@openwop/openwop-conformance 2.40.2 → 2.41.0

package/CHANGELOG.md

@@ -1,5 +1,22 @@
 # `@openwop/openwop-conformance` Changelog
 
+## [2.41.0] — 2026-09-26 — RFC 0111's context-budget witness can fail
+
+- **Conformance Soak is green again: the receiver modules count as harness doubles.** `host-callback-declaration` keys on imports of modules that stand up a server the host must reach, and its list omitted `scoped-receiver` and `effect-receiver`. So `v2-a2a-push-delivery`'s correct `REQUIRES_HOST_CALLBACK` read as a callback it "does not make", and the post-merge Soak failed on every run from #1546 (2026-09-24) on, which no PR check saw. Both modules join the list, and the eight scenarios that take a host callback through them now declare it (`v2-bound-id-kinds`, `v2-webhook-delivery-isolation`, `v2-webhook-delivery-shape`, `v2-webhook-durable-delivery`, `v2-webhook-unregister-stops-delivery`, `webhook-signed-delivery`, `v2-durability-recovery`, `v2-terminal-event-once`), so an off-process consumer can list them before running. Sabotage: dropping `scoped-receiver` again reproduces the Soak failure.
+- **`context-budget-transcript-bound` can now fail in the ways RFC 0111 names.** It checked that each fed event id existed and was unique, so a wrong `tokenCount`, a window that was not a recent tail, or a budget of 10⁹ the run never reached all passed. The checks now live in the pure checker `src/lib/context-budget.ts` (15 self-tests, one per sabotage). For `tokenCounter: "chars"`, the transcript-window seam must return `entries[] { eventId, rendered }` and the suite sums the rendered Unicode code points, which must equal `tokenCount`. It also checks every id is a real event and the verbatim entries equal `eventIds` in order. For the recent tail, fed events are in log order, none at or after the turn's own `runOrchestrator.decided`, and no newer eligible event (same `(type, nodeId)`) is dropped. A turn with no budget pressure records `partial-witness:`, not a pass. Other token units are advertise-and-attest. See the COMPATIBILITY.md correction on record of 2026-09-26.
+- **`context-summarization-replay` compares the summary text the model saw.** Equal `summaryRef`s showed the fork cited the same artifact. When the seam serves `entries[]` for both runs, the fork's fed summary text must now equal the source's byte for byte.
+- **Both scenarios run at major 2.** They are added to `BOTH_MAJORS`, so v2-applicable files go from 121 to 123. At major 2 they read the `multiAgent` record by presence and record `inapplicable` when `contextBudget` or `summarization` is absent. That is never a strict-mode failure, so no currently certified host is affected. Paths resolve through `runsPath()` and the driver's seam rewrite, and the seam is added to `api/seams-v2.yaml` at `/conformance/seams/sample/agent/transcript-window`.
+- **New fixture `conformance-context-budget-live`.** A supervisor driven by the host's live model, with no mock keys and bounded by `settings.maxLoopbackIterations: 8`. Both scenarios gate on it. `conformance-context-budget-multiturn` is kept but no longer gates anything: it scripts a mock supervisor, which RFC 0111 §Scope forbids from advertising `contextBudget`.
+- **RFC 0121 gap G1 closes** (UQ1 resolved for GitHub Copilot individual plans): `spec/v1/gaps.json` in `@openwop/spec-artifacts` moves. No scenario changes.
+- **Version moved ahead of publication.** `@openwop/openwop-conformance` and its exact-pinned peer `@openwop/spec-artifacts` move to `2.41.0`, absorbing the unpublished `2.40.4` cycle (its one entry, RFC 0121 G1, is above): the major-2 scenario promotion makes this a minor. Not tagged, not published.
+
+## [2.40.3] — 2026-09-26 — a cookie-borne lane can be revoke-witnessed, and a secret-named setting no longer corrupts the discovery digest
+
+- **`v2-revocation-honored` presents a cookie-lane credential as a cookie.** When the mint seam answers `presentation: { kind: "cookie", name }`, the probe sends `Cookie: <name>=<credential>` instead of a bearer token. The seams contract's `mintLaneCredential` response gains that optional member. Bearer stays the default when `presentation` is absent. The cookie branch is unexercised until a host serves it; openwop-app's `session` lane is the first expected.
+- **RFC 0215 gap G1 closes in the contract peer.** `@openwop/spec-artifacts`' `spec/v1/gaps.json` records MyndHyve's measurement (openwop#1595), and `spec/v2/core/webhooks.md`'s `Stable` banner cites RFC 0215 now that it is `Accepted`. No scenario change.
+- **A setting that names a secret is not scrubbed as one.** `evidenceSecretsFromEnv` sweeps every `OPENWOP_*` value whose NAME mentions a key/token/secret/password, and `scrubEvidence` then rewrites that value everywhere in the bundle. openwop-app's RFC 0201 setting `OPENWOP_WEBHOOK_SECRET_ROTATION_OVERLAP_S=60` was swept, the "60" inside `discovery.sha256` was rewritten, the bundle failed `^[0-9a-f]{64}$`, and its deploy certify refused to ship. Values guessed from names are now swept only when credential-shaped (at least 8 characters and not a bare integer); a credential the emitter hands over explicitly is scrubbed whatever its shape. Sabotage: without the floor the new redaction leg fails.
+- **Version moved ahead of publication.** `@openwop/openwop-conformance` and its exact-pinned peer `@openwop/spec-artifacts` move to `2.40.3` because `2.40.2` is tagged. Not tagged, not published.
+
 ## [2.40.2] — 2026-09-26 — production bundles stop carrying harness-issuer blocked rows, and a colocated companion is marked
 
 - **RFC 0216: `--as-colocated-companion`, and a companion witnesses only the harness-trust-anchor rows.** The flag marks a v3 bundle `host.deployment: "colocated-companion"`. It is refused without `--certify` and bundle version 3. `witnessDigest` takes the marker into the preimage when present (`{ rows, relaxations?, deployment }`), so a signed companion cannot shed it: `--verify` then refuses with `witness-digest`. New coherence scenario `v2-colocated-companion` (7 legs, one id each) drives `scripts/check-harness-trust-anchors.mjs` and `scripts/check-companion-pairing.mjs` on tmpdir fixtures signed by the suite's own `signBundleV3`. Five sabotages were each run red.

package/README.md

@@ -11,7 +11,7 @@
 # --legacy-peer-deps is REQUIRED, not optional: the exact peer pin is what npm's
 # default resolver refuses. npm 10.9 fails outright with
 # "Cannot read properties of null (reading 'edgesOut')" — use npm >= 11.
-npm install --legacy-peer-deps @openwop/openwop-conformance@2.40.2 @openwop/spec-artifacts@2.40.2
+npm install --legacy-peer-deps @openwop/openwop-conformance@2.41.0 @openwop/spec-artifacts@2.41.0
 # or run without install:
 npx @openwop/openwop-conformance --base-url https://api.example.com --api-key hk_test_...
 ```

package/dist/lib/certification-bundle-verify.js

@@ -267,6 +267,25 @@ export function scrubEvidence(value, secrets) {
  * are excluded, and a caller may name values that MUST stay visible (`except`)
  * regardless of which variable carried them.
  */
+/**
+ * Whether an environment VALUE can be a credential at all (suite 2.40.3).
+ *
+ * The name filter above matches settings ABOUT secrets as well as secrets:
+ * openwop-app's `OPENWOP_WEBHOOK_SECRET_ROTATION_OVERLAP_S=60` (RFC 0201's
+ * rotation overlap) was swept in, and `scrubEvidence` then rewrote every "60"
+ * in the bundle — including the one inside `discovery.sha256`, which stopped
+ * matching `^[0-9a-f]{64}$` and failed the host's own deploy certify. A short
+ * value occurs by chance inside digests, counts and timestamps, so scrubbing it
+ * corrupts the evidence without protecting anything: no credential is a bare
+ * integer or under 8 characters (NIST SP 800-63B's floor for a memorized
+ * secret). The floor applies only to values GUESSED from variable names; a
+ * credential the emitter hands over explicitly (`extra`) is always scrubbed,
+ * whatever its shape.
+ */
+export function credentialShaped(v) {
+    const t = v.trim();
+    return t.length >= 8 && !/^\d+$/.test(t);
+}
 export function evidenceSecretsFromEnv(env, extra = [], except = []) {
     const keep = new Set(except.filter((s) => s !== undefined && s.trim() !== ''));
     const out = new Set();
@@ -277,8 +296,11 @@ export function evidenceSecretsFromEnv(env, extra = [], except = []) {
             continue;
         if (/_ID$/.test(k))
             continue; // a key's identifier is published, not secret
-        if (v !== undefined && v.trim() !== '' && !keep.has(v))
-            out.add(v);
+        if (v === undefined || v.trim() === '' || keep.has(v))
+            continue;
+        if (!credentialShaped(v))
+            continue; // a setting that names a secret, not a secret
+        out.add(v);
     }
     for (const s of extra)
         if (s !== undefined && s.trim() !== '' && !keep.has(s))

package/dist/spec-artifacts.lock.json

@@ -1,5 +1,5 @@
 {
   "package": "@openwop/spec-artifacts",
-  "version": "2.40.2",
-  "stampSha256": "5b784776f9088d58f1fd29e14efa113715876aac5940b728f775a779c59a962a"
+  "version": "2.41.0",
+  "stampSha256": "6b53adc27c8d6212c78dd08bf0be7b3c2d740d4290a72ce55324fd1fb4a6fe46"
 }

package/fixtures/conformance-context-budget-live.json

@@ -0,0 +1,44 @@
+{
+  "id": "conformance-context-budget-live",
+  "name": "Conformance: Context-Budget Live Orchestrator (RFC 0111)",
+  "version": "1.0",
+  "description": "RFC 0111, the witness fixture (suite 2.41.0). A `core.orchestrator.supervisor` loop driven by the host's LIVE model — no `mockDecisions`, `mockDispatchPlan` or `mockPendingDecision` — because RFC 0111 §Scope forbids a mock supervisor from advertising `contextBudget`. The supervisor receives `config.instructions` as its objective and dispatches `conformance-noop` until it terminates, bounded by `settings.maxLoopbackIterations`. Each iteration's transcript grows; a host advertising `multiAgent.executionModel.contextBudget` reports what it fed the model through `GET /v1/host/sample/agent/transcript-window` (`host-sample-test-seams.md` §14), including `entries[] { eventId, rendered }`. For a full witness the run MUST put the budget under pressure (an older eligible event evicted or summarized); a host whose advertised budget this run never reaches records `partial-witness`. Hosts SHOULD run the live supervisor at temperature 0 with structured output so the decision shape is stable; the scenario asserts the host's accounting, never the model's words.",
+  "nodes": [
+    {
+      "id": "supervisor",
+      "typeId": "core.orchestrator.supervisor",
+      "name": "Supervisor",
+      "position": { "x": 0, "y": 0 },
+      "config": {
+        "instructions": "You are testing a transcript budget. Dispatch the worker conformance-noop once per turn. After it has completed six times, terminate with reason goal-reached. Do not ask the user anything."
+      },
+      "inputs": {},
+      "agent": {
+        "agentId": "core.conformance.supervisor",
+        "modelClass": "reasoning"
+      }
+    },
+    {
+      "id": "dispatch",
+      "typeId": "core.dispatch",
+      "name": "Dispatch",
+      "position": { "x": 200, "y": 0 },
+      "config": {
+        "askUserRouting": "auto",
+        "workerDispatchModel": "child-run",
+        "fanOutPolicy": "sequential"
+      },
+      "inputs": {}
+    }
+  ],
+  "edges": [
+    { "id": "e1", "sourceNodeId": "supervisor", "targetNodeId": "dispatch" },
+    { "id": "e2", "sourceNodeId": "dispatch", "targetNodeId": "supervisor" }
+  ],
+  "triggers": [
+    { "id": "manual", "type": "manual", "enabled": true }
+  ],
+  "variables": [],
+  "metadata": { "tags": ["conformance", "multi-agent", "orchestrator", "context-budget", "rfc-0111", "live-model"] },
+  "settings": { "timeout": 300000, "maxLoopbackIterations": 8 }
+}

package/fixtures/conformance-context-budget-multiturn.json

@@ -2,7 +2,7 @@
   "id": "conformance-context-budget-multiturn",
   "name": "Conformance: Context-Budget Multi-Turn Orchestrator (RFC 0111)",
   "version": "1.0",
-  "description": "RFC 0111. Drives a multi-iteration `core.orchestrator.supervisor` loop so the per-turn transcript grows across iterations — enough turns that an UNBOUNDED transcript would exceed a host's advertised `multiAgent.executionModel.contextBudget.transcriptTokenBudget`. A host that bounds the transcript exposes its per-iteration accounting via the OPTIONAL seam `GET /v1/host/sample/agent/transcript-window?runId=…&iteration=N` (`host-sample-test-seams.md` §14): `{ tokenCounter, tokenCount, eventIds, summarizedRanges }`. The conformance scenario `context-budget-transcript-bound.test.ts` asserts `tokenCount ≤ budget`, independently re-counts the events named in `eventIds` from the run event-log, confirms the recent-tail + keepLastTurns-verbatim rules, and (for `summarization.supported` hosts) `context-summarization-replay.test.ts` replays the run and asserts the recorded `context.summarized` → `summaryRef` is reused (never re-summarized). Reference hosts that do not advertise `contextBudget` soft-skip both scenarios.",
+  "description": "RFC 0111 — SUPERSEDED as the witness gate in suite 2.41.0. This fixture scripts the supervisor with `mockDecisions`, i.e. a mock supervisor, which RFC 0111 §Scope forbids from advertising `contextBudget`; no honest host can witness the capability against it. Kept so hosts that still list it keep validating. Both RFC 0111 scenarios (`context-budget-transcript-bound`, `context-summarization-replay`) gate on `conformance-context-budget-live`, which drives the host's live model.",
   "nodes": [
     {
       "id": "supervisor",

package/fixtures.md

@@ -99,7 +99,8 @@ All fixtures MUST advertise:
 | Agent Memory Redaction                    | `conformance-agent-memory-redaction`                                                            | Phase 3 / SR-1 — BYOK plaintext surfaces as `[REDACTED:<id>]` on read                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                     | `completed`                                                                         | ≤ 15s                        |
 | Agent Memory TTL                          | `conformance-agent-memory-ttl`                                                                  | Phase 3 — `expiresAt` excludes expired entries from list/get; TWO-SIDED (S35, 2026-08-17): host lands `freshId` / `expiredId`, `memoryList` MUST include the fresh id and exclude the expired id                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           | `completed`                                                                         | ≤ 10s                        |
 | Agent Memory Injection Budget             | `conformance-agent-memory-injection-budget`                                                     | RFC 0113 — token-budgeted `MemoryAdapter.list` (`tokenBudget`/`rank`/`query`): cumulative tokens ≤ budget, over-budget single entry omitted, SR-1 + CTI-1 re-asserted on the budgeted path; `rank:'relevance'` delegates to `memory.search` semantic                                                                                                                                                                                                                                                                                                                                                                                                       | `completed`                                                                         | ≤ 15s                        |
-| Context-Budget Multi-Turn                 | `conformance-context-budget-multiturn`                                                          | RFC 0111 — multi-iteration `core.orchestrator.supervisor` loop whose per-turn transcript grows; a host bounding it exposes `{ tokenCounter, tokenCount, eventIds, summarizedRanges }` via the `GET /v1/host/sample/agent/transcript-window` seam (`tokenCount ≤ transcriptTokenBudget`, internally consistent, keepLastTurns verbatim); the summarization leg replays and asserts `context.summarized` → `summaryRef` reuse. Reference hosts that do not advertise `contextBudget` soft-skip                                                                                                                                              | `completed`                                                                         | ≤ 60s                        |
+| Context-Budget Multi-Turn                 | `conformance-context-budget-multiturn`                                                          | RFC 0111 — **superseded as the witness gate in suite 2.41.0.** Drives the supervisor through `mockDecisions`, i.e. a mock supervisor — exactly what RFC 0111 §Scope forbids from advertising `contextBudget`, so no honest host can witness against it. Kept (v1 compatibility: hosts may still list it); both RFC 0111 scenarios now gate on `conformance-context-budget-live`                                                                                                                                              | `completed`                                                                         | ≤ 60s                        |
+| Context-Budget Live Orchestrator | `conformance-context-budget-live` | RFC 0111 witness (suite 2.41.0) — `core.orchestrator.supervisor` driven by the host's LIVE model (no mock keys; `config.instructions` is the objective), dispatching `conformance-noop` until it terminates, bounded by `settings.maxLoopbackIterations: 8`. The host reports each iteration's fed transcript through `GET /v1/host/sample/agent/transcript-window` including `entries[] { eventId, rendered }`; `context-budget-transcript-bound` recounts `chars`, checks real events + the recent tail + budget PRESSURE (no pressure ⇒ `partial-witness`); `context-summarization-replay` replays and compares the fed summary text byte for byte. A host whose supervisor has no live model MUST NOT advertise this fixture | `completed` | ≤ 300s |
 | Conversation Lifecycle                    | `conformance-conversation-lifecycle`                                                            | Phase 4 — open → exchange → close event ordering                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                          | `completed`                                                                         | ≤ 20s                        |
 | Conversation vs Clarification             | `conformance-conversation-vs-clarification`                                                     | Phase 4 — conversation suspend emits `conversation.*`, NOT `clarification.*`                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                              | `completed`                                                                         | ≤ 15s                        |
 | Conversation Replay                       | `conformance-conversation-replay`                                                               | Phase 4 — `:fork` preserves conversation channel projection                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               | `completed`                                                                         | ≤ 30s                        |

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@openwop/openwop-conformance",
-  "version": "2.40.2",
+  "version": "2.41.0",
   "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": "2.40.2"
+    "@openwop/spec-artifacts": "2.41.0"
   }
 }

package/requirements.json

@@ -3,10 +3,10 @@
   "generatedFrom": "src/scenarios/*.test.ts",
   "counts": {
     "files": 636,
-    "tests": 2437,
-    "withStableId": 2437,
+    "tests": 2438,
+    "withStableId": 2438,
     "interpolatedTitles": 0,
-    "explicitIds": 2331
+    "explicitIds": 2332
   },
   "records": [
     {
@@ -7170,9 +7170,38 @@
       ]
     },
     {
-      "id": "openwop.it.certification-bundle-redaction.evidencesecretsfromenv-never-scrubs-a-key-identifier-id-variables-are-excluded-a",
+      "id": "openwop.it.certification-bundle-redaction.evidencesecretsfromenv-never-treats-a-setting-about-a-secret-as-one-a-bare-integ",
       "file": "certification-bundle-redaction.test.ts",
       "line": 157,
+      "title": "evidenceSecretsFromEnv never treats a setting ABOUT a secret as one: a bare integer or a value under 8 characters is not swept from the environment",
+      "explicitId": "openwop.it.certification-bundle-redaction.setting-about-a-secret-not-swept",
+      "citations": [
+        {
+          "section": "RFC 0148 §C",
+          "requirement": "a bare-integer setting whose NAME mentions a secret (a rotation overlap, a TTL) MUST NOT be swept as a credential"
+        },
+        {
+          "section": "RFC 0148 §C",
+          "requirement": "a value under 8 characters is not credential-shaped and is not swept from the environment"
+        },
+        {
+          "section": "RFC 0148 §C",
+          "requirement": "a credential-shaped value is still swept"
+        },
+        {
+          "section": "RFC 0148 §C",
+          "requirement": "a credential the emitter hands over explicitly is scrubbed whatever its shape"
+        },
+        {
+          "section": "RFC 0148 §C",
+          "requirement": "scrubbing MUST NOT corrupt a digest the bundle carries"
+        }
+      ]
+    },
+    {
+      "id": "openwop.it.certification-bundle-redaction.evidencesecretsfromenv-never-scrubs-a-key-identifier-id-variables-are-excluded-a",
+      "file": "certification-bundle-redaction.test.ts",
+      "line": 180,
       "title": "evidenceSecretsFromEnv never scrubs a key IDENTIFIER: *_ID variables are excluded and `except` values stay visible whichever variable carried them",
       "explicitId": "openwop.it.certification-bundle-redaction.key-identifier-never-scrubbed",
       "citations": [
@@ -7189,7 +7218,7 @@
     {
       "id": "openwop.it.certification-bundle-redaction.verifybundlev2-rejects-a-bundle-carrying-the-conformance-canary-anywhere-and-acc",
       "file": "certification-bundle-redaction.test.ts",
-      "line": 177,
+      "line": 200,
       "title": "verifyBundleV2 REJECTS a bundle carrying the conformance canary anywhere, and accepts the scrubbed twin",
       "explicitId": "openwop.it.certification-bundle-redaction.verifybundlev2-rejects-a-bundle-carrying-the-conformance-canary-anywhere-and-acc",
       "citations": [
@@ -7202,7 +7231,7 @@
     {
       "id": "openwop.it.certification-bundle-redaction.the-canary-in-a-discovery-document-key-is-still-a-rejection-keys-are-evidence",
       "file": "certification-bundle-redaction.test.ts",
-      "line": 192,
+      "line": 215,
       "title": "the canary in a discovery-document KEY is still a rejection (keys are evidence)",
       "explicitId": "openwop.it.certification-bundle-redaction.the-canary-in-a-discovery-document-key-is-still-a-rejection-keys-are-evidence",
       "citations": [
@@ -7215,7 +7244,7 @@
     {
       "id": "openwop.it.certification-bundle-redaction.the-target-configuration-digest-is-a-digest-the-api-key-is-not-a-substring-of-an",
       "file": "certification-bundle-redaction.test.ts",
-      "line": 198,
+      "line": 221,
       "title": "the target-configuration digest is a digest: the API key is not a substring of any string in the document",
       "explicitId": "openwop.it.certification-bundle-redaction.the-target-configuration-digest-is-a-digest-the-api-key-is-not-a-substring-of-an",
       "citations": [
@@ -9150,7 +9179,7 @@
     {
       "id": "openwop.it.context-budget-transcript-bound.bounds-the-per-turn-transcript-to-transcripttokenbudget-with-an-internally-consi",
       "file": "context-budget-transcript-bound.test.ts",
-      "line": 135,
+      "line": 89,
       "title": "bounds the per-turn transcript to transcriptTokenBudget with an internally-consistent, recent-tail accounting",
       "explicitId": "openwop.it.context-budget-transcript-bound.bounds-the-per-turn-transcript-to-transcripttokenbudget-with-an-internally-consi",
       "citations": [
@@ -9158,22 +9187,6 @@
           "section": "RFC 0111",
           "requirement": "tokenCounter MUST be advertised when transcriptTokenBudget is present (schema if/then)"
         },
-        {
-          "section": "RFC 0111",
-          "requirement": "POST /v1/runs MUST return a runId"
-        },
-        {
-          "section": "host-sample-test-seams.md §14",
-          "requirement": "the transcript-window seam MUST return 200 for a valid iteration"
-        },
-        {
-          "section": "host-sample-test-seams.md §14",
-          "requirement": "the seam MUST return { tokenCounter, tokenCount, eventIds, summarizedRanges }"
-        },
-        {
-          "section": "host-sample-test-seams.md §14",
-          "requirement": "a wired transcript-window seam MUST report at least one orchestrator iteration"
-        },
         {
           "section": "RFC 0111",
           "requirement": null,
@@ -9181,21 +9194,24 @@
         },
         {
           "section": "RFC 0111",
-          "requirement": null,
-          "interpolated": true
+          "requirement": "the create response MUST carry a runId"
         },
         {
-          "section": "RFC 0111 §\"Conformance seam\"",
+          "section": "host-sample-test-seams.md §14",
           "requirement": null,
           "interpolated": true
         },
         {
-          "section": "RFC 0111 §\"Conformance seam\"",
+          "section": "host-sample-test-seams.md §14",
           "requirement": null,
           "interpolated": true
         },
         {
-          "section": "RFC 0111",
+          "section": "host-sample-test-seams.md §14",
+          "requirement": "a wired transcript-window seam MUST report at least one orchestrator iteration"
+        },
+        {
+          "section": "RFC 0111 §\"Conformance seam\" (b)/(c), host-sample-test-seams.md §14",
           "requirement": null,
           "interpolated": true
         },
@@ -9209,13 +9225,18 @@
     {
       "id": "openwop.it.context-summarization-replay.replay-reuses-the-recorded-context-summarized-summaryref-never-re-summarizes",
       "file": "context-summarization-replay.test.ts",
-      "line": 99,
+      "line": 106,
       "title": "replay reuses the recorded context.summarized summaryRef — never re-summarizes",
       "explicitId": "openwop.it.context-summarization-replay.replay-reuses-the-recorded-context-summarized-summaryref-never-re-summarizes",
       "citations": [
         {
-          "section": "rest-endpoints.md POST /v1/runs/{runId}:fork",
-          "requirement": "POST /v1/runs MUST return a runId"
+          "section": "RFC 0111",
+          "requirement": null,
+          "interpolated": true
+        },
+        {
+          "section": "rest-endpoints.md POST /v1/runs",
+          "requirement": "the create response MUST carry a runId"
         },
         {
           "section": "rest-endpoints.md POST /v1/runs/{runId}:fork",
@@ -9228,6 +9249,14 @@
         {
           "section": "RFC 0111 §\"Replay determinism\"",
           "requirement": "a replay fork MUST reuse the recorded context.summarized summaryRef (never re-summarize to a different transcript)"
+        },
+        {
+          "section": "RFC 0111 §\"Replay determinism\"",
+          "requirement": "the source run summarized, so its transcript windows MUST carry the summary text it fed"
+        },
+        {
+          "section": "RFC 0111 §\"Replay determinism\"",
+          "requirement": "the replay MUST feed the model the recorded summary text, byte for byte — never a re-summarization"
         }
       ]
     },
@@ -27657,7 +27686,7 @@
     {
       "id": "openwop.it.v2-bound-id-kinds.subscriptionid-post-webhooks-mints-a-bound-webhookid-the-projected-segment-is-ac",
       "file": "v2-bound-id-kinds.test.ts",
-      "line": 89,
+      "line": 91,
       "title": "subscriptionId: POST /webhooks mints a bound webhookId, the projected segment is accepted, and a foreign tenant segment is refused 403",
       "explicitId": "openwop.requirement.0187.bound-id-kinds.webhook",
       "citations": [
@@ -27684,7 +27713,7 @@
     {
       "id": "openwop.it.v2-bound-id-kinds.subscriptionid-the-webhookid-a-host-emits-on-a-delivery-is-the-one-it-minted",
       "file": "v2-bound-id-kinds.test.ts",
-      "line": 140,
+      "line": 142,
       "title": "subscriptionId: the webhookId a host EMITS on a delivery is the one it minted",
       "explicitId": "openwop.requirement.0187.bound-id-kinds.webhook-emitted",
       "citations": [
@@ -27703,7 +27732,7 @@
     {
       "id": "openwop.it.v2-bound-id-kinds.interruptid-the-id-a-suspension-puts-on-the-wire-is-bound",
       "file": "v2-bound-id-kinds.test.ts",
-      "line": 192,
+      "line": 194,
       "title": "interruptId: the id a suspension puts on the wire is bound",
       "explicitId": null,
       "citations": []
@@ -27711,7 +27740,7 @@
     {
       "id": "openwop.it.v2-bound-id-kinds.effectid-every-row-get-runs-runid-effects-serves-is-bound",
       "file": "v2-bound-id-kinds.test.ts",
-      "line": 213,
+      "line": 215,
       "title": "effectId: every row GET /runs/{runId}/effects serves is bound",
       "explicitId": "openwop.requirement.0187.bound-id-kinds.per-kind",
       "citations": [
@@ -27724,7 +27753,7 @@
     {
       "id": "openwop.it.v2-bound-id-kinds.deliveryid-every-record-in-the-dead-letter-read-carries-a-bound-id-in-the-caller",
       "file": "v2-bound-id-kinds.test.ts",
-      "line": 228,
+      "line": 230,
       "title": "deliveryId: every record in the dead-letter read carries a bound id in the caller's tenant",
       "explicitId": "openwop.requirement.0187.bound-id-kinds.per-kind",
       "citations": [
@@ -28647,7 +28676,7 @@
     {
       "id": "openwop.it.v2-durability-recovery.accepted-work-survives-a-kill-before-dispatch-and-dispatches-on-resume",
       "file": "v2-durability-recovery.test.ts",
-      "line": 280,
+      "line": 282,
       "title": "accepted work survives a kill before dispatch and dispatches on resume",
       "explicitId": "openwop.requirement.0158.kill-after-accept",
       "citations": [
@@ -28661,7 +28690,7 @@
     {
       "id": "openwop.it.v2-durability-recovery.work-executing-at-a-real-process-death-is-never-reported-complete-and-resumes",
       "file": "v2-durability-recovery.test.ts",
-      "line": 334,
+      "line": 336,
       "title": "work executing at a real process death is never reported complete, and resumes",
       "explicitId": "openwop.requirement.0158.kill-during-execution",
       "citations": [
@@ -28680,7 +28709,7 @@
     {
       "id": "openwop.it.v2-durability-recovery.the-same-accepted-work-delivered-twice-fires-each-effect-exactly-once",
       "file": "v2-durability-recovery.test.ts",
-      "line": 390,
+      "line": 392,
       "title": "the same accepted work delivered twice fires each effect exactly once",
       "explicitId": "openwop.requirement.0158.duplicate-delivery",
       "citations": [
@@ -28699,7 +28728,7 @@
     {
       "id": "openwop.it.v2-durability-recovery.the-declared-recovery-bound-is-derived-from-the-mechanism-that-enforces-it",
       "file": "v2-durability-recovery.test.ts",
-      "line": 524,
+      "line": 526,
       "title": "the declared recovery bound is derived from the mechanism that enforces it",
       "explicitId": "openwop.requirement.0158.bound-is-derived",
       "citations": [
@@ -28713,7 +28742,7 @@
     {
       "id": "openwop.it.v2-durability-recovery.deterministically-failing-work-reaches-a-terminal-state-and-stops-being-retried",
       "file": "v2-durability-recovery.test.ts",
-      "line": 555,
+      "line": 557,
       "title": "deterministically failing work reaches a terminal state and stops being retried",
       "explicitId": "openwop.requirement.0158.poison-exhaustion",
       "citations": [
@@ -32080,7 +32109,7 @@
     {
       "id": "openwop.it.v2-revocation-honored.a-revoked-next-request-credential-is-refused-on-the-next-request-with-credential",
       "file": "v2-revocation-honored.test.ts",
-      "line": 38,
+      "line": 40,
       "title": "a revoked next-request credential is refused on the next request with credential_revoked",
       "explicitId": "openwop.requirement.0170.revocation-honored",
       "citations": [
@@ -32935,7 +32964,7 @@
     {
       "id": "openwop.it.v2-terminal-event-once.every-run-observed-to-a-terminal-state-has-exactly-one-terminal-event-and-no-for",
       "file": "v2-terminal-event-once.test.ts",
-      "line": 81,
+      "line": 83,
       "title": "every run observed to a terminal state has exactly one terminal event and no forward execution after it",
       "explicitId": "openwop.requirement.0194.terminal-once",
       "citations": [
@@ -32953,7 +32982,7 @@
     {
       "id": "openwop.it.v2-terminal-event-once.work-delivered-twice-still-leaves-one-terminal-event-and-no-forward-execution-af",
       "file": "v2-terminal-event-once.test.ts",
-      "line": 109,
+      "line": 111,
       "title": "work delivered twice still leaves one terminal event and no forward execution after it",
       "explicitId": "openwop.requirement.0194.terminal-once.duplicate-delivery",
       "citations": [
@@ -33192,7 +33221,7 @@
     {
       "id": "openwop.requirement.0215.no-head-of-line",
       "file": "v2-webhook-delivery-isolation.test.ts",
-      "line": 109,
+      "line": 111,
       "title": null,
       "explicitId": "openwop.requirement.0215.no-head-of-line",
       "citations": [
@@ -33233,7 +33262,7 @@
     {
       "id": "openwop.it.v2-webhook-delivery-shape.a-major-2-subscriber-receives-the-v2-rendering-the-delivery-validates-and-run-st",
       "file": "v2-webhook-delivery-shape.test.ts",
-      "line": 217,
+      "line": 219,
       "title": "a major-2 subscriber receives the v2 rendering: the delivery validates, and run.started.owner carries subject, never principal",
       "explicitId": "openwop.requirement.0171.webhook-delivery-shape",
       "citations": [
@@ -33264,7 +33293,7 @@
     {
       "id": "openwop.it.v2-webhook-delivery-shape.a-major-1-subscriber-still-receives-the-v1-rendering-the-v1-wire-does-not-move-m",
       "file": "v2-webhook-delivery-shape.test.ts",
-      "line": 237,
+      "line": 239,
       "title": "a major-1 subscriber still receives the v1 rendering — the v1 wire does not move mid-overlap",
       "explicitId": "openwop.requirement.0171.webhook-delivery-shape",
       "citations": [
@@ -33282,7 +33311,7 @@
     {
       "id": "openwop.it.v2-webhook-delivery-shape.a-seeded-era-2-run-is-delivered-projected-too-the-fan-out-branch-a-fresh-run-can",
       "file": "v2-webhook-delivery-shape.test.ts",
-      "line": 261,
+      "line": 263,
       "title": "a seeded era-2 run is delivered projected too — the fan-out branch a fresh run cannot reach (seam-gated)",
       "explicitId": "openwop.requirement.0171.webhook-delivery-shape",
       "citations": [
@@ -33300,7 +33329,7 @@
     {
       "id": "openwop.it.v2-webhook-durable-delivery.a-failed-attempt-is-retried-and-the-event-is-delivered-at-least-once",
       "file": "v2-webhook-durable-delivery.test.ts",
-      "line": 256,
+      "line": 258,
       "title": "a failed attempt is retried and the event is delivered at least once",
       "explicitId": "openwop.requirement.0173.webhook-durable-delivery",
       "citations": [
@@ -33332,7 +33361,7 @@
     {
       "id": "openwop.it.v2-webhook-durable-delivery.an-exhausted-delivery-is-dead-lettered-never-dropped",
       "file": "v2-webhook-durable-delivery.test.ts",
-      "line": 358,
+      "line": 360,
       "title": "an exhausted delivery is dead-lettered, never dropped",
       "explicitId": "openwop.requirement.0173.webhook-durable-delivery.dead-letter",
       "citations": [
@@ -33366,7 +33395,7 @@
     {
       "id": "openwop.it.v2-webhook-durable-delivery.the-dead-letter-read-is-served-and-its-records-carry-no-payload",
       "file": "v2-webhook-durable-delivery.test.ts",
-      "line": 452,
+      "line": 454,
       "title": "the dead-letter read is served and its records carry no payload",
       "explicitId": "openwop.requirement.0188.dead-letter-read",
       "citations": [
@@ -33379,7 +33408,7 @@
     {
       "id": "openwop.it.v2-webhook-durable-delivery.a-dead-letter-record-carries-no-delivered-payload",
       "file": "v2-webhook-durable-delivery.test.ts",
-      "line": 473,
+      "line": 475,
       "title": "a dead-letter record carries no delivered payload",
       "explicitId": "openwop.requirement.0188.dead-letter-content-free",
       "citations": [
@@ -33723,7 +33752,7 @@
     {
       "id": "openwop.it.v2-webhook-unregister-stops-delivery.no-attempt-for-an-unregistered-subscription-starts-after-the-204-while-its-contr",
       "file": "v2-webhook-unregister-stops-delivery.test.ts",
-      "line": 83,
+      "line": 85,
       "title": "no attempt for an unregistered subscription starts after the 204, while its control keeps being retried",
       "explicitId": "openwop.requirement.0215.unregister-stops-delivery",
       "citations": [
@@ -34537,7 +34566,7 @@
     {
       "id": "openwop.it.webhook-signed-delivery.host-posts-run-events-to-subscriber-with-valid-x-openwop-signature",
       "file": "webhook-signed-delivery.test.ts",
-      "line": 168,
+      "line": 170,
       "title": "host POSTs run events to subscriber with valid X-openwop-Signature",
       "explicitId": "openwop.it.webhook-signed-delivery.host-posts-run-events-to-subscriber-with-valid-x-openwop-signature",
       "citations": [
@@ -34588,7 +34617,7 @@
     {
       "id": "openwop.it.webhook-signed-delivery.control-with-a-tunnel-wired-the-host-still-refuses-the-loopback-receiver",
       "file": "webhook-signed-delivery.test.ts",
-      "line": 423,
+      "line": 425,
       "title": "control: with a tunnel wired, the host still refuses the loopback receiver",
       "explicitId": "openwop.it.webhook-signed-delivery.control-with-a-tunnel-wired-the-host-still-refuses-the-loopback-receiver",
       "citations": [

package/scenario-majors.json

@@ -3,7 +3,7 @@
   "counts": {
     "files": 561,
     "v1": 449,
-    "v2": 121
+    "v2": 123
   },
   "majors": {
     "a2a-1-0-agent-card.test.ts": [
@@ -410,10 +410,12 @@
       1
     ],
     "context-budget-transcript-bound.test.ts": [
-      1
+      1,
+      2
     ],
     "context-summarization-replay.test.ts": [
-      1
+      1,
+      2
     ],
     "contract-provenance.test.ts": [
       1