized output;\n * this is what makes a content-hash stable across encoders, key insertion\n * orders, and runtime versions. Exported for any consumer that needs the same\n * canonicalization guarantee outside the manifest-signing path (e.g., signing\n * an artifact bundle, hashing a dataset version, etc.).\n */\nexport function canonicalize(v: unknown): unknown {\n  if (v === null || typeof v !== 'object') return v\n  if (Array.isArray(v)) return v.map(canonicalize)\n  const keys = Object.keys(v as Record<string, unknown>).sort()\n  const out: Record<string, unknown> = {}\n  for (const k of keys) out[k] = canonicalize((v as Record<string, unknown>)[k])\n  return out\n}\n\n/**\n * SHA-256 hex (full 64 chars) over the canonicalized JSON encoding of `obj`.\n *\n * The same primitive `signManifest` and `verifyManifest` are built on, exposed\n * directly so consumers signing arbitrary structured content (artifact bundles,\n * production packets, dataset manifests, etc.) don't have to re-derive\n * canonicalize+sha256 from scratch.\n *\n * Stable across:\n *   - object key insertion order (canonicalization sorts keys recursively)\n *   - encoder choice (UTF-8 via TextEncoder, fixed)\n *   - runtime (uses the Web Crypto subtle digest, present in Node ≥18 and browsers)\n *\n * Naming note: `hashJson` rather than `hashContent` because `hashContent` is\n * already taken in `prompt-registry.ts` for the truncated 12-char prompt-id\n * helper, which has different semantics (string input, short return). Both\n * coexist; `hashJson` is the right name when you mean \"canonicalize then hash.\"\n *\n * @example\n *   const hash = await hashJson({ id: '1', kind: 'spec' })\n *   // 'a3f1...' (64 hex chars)\n */\nexport async function hashJson<T>(obj: T): Promise<string> {\n  const canonical = canonicalize(obj)\n  const bytes = new TextEncoder().encode(JSON.stringify(canonical))\n  const digest = await globalThis.crypto.subtle.digest('SHA-256', bytes)\n  return Array.from(new Uint8Array(digest))\n    .map((b) => b.toString(16).padStart(2, '0'))\n    .join('')\n}\n\n/**\n * Sign a manifest with a SHA-256 content hash.\n *\n * The hash covers the canonicalized manifest with the `contentHash`\n * and `algo` fields stripped; this lets verifiers re-sign the rest and\n * compare. Returned manifest always carries `algo: 'sha256-content'`\n * so downstream consumers can identify the scheme; legacy serialized\n * manifests without `algo` still verify because it is stripped before\n * hashing on both sides.\n */\nexport async function signManifest(m: HypothesisManifest): Promise<SignedManifest> {\n  const hash = await hashJson(m)\n  return { ...m, contentHash: hash, algo: 'sha256-content' }\n}\n\n/**\n * Verify that a signed manifest has not been tampered with.\n *\n * Strips `contentHash` and `algo` before re-signing so legacy manifests\n * (written before `algo` was emitted) verify identically to current\n * ones.\n */\nexport async function verifyManifest(m: SignedManifest): Promise<boolean> {\n  const { contentHash, algo: _algo, ...rest } = m\n  void _algo\n  const resigned = await signManifest(rest)\n  return resigned.contentHash === contentHash\n}\n\n/**\n * Evaluate a pre-registered hypothesis against observed results.\n * Mechanical — no re-interpretation permitted.\n */\nexport async function evaluateHypothesis(\n  manifest: SignedManifest,\n  observed: { n: number; effect: number; pValue: number },\n): Promise<HypothesisResult> {\n  if (!(await verifyManifest(manifest))) {\n    throw new Error('evaluateHypothesis: manifest content hash mismatch (tampered)')\n  }\n  const reasons: HypothesisResult['rejectionReasons'] = []\n  const directionOk =\n    manifest.direction === 'increase' ? observed.effect > 0 : observed.effect < 0\n  if (!directionOk) reasons.push('wrong_direction')\n  if (Math.abs(observed.effect) < manifest.minEffect) reasons.push('effect_too_small')\n  if (observed.pValue >= manifest.alpha) reasons.push('not_significant')\n  if (observed.n < manifest.preRegisteredN) reasons.push('undersampled')\n  return {\n    manifest,\n    observedN: observed.n,\n    observedEffect: observed.effect,\n    observedPValue: observed.pValue,\n    confirmed: reasons.length === 0,\n    rejectionReasons: reasons,\n  }\n}\n"],"mappings":";AAmFO,SAAS,aAAa,GAAqB;AAChD,MAAI,MAAM,QAAQ,OAAO,MAAM,SAAU,QAAO;AAChD,MAAI,MAAM,QAAQ,CAAC,EAAG,QAAO,EAAE,IAAI,YAAY;AAC/C,QAAM,OAAO,OAAO,KAAK,CAA4B,EAAE,KAAK;AAC5D,QAAM,MAA+B,CAAC;AACtC,aAAW,KAAK,KAAM,KAAI,CAAC,IAAI,aAAc,EAA8B,CAAC,CAAC;AAC7E,SAAO;AACT;AAwBA,eAAsB,SAAY,KAAyB;AACzD,QAAM,YAAY,aAAa,GAAG;AAClC,QAAM,QAAQ,IAAI,YAAY,EAAE,OAAO,KAAK,UAAU,SAAS,CAAC;AAChE,QAAM,SAAS,MAAM,WAAW,OAAO,OAAO,OAAO,WAAW,KAAK;AACrE,SAAO,MAAM,KAAK,IAAI,WAAW,MAAM,CAAC,EACrC,IAAI,CAAC,MAAM,EAAE,SAAS,EAAE,EAAE,SAAS,GAAG,GAAG,CAAC,EAC1C,KAAK,EAAE;AACZ;AAYA,eAAsB,aAAa,GAAgD;AACjF,QAAM,OAAO,MAAM,SAAS,CAAC;AAC7B,SAAO,EAAE,GAAG,GAAG,aAAa,MAAM,MAAM,iBAAiB;AAC3D;AASA,eAAsB,eAAe,GAAqC;AACxE,QAAM,EAAE,aAAa,MAAM,OAAO,GAAG,KAAK,IAAI;AAC9C,OAAK;AACL,QAAM,WAAW,MAAM,aAAa,IAAI;AACxC,SAAO,SAAS,gBAAgB;AAClC;AAMA,eAAsB,mBACpB,UACA,UAC2B;AAC3B,MAAI,CAAE,MAAM,eAAe,QAAQ,GAAI;AACrC,UAAM,IAAI,MAAM,+DAA+D;AAAA,EACjF;AACA,QAAM,UAAgD,CAAC;AACvD,QAAM,cACJ,SAAS,cAAc,aAAa,SAAS,SAAS,IAAI,SAAS,SAAS;AAC9E,MAAI,CAAC,YAAa,SAAQ,KAAK,iBAAiB;AAChD,MAAI,KAAK,IAAI,SAAS,MAAM,IAAI,SAAS,UAAW,SAAQ,KAAK,kBAAkB;AACnF,MAAI,SAAS,UAAU,SAAS,MAAO,SAAQ,KAAK,iBAAiB;AACrE,MAAI,SAAS,IAAI,SAAS,eAAgB,SAAQ,KAAK,cAAc;AACrE,SAAO;AAAA,IACL;AAAA,IACA,WAAW,SAAS;AAAA,IACpB,gBAAgB,SAAS;AAAA,IACzB,gBAAgB,SAAS;AAAA,IACzB,WAAW,QAAQ,WAAW;AAAA,IAC9B,kBAAkB;AAAA,EACpB;AACF;","names":[]}

package/dist/{chunk-MCMV7DUL.js → chunk-ARZ6BEV6.js}

@@ -3,7 +3,7 @@ import {
 } from "./chunk-YUFXO3TU.js";
 import {
   TraceEmitter
-} from "./chunk-PKCVBYTQ.js";
+} from "./chunk-5IIQKMD5.js";
 
 // src/control-runtime.ts
 var DEFAULT_BUDGET = {
@@ -1307,4 +1307,4 @@ export {
   runProposeReviewAsControlLoop,
   controlFailureClassFromVerification
 };
-//# sourceMappingURL=chunk-MCMV7DUL.js.map
+//# sourceMappingURL=chunk-ARZ6BEV6.js.map