@openwop/openwop-conformance 2.1.3 → 2.1.4

package/CHANGELOG.md

@@ -1,5 +1,14 @@
 # `@openwop/openwop-conformance` Changelog
 
+## [2.1.4] — 2026-09-12 — the replay reds now say why
+
+`replay-observable-sequence-determinism` and `replay-divergence-at-refusal`
+asserted `status === 'completed'` bare, discarding the run's `error` and
+`currentNodeId`, so a red reported `expected 'failed' to be 'completed'` and
+nothing else. Both now attach the terminal detail, and both honour
+`OPENWOP_POLL_TIMEOUT_SCALE` — they were the only two scenarios that rolled
+their own poll loop and never read it. No assertion changed meaning.
+
 ## [2.1.3] — 2026-09-11 — `since` is the host's minor, said in the schema
 
 No scenario changed. The contract peer's generated

package/dist/spec-artifacts.lock.json

@@ -1,5 +1,5 @@
 {
   "package": "@openwop/spec-artifacts",
-  "version": "2.1.3",
-  "stampSha256": "05c0632125a5265905554cd3203217510192e15dafd6c8cd5654e10120d2265a"
+  "version": "2.1.4",
+  "stampSha256": "6f496326b394df65a2a9e89547f36693549d59cd13a9a5491dfb3ad34bb78596"
 }

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@openwop/openwop-conformance",
-  "version": "2.1.3",
+  "version": "2.1.4",
   "description": "Production-ready black-box conformance suite for OpenWOP v1.0 compliant servers.",
   "repository": {
     "type": "git",
@@ -56,6 +56,6 @@
     "@openwop/spec-artifacts": "file:../spec-artifacts"
   },
   "peerDependencies": {
-    "@openwop/spec-artifacts": "2.1.3"
+    "@openwop/spec-artifacts": "2.1.4"
   }
 }

package/requirements.json

@@ -21552,7 +21552,7 @@
     {
       "id": "openwop.it.replay-divergence-at-refusal.replaydeterminism-when-present-conforms-to-rfc-0041-d",
       "file": "replay-divergence-at-refusal.test.ts",
-      "line": 79,
+      "line": 111,
       "title": "replayDeterminism (when present) conforms to RFC 0041 §D",
       "explicitId": "openwop.it.replay-divergence-at-refusal.replaydeterminism-when-present-conforms-to-rfc-0041-d",
       "citations": [
@@ -21573,10 +21573,15 @@
     {
       "id": "openwop.it.replay-divergence-at-refusal.phase-4-host-must-emit-replay-divergedatrefusal-fail-with-replay-diverged-at-ref",
       "file": "replay-divergence-at-refusal.test.ts",
-      "line": 186,
+      "line": 227,
       "title": "Phase 4 host MUST emit replay.divergedAtRefusal + fail with replay_diverged_at_refusal when original=valid + replay=refusal",
       "explicitId": "openwop.it.replay-divergence-at-refusal.phase-4-host-must-emit-replay-divergedatrefusal-fail-with-replay-diverged-at-ref",
       "citations": [
+        {
+          "section": "RFCS/0041-multi-agent-replay-under-nondeterminism.md §D",
+          "requirement": null,
+          "interpolated": true
+        },
         {
           "section": "RFCS/0041-multi-agent-replay-under-nondeterminism.md §B + spec/v1/rest-endpoints.md §\"Common error codes\"",
           "requirement": "replay MUST terminate `failed` when refusal-divergence is detected (silent substitution is non-conformant)"
@@ -21602,7 +21607,7 @@
     {
       "id": "openwop.it.replay-divergence-at-refusal.phase-4-host-must-emit-replay-divergedatrefusal-fail-with-replay-diverged-at-ref~2",
       "file": "replay-divergence-at-refusal.test.ts",
-      "line": 272,
+      "line": 320,
       "title": "Phase 4 host MUST emit replay.divergedAtRefusal + fail with replay_diverged_at_refusal when original=refusal + replay=valid (symmetric case)",
       "explicitId": "openwop.it.replay-divergence-at-refusal.phase-4-host-must-emit-replay-divergedatrefusal-fail-with-replay-diverged-at-ref~2",
       "citations": [
@@ -22001,10 +22006,15 @@
     {
       "id": "openwop.it.replay-observable-sequence-determinism.original-and-replay-event-log-prefixes-must-be-byte-equivalent-modulo-per-event",
       "file": "replay-observable-sequence-determinism.test.ts",
-      "line": 168,
+      "line": 210,
       "title": "original and replay event-log prefixes MUST be byte-equivalent (modulo per-event clock + ULID entropy)",
       "explicitId": "openwop.it.replay-observable-sequence-determinism.original-and-replay-event-log-prefixes-must-be-byte-equivalent-modulo-per-event",
       "citations": [
+        {
+          "section": "RFCS/0041-multi-agent-replay-under-nondeterminism.md §C",
+          "requirement": null,
+          "interpolated": true
+        },
         {
           "section": "RFCS/0041-multi-agent-replay-under-nondeterminism.md §C",
           "requirement": "a mode:replay fork MUST reproduce the original observable event-log sequence byte-for-byte modulo volatile per-event fields (eventId/ULID entropy, per-region observedAt clock)"
@@ -22014,7 +22024,7 @@
     {
       "id": "openwop.it.replay-observable-sequence-determinism.replay-of-a-nondeterministic-tool-node-reproduces-the-original-observable-result",
       "file": "replay-observable-sequence-determinism.test.ts",
-      "line": 202,
+      "line": 251,
       "title": "replay of a nondeterministic tool node reproduces the ORIGINAL observable result, NOT a fresh call",
       "explicitId": "openwop.it.replay-observable-sequence-determinism.replay-of-a-nondeterministic-tool-node-reproduces-the-original-observable-result",
       "citations": [
@@ -22023,6 +22033,11 @@
           "requirement": null,
           "interpolated": true
         },
+        {
+          "section": "RFCS/0041-multi-agent-replay-under-nondeterminism.md §C",
+          "requirement": null,
+          "interpolated": true
+        },
         {
           "section": "RFCS/0041-multi-agent-replay-under-nondeterminism.md §C",
           "requirement": "the nondeterministic tool node MUST replay its ORIGINAL observable result (cached event-log entry) rather than re-executing — bit-equivalent re-execution would require unbounded caching, rejected per RFC 0041 §\"Alternatives considered\" #2"

package/schemas/CORPUS-STAMP.json

@@ -1,17 +1,17 @@
 {
   "_comment": "Provenance of @openwop/spec-artifacts (RFC 0168 §D.2). files: SHA-256 per file; the conformance suite compares the installed peer against dist/spec-artifacts.lock.json at start.",
   "package": "@openwop/spec-artifacts",
-  "version": "2.1.3",
-  "corpusTag": "v2.1.3",
+  "version": "2.1.4",
+  "corpusTag": "v2.1.4",
   "files": {
     "api/.redocly.lint-ignore.yaml": "bf5a8350b88a72fa43f59605ed8d903ed24b6cfccda5e45509c9f6ed9ee4e712",
     "api/asyncapi.yaml": "d5ecb9ee6114582be3b1f662c84bfac9ae96dae7bacb853e461168f70a8e1c7d",
     "api/grpc/openwop.proto": "c3e72bb17cba514ee98feb6434e6c9b6ea6795bfd086489ec69fd882dd1ad977",
     "api/openapi.yaml": "39081c59fb696159806b0f2f9a42e7e9ff830d622fcf2ad4159b21357580a955",
     "api/redocly.yaml": "b0604c89b2ca6d5076ec25725c539dad44a741a811fe524439ee6daef8baa09f",
-    "api/seams-v2.yaml": "af7bb78f590996cf16ecb5669b34fbafa700177ed06a2ece342574edd25c95b1",
-    "api/v2/asyncapi.yaml": "a3d6954f17ddc7d4e6ede42a9766a6bdddb20d4b642967358926e488af8ce082",
-    "api/v2/openapi.yaml": "d8b0526a61dc11ef8d94018e48bd679b7f1c28a2b32a1e914a4824435072150d",
+    "api/seams-v2.yaml": "277fb0c788ff5a085e3bb145e55db05ce002202f6c89a88429cf06187d07f610",
+    "api/v2/asyncapi.yaml": "7b8dcd7ca9b6be23b6e81ca19ceb6620c97b7df2840d68c547773895d17220aa",
+    "api/v2/openapi.yaml": "6afad97fbcfdfeef0776651f6264a27f5598ac8301a20842a9c739c8ccef3ad4",
     "api/v2/redocly.yaml": "1e66b60e6118ad11a823bb620678be464d99dfe50a40e3e6f93ec9429b88b34c",
     "schemas/README.md": "0c0b737ffcf8f30e7d2809cec8a498232de710f41443212922ad8337cdde0b51",
     "schemas/a2a-task-state.schema.json": "c9365918f993f943b4b619d42551eb066a1ed33a08d895d51b395432a5b1f1bc",
@@ -200,7 +200,7 @@
     "schemas/workspace-file.schema.json": "464de85c2a068243084ee9c1d969bc7cd5d8f7948574e58450d6493c38a0e1e4",
     "spec/v1/alias-detectors.json": "fee4594ef49953953ffcd0b3813300067d16b3e65ebff2aac722034ac9b3f545",
     "spec/v1/capability-declaration-classes.json": "e7729aed5c4b4e1dd02abab0530f14cc95f5d4070fe51fb139e7f5cccefa00c6",
-    "spec/v1/core-standard-manifest.json": "52a1aa95c365260c58cbbe2cf5e4560808814e066ba15811f9be534ef5c9bbd7",
+    "spec/v1/core-standard-manifest.json": "9b9a20963e9e364b0bb784a2fb38c376208393214e7ece962c9961fce03b1623",
     "spec/v1/deprecations.json": "1d5acb69a9b8ccb57275a95605f74aef1d920685f8407c9d382a46b59dc803bb",
     "spec/v1/deprecations.schema.json": "18c87e78bedc210431f795ae44c5b5d202f2f3317850d5cf86867d4f1fa1cdfb",
     "spec/v1/event-codemap.json": "3da60d884157793a360da532a9fcbbfb5285636db325a74cec94b34622186d97",
@@ -212,14 +212,14 @@
     "spec/v1/migrations.schema.json": "886779aa6c22e646db097f5df210adb018a4dd14a7b815465a18c8a7056c8f72",
     "spec/v1/operation-path-manifest.json": "5f5f4e3842669371730ebbd1aace3fb794192615f0018dc144f484ba5db82ac3",
     "spec/v1/spec-gaps.json": "6cc9962c6b969f632e07a78b52a4f61447ff579e2990cbae989866f584a86042",
-    "spec/v2/README.md": "69cffb23e28308482a552b71efaa35d8d8a40550fa8fcc35a8c20c9476ab59a3",
+    "spec/v2/README.md": "56a157cb6e35e7313d05514b438b4e148d679f6e084b7022431d8f2e81a519b0",
     "spec/v2/core/capabilities.md": "c5107a4964e487c20bbb5254cfe1f706e0d9ba7e34da9070bc43b50b385e7948",
     "spec/v2/core/conformance.md": "622084e3c74c4549087f4d861b1bb98cec201a0c5aec23788d27f53a355fde3b",
     "spec/v2/core/connection-packs.md": "1d759681e2962f5103228c52c45f24607ed4bb9ae73e30605f690c1d5a59e9ee",
     "spec/v2/core/errors.md": "74fdb5862f432a831cf0c2eadbe0c8cca6b2660ec578180ae33db54773a1df28",
     "spec/v2/core/events.md": "defe0969d6ec7c980f415cc0d78b9d065782ac7e5638b12a44aa4895a35d6126",
     "spec/v2/core/form-content-packs.md": "6ffaae85d4550be48dd53fb6e2959af8aceb35bd87c5343b722885084f197713",
-    "spec/v2/core/headers.md": "43c104a39f8c809550553f12d4f80d044006b37a95dc720a39914989de1c3ef5",
+    "spec/v2/core/headers.md": "35f9b7e912080bb4ff28d06dc1679a9b2857fcff6ce810d6df266bc66d2b13f2",
     "spec/v2/core/idempotency.md": "b86c352d964aebad532727f314e26ed2177d49de382076fa64d3162e2f56118d",
     "spec/v2/core/identity.md": "dd5299f9cdf54640870e1fffae11edb820cf6396ea298c4ebd14c649c4bd4a87",
     "spec/v2/core/interop.md": "d4c404e469e1d9d9be88bc2c939d777e7d7c1d0419b1251c18770f039838a401",
@@ -270,8 +270,8 @@
     "spec/v2/path-manifest.json": "a6f4e652e866889e8bde815874062f62929aee95dc8e58e7152c56adee262163",
     "spec/v2/peer-dependency-aliases.json": "d10299280abee08258502925bc327293ee413e0108cd6e6ec75ff6110653308d",
     "spec/v2/profiles.json": "0636f19fceae625390003a347e70ef4797d84766b5c24ce8a02cea52aadebca4",
-    "spec/v2/release.json": "2293c3f7bd204432f892936df8625a17f557283c252383b2553e92a18729fb19",
+    "spec/v2/release.json": "31797591070c2c06955915a0348dc9964d162e5fd1b544e073d667cbf78d9756",
     "spec/v2/retention-floors.json": "eaf3722d95c79947af1d4269ef85117e126518c588cfcf1a2b21b97269f51624"
   },
-  "corpusCommit": "9bc425c013fcba8d17329df901d384943283b168"
+  "corpusCommit": "5e9633c7297c993efc5a8198cd07dffcd2213e11"
 }

package/src/scenarios/replay-divergence-at-refusal.test.ts

@@ -75,6 +75,38 @@ async function readDiscovery(): Promise<DiscoveryDoc | null> {
   } catch { return null; }
 }
 
+/**
+ * Why this says more than `expect(status).toBe('completed')`.
+ *
+ * A terminal run that is `failed` carries `error` (`schemas/run-snapshot.schema.json`),
+ * and the bare equality assertion threw it away: every occurrence of this red
+ * reported only `expected 'failed' to be 'completed'`, with no node, no code and
+ * no message. A tier-1 host carried exactly that line in its notes across three
+ * lanes without ever being able to diagnose it (reported 2026-09-12), because
+ * the suite had recorded that something went wrong and nothing about what.
+ *
+ * That is the suite owing hosts what RFC 0064 §F makes a host owe the wire: a
+ * failure MUST be self-describing. This does not make the flake go away — the
+ * cause is still unknown, and pretending otherwise would be the third instrument
+ * this week that answers a narrower question than it appears to. It makes the
+ * next occurrence carry its own evidence.
+ */
+function terminalDetail(snap: RunSnapshot): string {
+  if (snap.status === 'completed') return '';
+  const err = snap.error as Record<string, unknown> | undefined;
+  const parts: string[] = [`status=${String(snap.status)}`];
+  if (snap.currentNodeId) parts.push(`currentNodeId=${String(snap.currentNodeId)}`);
+  if (err && typeof err === 'object') {
+    for (const k of ['code', 'message', 'nodeId', 'retriable']) {
+      if (err[k] !== undefined) parts.push(`error.${k}=${JSON.stringify(err[k])}`);
+    }
+    if (!('code' in err) && !('message' in err)) parts.push(`error=${JSON.stringify(err).slice(0, 300)}`);
+  } else {
+    parts.push('error=(absent — the host reported a non-completed terminal state with no error object)');
+  }
+  return ` — ${parts.join(' ')}`;
+}
+
 describe.skipIf(HTTP_SKIP)('replay-divergence-at-refusal: advertisement shape (RFC 0041 §D)', () => {
   it('replayDeterminism (when present) conforms to RFC 0041 §D', async (ctx) => {
     const d = await readDiscovery();
@@ -127,6 +159,7 @@ describe.skipIf(HTTP_SKIP)('replay-divergence-at-refusal: advertisement shape (R
 interface RunSnapshot {
   status?: string;
   error?: { code?: string; message?: string };
+  currentNodeId?: string;
 }
 interface RunEventDoc {
   type: string;
@@ -135,8 +168,16 @@ interface RunEventDoc {
   payload?: Record<string, unknown>;
 }
 
+// The 5s budget below is scaled by OPENWOP_POLL_TIMEOUT_SCALE like every
+// other bound in the suite. `lib/polling.ts` exists so a timeout can say
+// whether it measured the host or the harness — *"observe no change in
+// those scenarios, and record a failure that measured the environment
+// rather than the host"* — and these two replay files were the only ones
+// that rolled their own loop and never read it.
 async function pollUntilTerminal(runId: string): Promise<RunSnapshot> {
-  for (let i = 0; i < 50; i++) {
+  const scale = Number(process.env['OPENWOP_POLL_TIMEOUT_SCALE'] ?? '1');
+  const rounds = Math.max(1, Math.round(50 * (Number.isFinite(scale) && scale > 0 ? scale : 1)));
+  for (let i = 0; i < rounds; i++) {
     const r = await driver.get(`/v1/runs/${encodeURIComponent(runId)}`);
     const snap = r.json as RunSnapshot;
     if (snap.status === 'completed' || snap.status === 'failed' || snap.status === 'cancelled') {
@@ -144,7 +185,7 @@ async function pollUntilTerminal(runId: string): Promise<RunSnapshot> {
     }
     await new Promise((resolve) => setTimeout(resolve, 100));
   }
-  throw new Error(`run ${runId} did not reach terminal within 5s`);
+  throw new Error(`run ${runId} did not reach a terminal state within ${rounds * 100}ms (OPENWOP_POLL_TIMEOUT_SCALE=${process.env['OPENWOP_POLL_TIMEOUT_SCALE'] ?? '1'})`);
 }
 
 async function readEvents(runId: string): Promise<RunEventDoc[]> {
@@ -212,7 +253,14 @@ describe.skipIf(HTTP_SKIP)('replay-divergence-at-refusal: behavioral (RFC 0041
     expect(createRes.status).toBe(201);
     const sourceRunId = (createRes.json as { runId: string }).runId;
     const sourceTerminal = await pollUntilTerminal(sourceRunId);
-    expect(sourceTerminal.status).toBe('completed');
+    expect(
+        sourceTerminal.status,
+        req(
+          'openwop.it.replay-divergence-at-refusal.phase-4-host-must-emit-replay-divergedatrefusal-fail-with-replay-diverged-at-ref',
+          'RFCS/0041-multi-agent-replay-under-nondeterminism.md §D',
+          `the source run MUST reach \`completed\` for the replay comparison to mean anything${terminalDetail(sourceTerminal)}`,
+        ),
+      ).toBe('completed');
 
     // Stage refusal for the replay's mock-AI dispatch.
     await programMock(NODE_ID, [

package/src/scenarios/replay-observable-sequence-determinism.test.ts

@@ -62,6 +62,8 @@ interface DiscoveryDoc {
 
 interface RunSnapshot {
   status?: string;
+  error?: unknown;
+  currentNodeId?: string;
 }
 interface RunEventDoc {
   type: string;
@@ -93,8 +95,16 @@ async function gateOnPhase4(ctx: { skip: () => void }): Promise<boolean> {
   return true;
 }
 
+// The 5s budget below is scaled by OPENWOP_POLL_TIMEOUT_SCALE like every
+// other bound in the suite. `lib/polling.ts` exists so a timeout can say
+// whether it measured the host or the harness — *"observe no change in
+// those scenarios, and record a failure that measured the environment
+// rather than the host"* — and these two replay files were the only ones
+// that rolled their own loop and never read it.
 async function pollUntilTerminal(runId: string): Promise<RunSnapshot> {
-  for (let i = 0; i < 50; i++) {
+  const scale = Number(process.env['OPENWOP_POLL_TIMEOUT_SCALE'] ?? '1');
+  const rounds = Math.max(1, Math.round(50 * (Number.isFinite(scale) && scale > 0 ? scale : 1)));
+  for (let i = 0; i < rounds; i++) {
     const r = await driver.get(`/v1/runs/${encodeURIComponent(runId)}`);
     const snap = r.json as RunSnapshot;
     if (snap.status === 'completed' || snap.status === 'failed' || snap.status === 'cancelled') {
@@ -102,7 +112,7 @@ async function pollUntilTerminal(runId: string): Promise<RunSnapshot> {
     }
     await new Promise((resolve) => setTimeout(resolve, 100));
   }
-  throw new Error(`run ${runId} did not reach terminal within 5s`);
+  throw new Error(`run ${runId} did not reach a terminal state within ${rounds * 100}ms (OPENWOP_POLL_TIMEOUT_SCALE=${process.env['OPENWOP_POLL_TIMEOUT_SCALE'] ?? '1'})`);
 }
 
 async function readEvents(runId: string): Promise<RunEventDoc[]> {
@@ -162,6 +172,38 @@ async function startFixtureRun(ctx: { skip: () => void }): Promise<string | null
   return (create.json as { runId: string }).runId;
 }
 
+/**
+ * Why this says more than `expect(status).toBe('completed')`.
+ *
+ * A terminal run that is `failed` carries `error` (`schemas/run-snapshot.schema.json`),
+ * and the bare equality assertion threw it away: every occurrence of this red
+ * reported only `expected 'failed' to be 'completed'`, with no node, no code and
+ * no message. A tier-1 host carried exactly that line in its notes across three
+ * lanes without ever being able to diagnose it (reported 2026-09-12), because
+ * the suite had recorded that something went wrong and nothing about what.
+ *
+ * That is the suite owing hosts what RFC 0064 §F makes a host owe the wire: a
+ * failure MUST be self-describing. This does not make the flake go away — the
+ * cause is still unknown, and pretending otherwise would be the third instrument
+ * this week that answers a narrower question than it appears to. It makes the
+ * next occurrence carry its own evidence.
+ */
+function terminalDetail(snap: RunSnapshot): string {
+  if (snap.status === 'completed') return '';
+  const err = snap.error as Record<string, unknown> | undefined;
+  const parts: string[] = [`status=${String(snap.status)}`];
+  if (snap.currentNodeId) parts.push(`currentNodeId=${String(snap.currentNodeId)}`);
+  if (err && typeof err === 'object') {
+    for (const k of ['code', 'message', 'nodeId', 'retriable']) {
+      if (err[k] !== undefined) parts.push(`error.${k}=${JSON.stringify(err[k])}`);
+    }
+    if (!('code' in err) && !('message' in err)) parts.push(`error=${JSON.stringify(err).slice(0, 300)}`);
+  } else {
+    parts.push('error=(absent — the host reported a non-completed terminal state with no error object)');
+  }
+  return ` — ${parts.join(' ')}`;
+}
+
 describe.skipIf(HTTP_SKIP)(
   'replay-observable-sequence-determinism: prefix byte-equivalence (RFC 0041 §C)',
   () => {
@@ -171,7 +213,14 @@ describe.skipIf(HTTP_SKIP)(
       const sourceRunId = await startFixtureRun(ctx);
       if (sourceRunId === null) return softSkip('blocked', 'precondition not met — `sourceRunId === null` returned early (seam, prior step, or fixture unavailable)');
       const sourceTerminal = await pollUntilTerminal(sourceRunId);
-      expect(sourceTerminal.status).toBe('completed');
+      expect(
+        sourceTerminal.status,
+        req(
+          'openwop.it.replay-observable-sequence-determinism.original-and-replay-event-log-prefixes-must-be-byte-equivalent-modulo-per-event',
+          'RFCS/0041-multi-agent-replay-under-nondeterminism.md §C',
+          `the source run MUST reach \`completed\` for the replay comparison to mean anything${terminalDetail(sourceTerminal)}`,
+        ),
+      ).toBe('completed');
       const sourceEvents = await readEvents(sourceRunId);
 
       const forkRes = await driver.post(`/v1/runs/${encodeURIComponent(sourceRunId)}:fork`, {
@@ -204,7 +253,15 @@ describe.skipIf(HTTP_SKIP)(
 
       const sourceRunId = await startFixtureRun(ctx);
       if (sourceRunId === null) return softSkip('blocked', 'precondition not met — `sourceRunId === null` returned early (seam, prior step, or fixture unavailable)');
-      expect((await pollUntilTerminal(sourceRunId)).status).toBe('completed');
+      const t = await pollUntilTerminal(sourceRunId);
+      expect(
+        t.status,
+        req(
+          'openwop.it.replay-observable-sequence-determinism.replay-of-a-nondeterministic-tool-node-reproduces-the-original-observable-result',
+          'RFCS/0041-multi-agent-replay-under-nondeterminism.md §C',
+          `the source run MUST reach \`completed\` for the replay comparison to mean anything${terminalDetail(t)}`,
+        ),
+      ).toBe('completed');
       const sourceEvents = await readEvents(sourceRunId);
 
       // The terminal event(s) for the nondeterministic node carry its