@openwop/openwop-conformance 1.2.0 → 1.3.0

package/CHANGELOG.md

@@ -1,5 +1,70 @@
 # `@openwop/openwop-conformance` Changelog
 
+## [1.3.0] — 2026-05-19
+
+Minor bump per `PUBLISHING.md` §"Versioning alignment" — conformance scenario + fixture additions land on a minor. Closes the conformance-republish acceptance gate on RFC 0024 (§"Active → Accepted" criterion b) and bundles the wider behavioral-coverage push that converted ~50 `it.todo()` placeholders into live behavioral assertions across the RFC 0013 / 0016 / 0017 / 0022 / 0023 / 0024 surface.
+
+### Added — RFC 0024 streaming reasoning (headline)
+
+- **`conformance/fixtures/conformance-agent-reasoning-streaming.json`** — new RFC 0024 fixture. Drives the `core.conformance.mock-agent` typeId via `mockReasoning.streamChunks` (extended schema entry per RFC 0024 §"Conformance") to emit a deterministic sequence of `agent.reasoning.delta` events followed by exactly one closing `agent.reasoned`. The closing event's `reasoning` field MUST equal the concatenation of all chunks; deltas MUST carry monotonically-increasing `sequence` starting at 0 and MUST all precede the closing event in the event log.
+- **`conformance/src/scenarios/agentReasoningStreaming.test.ts`** — new scenario gating on `capabilities.agents.supported: true` AND `capabilities.agents.reasoning.streaming: true` AND `getReasoningVerbosity() !== 'off'`. Asserts: (a) at least one `agent.reasoning.delta` event surfaces; (b) deltas appear in monotonic sequence order; (c) concatenated deltas equal the closing `agent.reasoned.reasoning` byte-for-byte; (d) the last delta's `eventLogIdx` strictly precedes the closing event's `eventLogIdx`; (e) the closing `agent.reasoned.causationId` chains correctly per `replay.md` §"Determinism with non-deterministic agents". Hosts without the streaming flag skip cleanly.
+- **`schemas/core-conformance-mock-agent-config.schema.json`** updated with the `streamChunks` field on `mockReasoning` per RFC 0024 §"Conformance".
+- **`spec/v1/node-packs.md`** Authorized-Emitters table for the `agent.*` family extended with `agent.reasoning.delta` (RFC 0024 addendum).
+- **`spec/v1/capabilities.md`** §`agents.reasoning` documents the `streaming: boolean` field (default `false`); spec text + JSON Schema additions in `schemas/capabilities.schema.json`.
+
+### Added — RFC 0022 §C variable-mapping refusal contract
+
+- **3 new fixture variants** for the unset-default + per-worker-override + subWorkflow-unset edges: `conformance-dispatch-input-mapping-no-default.json`, `conformance-dispatch-per-worker-override.json`, `conformance-subworkflow-input-mapping-no-default.json`. Catalog rows added to `conformance/fixtures.md`.
+- **`dispatch-input-mapping.test.ts`** + **`subworkflow-input-mapping.test.ts`** + **`dispatch-cross-worker-handoff.test.ts`** behavioral assertions for HVMAP-1a-null + HVMAP-1a-refusal + HVMAP-1c-override + HVMAP-2-unset + HVMAP-2-refusal — exercising both the projection paths and the `validation_error` refusal contract from RFC 0022 §C when a host does NOT advertise the gating capability.
+
+### Added — Thread C: child-lifecycle fixtures
+
+- **`conformance-dispatch-cancellable-child.json`** + **`conformance-dispatch-deterministic-fail-child.json`** support `dispatch-cross-worker-handoff` outputMapping skip-on-failure semantics and child-cancellation propagation tests. 2 new behavioral assertions in `dispatch-output-mapping.test.ts`.
+
+### Added — Threads E.1 / E.2 / E.3: event-log query + OTel + debug-bundle seams
+
+- **`event-log-query.test.ts`** — `eventLogQuery({fromSeq, toSeq, types[], causationId})` projection seam exercised via 12 behavioral assertions previously `it.todo()`. Gates on the host's advertisement of `host.eventLog.query.supported: true`.
+- **`otel-trace-propagation-subworkflow.test.ts`** — W3C traceparent propagation across the dispatch boundary; reads context from the closing `runOrchestrator.decided`'s tracecontext and asserts the child `core.subWorkflow` inherits the parent trace.
+- **`debug-bundle.test.ts`** + **`debug-bundle-redaction.test.ts`** — assert the host's debug-bundle endpoint surfaces the redacted event projection per SR-1 (no BYOK credential material in the bundle's event-log slice).
+
+### Added — Thread F: stream/search/blob/queue/table/cache
+
+- **`stream.test.ts`** + **`search-knn.test.ts`** + **`blob-presign.test.ts`** — 3 todos converted to behavioral via the `/v1/host/sample/test/surface` seam.
+- **`queue-publish-consume-roundtrip.test.ts`** + **`queue-ack-nack-dlq.test.ts`** — full RFC 0017 §B point 2 ack/nack/DLQ state machine asserted end-to-end (consume → ack drop, consume → nack(requeue=true) re-queues at head with incremented deliveryCount, consume → deadLetter routes to `<subject>.dlq` with `{ original, deadLetterReason }` wrapper).
+- **`table-cursor-pagination.test.ts`** + **`table-schema-enforcement.test.ts`** — RFC 0016 §B points 2+3: first-insert declares per-column types, divergent-type insert throws `table_schema_violation`; opaque base64 cursor pagination with `nextCursor: null` on the final page.
+- **`cache-ttl-expiry.test.ts`** + **`kv-ttl-expiry.test.ts`** — TTL expiry assertions via the host-side test seam.
+- **`sql-transaction-atomicity.test.ts`** + **`vector-knn-roundtrip.test.ts`** — round-trip + atomicity assertions on the SQL + vector surfaces.
+- **`replay-llm-cache-key.test.ts`** — `replay.md` §"LLM cache-key recipe" §B asserted via the new `llm-cache-key` host seam (`POST /v1/host/sample/test/llm-cache-key`). Cross-host parity (B-suffixed runs against `OPENWOP_BASE_URL_B`) stays deferred awaiting cross-host plumbing.
+
+### Added — AI envelope behavioral assertions
+
+- **`aiEnvelope.{schemaDrift,redaction,correlationReplay}.test.ts`** — 9 todos converted to behavioral via the extended `host/envelopeAcceptor.ts` seam: (1) `schemaVersionFloor` + `envelopeStrictness` for below-floor refusal under `strict`; (2) `priorCorrelations` for same-correlationId re-emission returning the cached outcome AND same-correlationId-different-type refusing with `envelope_correlation_conflict`; (3) `byokCanaries` for recursive deep substitution of canary values with the canonical SR-1 `[REDACTED:<secretId>]` marker per `agent-memory.md:66`.
+
+### Added — Driver helpers + opt-out axes
+
+- **`conformance/src/lib/host-toggle.ts`** (NEW) — driver helpers `setHostCapability(name, value)`, `resetHostCapabilities()`, `isToggleAvailable()`. All operations soft-skip on HTTP 404 so non-seam hosts keep advertisement-shape coverage intact; scenarios MUST reset in `finally{}`. Backed by `POST /v1/host/sample/test/capability-toggle` on the reference workflow-engine.
+- **`OPENWOP_OPTED_OUT_FIXTURES`** + **`OPENWOP_OPTED_OUT_SCENARIOS`** — two new operator-side env axes (CSV + trailing-`*` glob support) for hosts that auto-load every `conformance-*.json` on disk but do NOT implement the gated feature for some of them. Symmetric to the existing `OPENWOP_OPTED_OUT_PROFILES`. `conformance/src/lib/env.ts` + `conformance/src/lib/fixtures.ts` carry the helpers; `fixtures-gating.test.ts` adds 12 parser-edge-case tests covering CSV + glob semantics.
+
+### Added — RFC 0013 Phase 3 reference host expansion
+
+- **`host-in-memory`** gains the Phase 3 surface; new scenarios assert the in-memory host now passes the same Phase 3 advertisement-shape suite as the SQLite host. Per `INTEROP-MATRIX.md`, both reference hosts now advertise Phase 3.
+
+### Tightened
+
+- **Persisted envelope-correlation dedup seam** (`be89f4d`) — `priorCorrelations` now reads from a real persisted store on the reference workflow-engine, not just an in-flight map. Scenario assertions strengthened to survive a process restart between the original `accept` and the replay.
+- **`apps/workflow-engine/.../host/capabilityOverlay.ts`** (NEW reference-host file) — process-local overlay over advertised capability flags, consulted by `validateDefinition` at workflow-register time. Defaults `agents.dispatchMapping` and `subWorkflow.inputMapping` to `false` per the honest-advertisement principle (the reference workflow-engine implements the RFC 0022 §C refusal contract but does NOT yet execute the mapping itself).
+- **`examples/hosts/sqlite/src/server.ts` artifact-route stub** — `checkAuth` now runs BEFORE any 404 across HEAD/POST/PUT/DELETE; non-GET → 405 method_not_allowed (per `rest-endpoints.md §getArtifact` advertising GET only); GET → 404 not_found. Closes an unauthenticated runId/artifactId probe surface that the prior catch-all 404 left open.
+
+### Compatibility
+
+**Additive** per `COMPATIBILITY.md §2.1`. New fixtures, new scenarios, new driver helpers, new env axes — no existing scenario was relaxed; no existing fixture was renamed or its semantics changed; no existing host-side contract changed. RFC 0024's new event type `agent.reasoning.delta` is gated on `capabilities.agents.reasoning.streaming: true` (default `false`); hosts that omit it advertise the existing non-streaming contract and skip the new scenario cleanly. The capability-overlay toggle endpoint and the new sample-namespaced test seams live under the sample-namespaced test-seam prefix per `host-extensions.md` §"Canonical prefixes" — explicitly not part of the v1 wire contract.
+
+### Notes
+
+- This bump does NOT change the v1.0 spec corpus surface; `@openwop/openwop@1.1.x`, `openwop-client@1.1.x`, and the Go SDK stay locked to their current versions per `PUBLISHING.md` §"Versioning alignment" ("Conformance scenario addition | @openwop/openwop-conformance minor bump; other artifacts unaffected").
+- Trigger: push `openwop-conformance/v1.3.0` per `.github/workflows/openwop-publish.yml` (only the `publish-conformance` job runs).
+- RFC 0024 §"Active → Accepted" — this republish resolves criterion (b) ("next `@openwop/openwop-conformance` republish carrying the new fixture+scenario to downstream consumers"). Criterion (a) (external host advertisement evidence) remains open.
+
 ## [1.2.0] — 2026-05-18
 
 Minor bump per `PUBLISHING.md` §"Versioning alignment" — conformance scenario additions land on a minor. Captures the RFC 0022 + RFC 0023 fixture/scenario surface that landed across `cf7df05`, `02a84e1`, `a8a8594`, `f94d2e1`, `87c5de7`, `22d9f92`, `a025a85`, and `a65ea0e`.

package/README.md

@@ -92,7 +92,7 @@ Exit code is non-zero on any failed assertion.
 
 ## What's Covered
 
-The current suite has 157 scenario files under `src/scenarios/`. 2026-05-18 (RFC 0022 `Draft` — runtime variable mapping) added four `it.todo()` placeholder scenarios covering the new mapping surfaces on `core.dispatch` (§A — `dispatch-input-mapping.test.ts`, `dispatch-output-mapping.test.ts`, `dispatch-cross-worker-handoff.test.ts`) and `core.subWorkflow` (§B — `subworkflow-input-mapping.test.ts`). Gated on `capabilities.agents.dispatchMapping` (dispatch trio) and `capabilities.subWorkflow.inputMapping` (subWorkflow). Promote to live assertions when RFC 0022 reaches `Active` + a reference host advertises the matching flags. 2026-05-17 (RFC 0003 §D handoff-schema enforcement, HV-1) added `agentPackHandoffSchemaValidation.test.ts` — verifies the host validates dispatch payloads against `handoff.taskSchemaRef` AND return payloads against `handoff.returnSchemaRef` per RFC 0003 §D. Paired with the new `agent-pack-handoff-schema-enforcement` row in `SECURITY/invariants.yaml`. 2026-05-17 (AI Envelope gap-closure, DRAFT v1.x — `spec/v1/ai-envelope.md`) added 7 advertisement-shape scenarios with `it.todo()` behavioral placeholders gated on `capabilities.envelopeContracts.advertised: true`: `aiEnvelope.universalKinds.test.ts`, `aiEnvelope.schemaDrift.test.ts`, `aiEnvelope.correlationReplay.test.ts`, `aiEnvelope.contractRefusal.test.ts`, `aiEnvelope.trustBoundaryPropagation.test.ts`, `aiEnvelope.redaction.test.ts`, `aiEnvelope.capBreached.test.ts`. Paired with the new `envelope-redaction-sr-1-carry-forward` row in `SECURITY/invariants.yaml`. 2026-05-17 (post-publish hardening, deep audit of `core.openwop.agents`) added `agents-run-tool-allowlist.test.ts` — server-free scenario locking in the `core.openwop.agents@1.0.1` safety-fix that closes `OPENWOP-AUDIT-2026-003` (function-typed `tool.handler` properties rejected at `validateTools()` with `INVALID_TOOL_DECLARATION`; tool-driven runs require `ctx.agentRuntime`; tool-less safe fallback preserved). Paired with the new `agents-run-no-raw-handler` row in `SECURITY/invariants.yaml`. Same-day post-publish hardening added `idempotency-key-determinism.test.ts` — server-free scenario locking in the `core.openwop.http@1.1.2` determinism safety-fix (default `composite` mode produces deterministic keys in `(runId, nodeId, payload)`; removed `uuid` mode rejects with `CONFIG_INVALID`; cross-impl vector test lets third-party reimplementations verify wire agreement). Paired with the new `idempotency-key-deterministic` row in `SECURITY/invariants.yaml`. 2026-05-17 (Phase 3 of RFC 0013) added three server-free scenarios exercising the reference workflow-chain expansion library (`conformance/src/lib/workflow-chain-expansion.ts`): `workflow-chain-expansion.test.ts` (parameter substitution + node id collision avoidance + edge rewriting + capability propagation + runtime-invariance contract), `workflow-chain-unresolvable-typeid.test.ts` (rejection with `chain_unresolvable_typeid` when a chain references an unknown typeId), and `workflow-chain-pack-signature-verification.test.ts` (Ed25519 verification recipe reuse from `node-packs.md §Signing`). Earlier that day (Phase 1) added `workflow-chain-pack-manifest-validation.test.ts` — server-free schema-validation scenario covering the new `workflow-chain-pack-manifest.schema.json` (positive sample + two negatives: kind/contents mismatch and invalid `chainId`). Closes RFC 0013 (`Workflow-chain packs`, `Draft`) Phases 1 + 3 alongside the new `spec/v1/workflow-chain-packs.md`, the `Capabilities.workflowChainPacks` block, and the registry build-index/conformance-check `kind` routing from Phase 2. Earlier that day, the suite added 27 `it.todo()` placeholder scenarios paired with RFCs 0014-0020 (host capability surfaces — fs, kvStorage, tableStorage, queueBus, sql/vector/search, blob/cache, mcp.serverMount). These promote to live assertions when each RFC reaches `Active` + the matching capability block lands in `schemas/capabilities.schema.json` + a reference host advertises the capability. Earlier additions include 18 Multi-Agent Shift scenarios (Phases 1-5) added 2026-05-10, the `registry-public.test.ts` public-registry healthcheck added 2026-05-11 (opt-in via `OPENWOP_TEST_PUBLIC_REGISTRY=true`), the `replay-llm-cache-key.test.ts` placeholder added 2026-05-11 (three `it.todo()` cases for the cross-host LLM cache-key recipe per `replay.md` §"LLM cache-key recipe"), the two `production-*.test.ts` scenarios added 2026-05-11 for the `openwop-production` profile per RFC 0009 (`production-backpressure.test.ts`, `production-retention-expiry.test.ts`), the four `auth-*.test.ts` scenarios added 2026-05-11/12 for the production-auth profiles per RFC 0010 (`auth-api-key-rotation.test.ts`, `auth-oauth2-client-credentials.test.ts`, `auth-oidc-user-bearer.test.ts`, `auth-mtls.test.ts` (opt-in via `OPENWOP_TEST_MTLS=1`)), `replay-retention-expiry.test.ts` added 2026-05-12 (capability shape + 410/422 envelope per `replay.md` §"Retention and garbage collection"), `bulk-cancel.test.ts` added 2026-05-12 (Phase B close-out of R1 — `POST /v1/runs:bulk-cancel`), the two Phase H launch-blocker advertisement-contract scenarios added 2026-05-12 (`mcp-toolcall-redaction.test.ts` for the MCP-1 invariant per `host-capabilities.md §host.mcp` + `threat-model-prompt-injection.md §UNTRUSTED`, and `http-client-ssrf.test.ts` for the SSRF + body-size cap advertisement contract on `capabilities.httpClient`), the `wasm-pack-abi-version-rejection.test.ts` Track 7 scenario added 2026-05-12 for the ABI-mismatch positive path via the `vendor.openwop.misbehaving-abi` pack per RFC 0008 §H, the `otel-trace-propagation-subworkflow.test.ts` Track 11 close-out added 2026-05-13 (parent + child run spans share the inbound traceparent's traceId across the `core.subWorkflow` dispatch boundary), and the three RFC 0012 (Memory Compaction Profile, `Active`) scenarios added 2026-05-13/14: `memory-compaction-sr1-carry-forward.test.ts` (load-bearing SR-1 §D), `memory-compaction-event-emitted.test.ts` (canonical §B payload shape), and `memory-compaction-provenance-tag.test.ts` (soft assertion on §C `compacted-from:<id>` convention). All three gate on `capabilities.memory.compaction.supported` + the host's test seam at `/v1/test/memory/{seed,compact}` (Postgres reference host enables both via `OPENWOP_MEMORY_COMPACTION=true OPENWOP_TEST_TRIGGER_COMPACTION=true`). 2026-05-15 (gap-closure CF-3) added `interrupt-token-matrix.test.ts` (malformed / unknown / replay / cross-run-id paths on `GET|POST /v1/interrupts/{token}`). The maintained scenario-to-spec map lives in [`coverage.md`](./coverage.md); this README keeps the operator quickstart and the historical scenario notes below.
+The current suite has 160 scenario files under `src/scenarios/`. 2026-05-18 (RFC 0022 `Draft` — runtime variable mapping) added four `it.todo()` placeholder scenarios covering the new mapping surfaces on `core.dispatch` (§A — `dispatch-input-mapping.test.ts`, `dispatch-output-mapping.test.ts`, `dispatch-cross-worker-handoff.test.ts`) and `core.subWorkflow` (§B — `subworkflow-input-mapping.test.ts`). Gated on `capabilities.agents.dispatchMapping` (dispatch trio) and `capabilities.subWorkflow.inputMapping` (subWorkflow). Promote to live assertions when RFC 0022 reaches `Active` + a reference host advertises the matching flags. 2026-05-17 (RFC 0003 §D handoff-schema enforcement, HV-1) added `agentPackHandoffSchemaValidation.test.ts` — verifies the host validates dispatch payloads against `handoff.taskSchemaRef` AND return payloads against `handoff.returnSchemaRef` per RFC 0003 §D. Paired with the new `agent-pack-handoff-schema-enforcement` row in `SECURITY/invariants.yaml`. 2026-05-17 (AI Envelope gap-closure, DRAFT v1.x — `spec/v1/ai-envelope.md`) added 7 advertisement-shape scenarios with `it.todo()` behavioral placeholders gated on `capabilities.envelopeContracts.advertised: true`: `aiEnvelope.universalKinds.test.ts`, `aiEnvelope.schemaDrift.test.ts`, `aiEnvelope.correlationReplay.test.ts`, `aiEnvelope.contractRefusal.test.ts`, `aiEnvelope.trustBoundaryPropagation.test.ts`, `aiEnvelope.redaction.test.ts`, `aiEnvelope.capBreached.test.ts`. Paired with the new `envelope-redaction-sr-1-carry-forward` row in `SECURITY/invariants.yaml`. 2026-05-17 (post-publish hardening, deep audit of `core.openwop.agents`) added `agents-run-tool-allowlist.test.ts` — server-free scenario locking in the `core.openwop.agents@1.0.1` safety-fix that closes `OPENWOP-AUDIT-2026-003` (function-typed `tool.handler` properties rejected at `validateTools()` with `INVALID_TOOL_DECLARATION`; tool-driven runs require `ctx.agentRuntime`; tool-less safe fallback preserved). Paired with the new `agents-run-no-raw-handler` row in `SECURITY/invariants.yaml`. Same-day post-publish hardening added `idempotency-key-determinism.test.ts` — server-free scenario locking in the `core.openwop.http@1.1.2` determinism safety-fix (default `composite` mode produces deterministic keys in `(runId, nodeId, payload)`; removed `uuid` mode rejects with `CONFIG_INVALID`; cross-impl vector test lets third-party reimplementations verify wire agreement). Paired with the new `idempotency-key-deterministic` row in `SECURITY/invariants.yaml`. 2026-05-17 (Phase 3 of RFC 0013) added three server-free scenarios exercising the reference workflow-chain expansion library (`conformance/src/lib/workflow-chain-expansion.ts`): `workflow-chain-expansion.test.ts` (parameter substitution + node id collision avoidance + edge rewriting + capability propagation + runtime-invariance contract), `workflow-chain-unresolvable-typeid.test.ts` (rejection with `chain_unresolvable_typeid` when a chain references an unknown typeId), and `workflow-chain-pack-signature-verification.test.ts` (Ed25519 verification recipe reuse from `node-packs.md §Signing`). Earlier that day (Phase 1) added `workflow-chain-pack-manifest-validation.test.ts` — server-free schema-validation scenario covering the new `workflow-chain-pack-manifest.schema.json` (positive sample + two negatives: kind/contents mismatch and invalid `chainId`). Closes RFC 0013 (`Workflow-chain packs`, `Draft`) Phases 1 + 3 alongside the new `spec/v1/workflow-chain-packs.md`, the `Capabilities.workflowChainPacks` block, and the registry build-index/conformance-check `kind` routing from Phase 2. Earlier that day, the suite added 27 `it.todo()` placeholder scenarios paired with RFCs 0014-0020 (host capability surfaces — fs, kvStorage, tableStorage, queueBus, sql/vector/search, blob/cache, mcp.serverMount). These promote to live assertions when each RFC reaches `Active` + the matching capability block lands in `schemas/capabilities.schema.json` + a reference host advertises the capability. Earlier additions include 18 Multi-Agent Shift scenarios (Phases 1-5) added 2026-05-10, the `registry-public.test.ts` public-registry healthcheck added 2026-05-11 (opt-in via `OPENWOP_TEST_PUBLIC_REGISTRY=true`), the `replay-llm-cache-key.test.ts` placeholder added 2026-05-11 (three `it.todo()` cases for the cross-host LLM cache-key recipe per `replay.md` §"LLM cache-key recipe"), the two `production-*.test.ts` scenarios added 2026-05-11 for the `openwop-production` profile per RFC 0009 (`production-backpressure.test.ts`, `production-retention-expiry.test.ts`), the four `auth-*.test.ts` scenarios added 2026-05-11/12 for the production-auth profiles per RFC 0010 (`auth-api-key-rotation.test.ts`, `auth-oauth2-client-credentials.test.ts`, `auth-oidc-user-bearer.test.ts`, `auth-mtls.test.ts` (opt-in via `OPENWOP_TEST_MTLS=1`)), `replay-retention-expiry.test.ts` added 2026-05-12 (capability shape + 410/422 envelope per `replay.md` §"Retention and garbage collection"), `bulk-cancel.test.ts` added 2026-05-12 (Phase B close-out of R1 — `POST /v1/runs:bulk-cancel`), the two Phase H launch-blocker advertisement-contract scenarios added 2026-05-12 (`mcp-toolcall-redaction.test.ts` for the MCP-1 invariant per `host-capabilities.md §host.mcp` + `threat-model-prompt-injection.md §UNTRUSTED`, and `http-client-ssrf.test.ts` for the SSRF + body-size cap advertisement contract on `capabilities.httpClient`), the `wasm-pack-abi-version-rejection.test.ts` Track 7 scenario added 2026-05-12 for the ABI-mismatch positive path via the `vendor.openwop.misbehaving-abi` pack per RFC 0008 §H, the `otel-trace-propagation-subworkflow.test.ts` Track 11 close-out added 2026-05-13 (parent + child run spans share the inbound traceparent's traceId across the `core.subWorkflow` dispatch boundary), and the three RFC 0012 (Memory Compaction Profile, `Active`) scenarios added 2026-05-13/14: `memory-compaction-sr1-carry-forward.test.ts` (load-bearing SR-1 §D), `memory-compaction-event-emitted.test.ts` (canonical §B payload shape), and `memory-compaction-provenance-tag.test.ts` (soft assertion on §C `compacted-from:<id>` convention). All three gate on `capabilities.memory.compaction.supported` + the host's test seam at `/v1/test/memory/{seed,compact}` (Postgres reference host enables both via `OPENWOP_MEMORY_COMPACTION=true OPENWOP_TEST_TRIGGER_COMPACTION=true`). 2026-05-15 (gap-closure CF-3) added `interrupt-token-matrix.test.ts` (malformed / unknown / replay / cross-run-id paths on `GET|POST /v1/interrupts/{token}`). The maintained scenario-to-spec map lives in [`coverage.md`](./coverage.md); this README keeps the operator quickstart and the historical scenario notes below.
 
 High-level coverage includes:
 
@@ -171,7 +171,7 @@ Server-required (added in 1.7.0):
 |---|---|---|
 | **Redaction** | [`capabilities.md`](../spec/v1/capabilities.md) §"Secrets" + NFR-7 + §"aiProviders" | Vendor-neutral assertions that the server doesn't leak secret material. Three scenario groups: (a) discovery shape contract — `secrets` + `aiProviders` advertisements are well-formed regardless of `secrets.supported`; when `supported === true`, scopes MUST be non-empty + `resolution === 'host-managed'`; `byok ⊆ supported`. (b) bearer-token redaction — invalid Bearer canary in `Authorization` header is not echoed in the 401 response body. (c) credentialRef echo control — gated on `secrets.supported === true`; canary planted in `configurable.ai.credentialRef` MUST NOT appear in any RunEvent payload (poll-based capture; transport-agnostic). Uses runtime-built canary fixtures (`lib/canaries.ts`) that defeat static secret scanners. 6 scenarios. |
 
-Current source tree: 157 scenario files. Use [`coverage.md`](./coverage.md) for current grade/gap tracking.
+Current source tree: 160 scenario files. Use [`coverage.md`](./coverage.md) for current grade/gap tracking.
 
 ## Remaining Gaps
 

package/api/redocly.yaml

@@ -6,3 +6,18 @@ rules:
   # and take no input — they have no meaningful 4XX response surface. 5XX is the
   # only honest failure mode and is documented inline.
   operation-4xx-response: off
+
+  # The rule flags `then: { required: [<name>] }` clauses where <name> is not
+  # defined in the `then` clause's OWN `properties` block — but in the standard
+  # JSON Schema 2020-12 `if/then` conditional-requirement idiom (used by
+  # `capabilities.schema.json` to tighten required-fields when `supported: true`),
+  # the property IS defined in the parent's `properties` block. Redocly's rule
+  # doesn't walk to parent scope, so it false-positives on every valid use of
+  # the idiom. Disabling globally is safe because:
+  #
+  #   - the typo-in-required protection the rule provides is also covered by
+  #     `scripts/check-required-properties-defined.mjs` (which walks parent
+  #     scope correctly), wired into `openwop-check.sh` step [7/9];
+  #   - our only `then/required` sites are the two in `capabilities.schema.json`,
+  #     both of which are well-formed.
+  no-required-schema-properties-undefined: off

package/coverage.md

@@ -34,6 +34,7 @@
 | Audit-log integrity profile | `audit-log-integrity.test.ts` | A | Profile claim + `/v1/audit/verify` shape + checkpoint-signature verification; chain re-walk with `chainValid` and `checkpointsValid` bits. Tamper detection covered host-internally at `examples/hosts/sqlite/test/audit-tamper.test.ts` (mutate-entry + forge-signature paths). CF-11 close-out 2026-05-15: cross-host checkpoint export via `examples/hosts/postgres/src/audit-export.ts` + standalone verifier at `scripts/verify-audit-checkpoints.mjs`; `examples/hosts/postgres/test/audit-checkpoint-export.test.ts` verifies 7 paths (positive + tampered-signature + non-monotonic-atSequence). |
 | Multi-region idempotency capability | `multi-region-idempotency.test.ts` | C | Discovery enum coverage; remaining: cross-region partition simulation (requires multi-region harness). |
 | Public hosted registry (`packs.openwop.dev`) | `registry-public.test.ts` | A− | Discovery, index, and per-pack manifest assertions against the public registry. Opt-in via `OPENWOP_TEST_PUBLIC_REGISTRY=true` so default conformance runs don't depend on outbound `packs.openwop.dev` reachability. Remaining: tarball-fetch + signature-verify roundtrip. |
+| Workflow-chain packs (RFC 0013 — `spec/v1/workflow-chain-packs.md`) | `workflow-chain-pack-manifest-validation.test.ts`, `workflow-chain-pack-signature-verification.test.ts`, `workflow-chain-expansion.test.ts`, `workflow-chain-unresolvable-typeid.test.ts`, `workflow-chain-host-expansion.test.ts` | A (4 server-free + 1 live-host) | RFC 0013 promoted Draft → Active 2026-05-18; live-host gate (`workflow-chain-host-expansion.test.ts`) added 2026-05-18 closing the Phase 3 acceptance criterion. The 4 server-free scenarios exercise the pure-library algorithm; the new live-host scenario exercises the reference in-memory host's HTTP wrapper (`POST /v1/host/sample/workflow-chain:expand`) under `OPENWOP_REQUIRE_BEHAVIOR=true`, gated on `capabilities.workflowChainPacks.supported: true`. 6 cases — discovery advertisement / 1-node positive / 2-node positive with edges / unknown-pack 404 / unknown-chain 404 / malformed-body 422. RFC promotes to `Accepted` once the live-host scenario passes against a published host run. |
 | AI Envelope (FINAL v1.1 — `spec/v1/ai-envelope.md`, RFC 0021) | `aiEnvelope.universalKinds.test.ts`, `aiEnvelope.schemaDrift.test.ts`, `aiEnvelope.correlationReplay.test.ts`, `aiEnvelope.contractRefusal.test.ts`, `aiEnvelope.trustBoundaryPropagation.test.ts`, `aiEnvelope.redaction.test.ts`, `aiEnvelope.capBreached.test.ts`, `ai-envelope-shape.test.ts` | B (shape + 8 behavioral accept-pipeline assertions) | DRAFT v1.x gap-closure landed 2026-05-17; RFC 0021 promotion to FINAL v1.1 landed 2026-05-18. Closes the long-standing gap where 8 v1 surfaces already referenced AI Envelopes (`Capabilities.supportedEnvelopes` + `schemaVersions` + the three per-turn limits; `host.aiEnvelope.generate`; `envelopeType` on workflow-chain packs; `profiles.md` derivation; `host-extensions.md` namespacing; `positioning.md`; reference host discovery) but the envelope's own wire shape, universal kinds, schema discipline, and Envelope Contract gate were never specified. The 7 `aiEnvelope.*` advertisement-shape probes (gated on `capabilities.envelopeContracts.advertised: true`) cover the host's CAPABILITY claim. `ai-envelope-shape.test.ts` adds 8 BEHAVIORAL assertions through the workflow-engine sample's env-gated `POST /v1/host/sample/envelope/accept` seam: accepted / invalid (shape) / invalid (ISO 8601 timestamp) / gated (vendor kind not advertised) / breached (counter at cap) / universal-kind-always-allowed / normalizedMeta.contentTrust propagation from runTrustBoundary / envelope-supplied contentTrust precedence over runTrustBoundary. Behavior-unlock dependency for the remaining `aiEnvelope.*` `it.todo()` blocks: a reference host wires schema drift + correlationId-replay short-circuit + redaction-carry-forward + contract-refusal mappings + trust-boundary propagation onto RunEventDocs. |
 
 ---
@@ -106,7 +107,7 @@ Every OpenAPI operation should have:
 | `resolveInterruptByRun` | `interrupt-approval.test.ts`, `interrupt-clarification.test.ts`, `approval-payload.test.ts`, `interruptRace.test.ts` | Invalid action, unknown node, race cases | Add auth-required and quorum profile scenarios. |
 | `inspectInterruptByToken` | `interrupt-token-matrix.test.ts` (CF-3, 2026-05-15) covers malformed + unknown token paths | Negative paths covered | Add explicit expired-token case when a host advertises a TTL seam. |
 | `resolveInterruptByToken` | `interrupt-token-matrix.test.ts` covers replay (already-resolved) + unknown token; `interrupt-external-event-correlation.test.ts` covers positive path | Replay path + unknown-token path covered with explicit assertions | Add wrong-action case once the host advertises a typed allowed-actions vocabulary in the interrupt manifest. |
-| `getArtifact` | Indirect through approval payload fixtures | `route-coverage.test.ts` covers unknown artifact `404`/`403` envelope; `artifact-auth.test.ts` (CF-4 close-out 2026-05-15) covers `401` unauthenticated path | Negative paths covered (401 + 404/403) | Add positive artifact-read scenario once a reference host implements `getArtifact` end-to-end. |
+| `getArtifact` | Indirect through approval payload fixtures | `route-coverage.test.ts` covers unknown artifact `404`/`403` envelope; `artifact-auth.test.ts` (CF-4 close-out 2026-05-15; SQLite host 401-before-404 stub landed 2026-05-19, closes the info-leak surface for every HTTP method) covers `401` unauthenticated path | Negative paths covered (401 + 405 non-GET + 404/403) | Add positive artifact-read scenario once a reference host implements `getArtifact` end-to-end. |
 | `registerWebhook` | Webhook spec exists | `route-coverage.test.ts` covers invalid URL validation envelope | Add positive registration with a test receiver when harness support exists. |
 | `unregisterWebhook` | Webhook spec exists | `route-coverage.test.ts` covers unknown subscription behavior | Add full register-then-unregister roundtrip with a test receiver. |
 

package/fixtures/conformance-agent-reasoning-streaming.json

@@ -0,0 +1,37 @@
+{
+  "id": "conformance-agent-reasoning-streaming",
+  "name": "Conformance: Streaming agent.reasoning.delta (RFC 0024)",
+  "version": "1.0",
+  "description": "RFC 0024. Drives the `core.conformance.mock-agent` node to emit incremental `agent.reasoning.delta` events (one per `mockReasoning.streamChunks` entry, sequence 0..N-1) followed by exactly one closing `agent.reasoned` whose `reasoning` equals the concatenation of all chunks. Gated on `capabilities.agents.reasoning.streaming: true`. See agentReasoningStreaming.test.ts.",
+  "nodes": [
+    {
+      "id": "streaming-reasoner",
+      "typeId": "core.conformance.mock-agent",
+      "name": "Streaming Reasoning Agent",
+      "position": { "x": 0, "y": 0 },
+      "config": {
+        "mockReasoning": {
+          "summary": "Streamed reasoning trace for RFC 0024 conformance.",
+          "tokenCount": 12,
+          "streamChunks": [
+            "Let me think about this. ",
+            "First, the user is asking a question. ",
+            "Therefore, I should respond clearly."
+          ]
+        }
+      },
+      "inputs": {},
+      "agent": {
+        "agentId": "core.conformance.streaming-reasoner",
+        "modelClass": "reasoning"
+      }
+    }
+  ],
+  "edges": [],
+  "triggers": [
+    { "id": "manual", "type": "manual", "enabled": true }
+  ],
+  "variables": [],
+  "metadata": { "tags": ["conformance", "multi-agent", "phase-1", "reasoning", "streaming", "rfc-0024"] },
+  "settings": { "timeout": 15000 }
+}

package/fixtures/conformance-dispatch-cancellable-child.json

@@ -0,0 +1,27 @@
+{
+  "id": "conformance-dispatch-cancellable-child",
+  "name": "Conformance: Dispatch cancellable child (RFC 0022 §B HVMAP-1b-cancelled)",
+  "version": "1.0",
+  "description": "RFC 0022 §B / HVMAP-1b-cancelled — child workflow with a long-running `core.delay` so the test can externally cancel it via `POST /v1/runs/{childRunId}/cancel`. Used by `conformance-dispatch-output-mapping` (parent) to verify that the parent's `outputMapping` MUST be SKIPPED when the child terminates `cancelled`; the parent's `parentResult` variable MUST remain at its pre-dispatch state.",
+  "nodes": [
+    {
+      "id": "wait",
+      "typeId": "core.delay",
+      "name": "Long delay (cancelled externally)",
+      "position": { "x": 0, "y": 0 },
+      "config": { "durationMs": 30000 },
+      "inputs": {}
+    }
+  ],
+  "edges": [],
+  "triggers": [
+    { "id": "manual", "type": "manual", "enabled": true }
+  ],
+  "variables": [
+    { "name": "childOutcome", "type": "string", "defaultValue": "this-should-not-be-harvested" }
+  ],
+  "metadata": {
+    "tags": ["conformance", "rfc-0022", "dispatch", "output-mapping", "hvmap-1b-cancelled", "child"]
+  },
+  "settings": { "timeout": 60000, "maxRetries": 0 }
+}

package/fixtures/conformance-dispatch-deterministic-fail-child.json

@@ -0,0 +1,30 @@
+{
+  "id": "conformance-dispatch-deterministic-fail-child",
+  "name": "Conformance: Dispatch deterministic-fail child (RFC 0022 §B HVMAP-1b-failed)",
+  "version": "1.0",
+  "description": "RFC 0022 §B / HVMAP-1b-failed — child workflow that ALWAYS terminates `failed` via the host's `core.fail` node. Used by `conformance-dispatch-output-mapping` (parent) to verify that the parent's `outputMapping` MUST be SKIPPED when the child terminates non-completed; the parent's `parentResult` variable MUST remain at its pre-dispatch state.",
+  "nodes": [
+    {
+      "id": "fail",
+      "typeId": "core.fail",
+      "name": "Deterministic fail",
+      "position": { "x": 0, "y": 0 },
+      "config": {
+        "code": "hvmap_1b_failed_fixture",
+        "message": "intentional terminal failure for outputMapping-skip assertion"
+      },
+      "inputs": {}
+    }
+  ],
+  "edges": [],
+  "triggers": [
+    { "id": "manual", "type": "manual", "enabled": true }
+  ],
+  "variables": [
+    { "name": "childOutcome", "type": "string", "defaultValue": "this-should-not-be-harvested" }
+  ],
+  "metadata": {
+    "tags": ["conformance", "rfc-0022", "dispatch", "output-mapping", "hvmap-1b-failed", "child"]
+  },
+  "settings": { "timeout": 5000, "maxRetries": 0 }
+}

package/fixtures/conformance-dispatch-input-mapping-no-default.json

@@ -0,0 +1,49 @@
+{
+  "id": "conformance-dispatch-input-mapping-no-default",
+  "name": "Conformance: Dispatch Input Mapping — parentName unset (RFC 0022 §A HVMAP-1a-null)",
+  "version": "1.0",
+  "description": "RFC 0022 §A / HVMAP-1a-null — parent variant of `conformance-dispatch-input-mapping` that DECLARES `parentName` in `variables[]` but OMITS its `defaultValue`. The dispatch's `inputMapping: { childGreeting: 'parentName' }` projects an unset parent variable. Per RFC 0022 §A normative bullet: child `inputs.childGreeting` MUST surface as `undefined` (NOT `null`, NOT omitted). Same supervisor + dispatch shape as the base fixture; only the `variables[].parentName.defaultValue` is removed. See dispatch-input-mapping.test.ts § HVMAP-1a-null.",
+  "nodes": [
+    {
+      "id": "supervisor",
+      "typeId": "core.orchestrator.supervisor",
+      "name": "Supervisor (mock plan)",
+      "position": { "x": 0, "y": 0 },
+      "config": {
+        "mockDispatchPlan": [
+          { "kind": "next-worker", "nextWorkerIds": ["conformance-dispatch-input-mapping-child"] },
+          { "kind": "terminate", "reason": "goal-reached" }
+        ]
+      },
+      "inputs": {}
+    },
+    {
+      "id": "dispatch",
+      "typeId": "core.dispatch",
+      "name": "Dispatch with inputMapping (parent var unset)",
+      "position": { "x": 200, "y": 0 },
+      "config": {
+        "askUserRouting": "auto",
+        "workerDispatchModel": "child-run",
+        "fanOutPolicy": "sequential",
+        "inputMapping": { "childGreeting": "parentName" }
+      },
+      "inputs": {}
+    }
+  ],
+  "edges": [
+    { "id": "e1", "sourceNodeId": "supervisor", "targetNodeId": "dispatch" },
+    { "id": "e2", "sourceNodeId": "dispatch", "targetNodeId": "supervisor" }
+  ],
+  "triggers": [
+    { "id": "manual", "type": "manual", "enabled": true }
+  ],
+  "variables": [
+    { "name": "parentName", "type": "string" }
+  ],
+  "metadata": {
+    "tags": ["conformance", "rfc-0022", "dispatch", "input-mapping", "hvmap-1a-null"],
+    "requiresCapability": "capabilities.agents.dispatchMapping"
+  },
+  "settings": { "timeout": 30000 }
+}

package/fixtures/conformance-dispatch-per-worker-override.json

@@ -0,0 +1,59 @@
+{
+  "id": "conformance-dispatch-per-worker-override",
+  "name": "Conformance: Dispatch Per-Worker Mapping Override (RFC 0022 §A HVMAP-1c-override)",
+  "version": "1.0",
+  "description": "RFC 0022 §A / HVMAP-1c-override — verifies that `perWorkerInputMappings[childId]` OVERRIDES the default `inputMapping` for the named child, while OTHER children continue to use the default. Supervisor plan: `next-worker: ['conformance-dispatch-cross-worker-handoff-child-a', 'conformance-dispatch-cross-worker-handoff-child-b']` → `terminate`. Default `inputMapping: { input: 'defaultX' }`; `perWorkerInputMappings.conformance-dispatch-cross-worker-handoff-child-b: { input: 'sharedVar' }`. child-a receives `input: <defaultX-value>`; child-b receives `input: <sharedVar-value>` (NOT defaultX) per RFC 0022 §A `effectiveInputMapping` precedence. See dispatch-cross-worker-handoff.test.ts § HVMAP-1c-override.",
+  "nodes": [
+    {
+      "id": "supervisor",
+      "typeId": "core.orchestrator.supervisor",
+      "name": "Supervisor (mock plan, multi-worker)",
+      "position": { "x": 0, "y": 0 },
+      "config": {
+        "mockDispatchPlan": [
+          {
+            "kind": "next-worker",
+            "nextWorkerIds": [
+              "conformance-dispatch-cross-worker-handoff-child-a",
+              "conformance-dispatch-cross-worker-handoff-child-b"
+            ]
+          },
+          { "kind": "terminate", "reason": "goal-reached" }
+        ]
+      },
+      "inputs": {}
+    },
+    {
+      "id": "dispatch",
+      "typeId": "core.dispatch",
+      "name": "Dispatch with default + per-worker override",
+      "position": { "x": 200, "y": 0 },
+      "config": {
+        "askUserRouting": "auto",
+        "workerDispatchModel": "child-run",
+        "fanOutPolicy": "sequential",
+        "inputMapping": { "input": "defaultX" },
+        "perWorkerInputMappings": {
+          "conformance-dispatch-cross-worker-handoff-child-b": { "input": "sharedVar" }
+        }
+      },
+      "inputs": {}
+    }
+  ],
+  "edges": [
+    { "id": "e1", "sourceNodeId": "supervisor", "targetNodeId": "dispatch" },
+    { "id": "e2", "sourceNodeId": "dispatch", "targetNodeId": "supervisor" }
+  ],
+  "triggers": [
+    { "id": "manual", "type": "manual", "enabled": true }
+  ],
+  "variables": [
+    { "name": "defaultX", "type": "string", "defaultValue": "default-x-value" },
+    { "name": "sharedVar", "type": "string", "defaultValue": "shared-value" }
+  ],
+  "metadata": {
+    "tags": ["conformance", "rfc-0022", "dispatch", "per-worker-mapping", "hvmap-1c-override"],
+    "requiresCapability": "capabilities.agents.dispatchMapping"
+  },
+  "settings": { "timeout": 30000 }
+}

package/fixtures/conformance-subworkflow-input-mapping-no-default.json

@@ -0,0 +1,33 @@
+{
+  "id": "conformance-subworkflow-input-mapping-no-default",
+  "name": "Conformance: subWorkflow Input Mapping — currentPrdId unset (RFC 0022 §B HVMAP-2-unset)",
+  "version": "1.0",
+  "description": "RFC 0022 §B / HVMAP-2-unset — parent variant of `conformance-subworkflow-input-mapping` that DECLARES `currentPrdId` in `variables[]` but OMITS its `defaultValue`. The `core.subWorkflow` node's `inputMapping: { receivedPrdId: 'currentPrdId' }` projects an unset parent variable. Per RFC 0022 §B normative bullet: child `receivedPrdId` MUST surface as `undefined` (NOT `null`, NOT omitted) — distinct from the child's own `defaultValue` fold. Same shape as the base fixture; only `currentPrdId.defaultValue` is removed. See subworkflow-input-mapping.test.ts § HVMAP-2-unset.",
+  "nodes": [
+    {
+      "id": "subwf-call",
+      "typeId": "core.subWorkflow",
+      "name": "Invoke child with inputMapping (parent var unset)",
+      "position": { "x": 0, "y": 0 },
+      "config": {
+        "workflowId": "conformance-subworkflow-input-mapping-child",
+        "waitForCompletion": true,
+        "onChildFailure": "fail-parent",
+        "inputMapping": { "receivedPrdId": "currentPrdId" }
+      },
+      "inputs": {}
+    }
+  ],
+  "edges": [],
+  "triggers": [
+    { "id": "manual", "type": "manual", "enabled": true }
+  ],
+  "variables": [
+    { "name": "currentPrdId", "type": "string" }
+  ],
+  "metadata": {
+    "tags": ["conformance", "rfc-0022", "subworkflow", "input-mapping", "hvmap-2-unset"],
+    "requiresCapability": "capabilities.subWorkflow.inputMapping"
+  },
+  "settings": { "timeout": 10000 }
+}

package/fixtures.md

@@ -59,6 +59,7 @@ All fixtures MUST advertise:
 | Interrupt — Parent/Child Cancel | `conformance-interrupt-parent-child-cancel` + `conformance-interrupt-parent-child-cancel-child` | Verifies `openwop-interrupt-parent-child` cancel cascade | `cancelled` (both runs) | ≤ 30s |
 | Agent Identity | `conformance-agent-identity` | Phase 1 — `RunSnapshot.agent` / `runOrchestrator` AgentRef wire-shape | `completed` | ≤ 10s |
 | Agent Reasoning | `conformance-agent-reasoning` | Phase 1 / RFC 0023 — `agent.*` event family emission + `callId` pairing on `core.conformance.mock-agent` | `completed` | ≤ 15s |
+| Agent Reasoning Streaming | `conformance-agent-reasoning-streaming` | RFC 0024 — `core.conformance.mock-agent` with `mockReasoning.streamChunks` drives incremental `agent.reasoning.delta` events (sequence 0..N-1) followed by exactly one closing `agent.reasoned` whose `reasoning` equals the concatenation. Gated on `capabilities.agents.reasoning.streaming: true`. | `completed` | ≤ 15s |
 | Agent Low-Confidence | `conformance-agent-low-confidence` | Phase 1 / CP-1 / RFC 0023 — `core.conformance.mock-agent` emits `agent.decided` with confidence < threshold; host MUST follow with `node.suspended { reason: 'low-confidence' }` | `waiting-approval` (suspends) | unbounded (suspends) |
 | Message Reducer | `conformance-message-reducer` | Phase 1 — `message` reducer idempotency on duplicate `messageId` | `completed` | ≤ 10s |
 | Agent Pack Install | `conformance-agent-pack-install` | Phase 2 — pack `agents[]` surface as AgentManifest at `GET /v1/packs` | `completed` | ≤ 5s |
@@ -74,6 +75,11 @@ All fixtures MUST advertise:
 | Dispatch Output Mapping (child) | `conformance-dispatch-output-mapping-child` | RFC 0022 §A / HVMAP-1b — child workflow for the dispatch output-mapping scenario. Declares `childOutcome.defaultValue='done'`; on terminal, parent's `outputMapping` harvests `childOutcome → parentResult`. | `completed` | ≤ 5s |
 | Dispatch Cross-Worker Handoff (child-a) | `conformance-dispatch-cross-worker-handoff-child-a` | RFC 0022 §A / HVMAP-1c — first child of the cross-worker-handoff scenario. Declares `output.defaultValue='hello'`; on terminal, parent's `perWorkerOutputMappings.child-a` harvests `output → sharedVar`. | `completed` | ≤ 5s |
 | Dispatch Cross-Worker Handoff (child-b) | `conformance-dispatch-cross-worker-handoff-child-b` | RFC 0022 §A + §D / HVMAP-1c — second child of the cross-worker-handoff scenario. Sequential fan-out — runs after child-a; receives parent's `sharedVar` via `perWorkerInputMappings.child-b` onto its `input` input. Scenario reads child-b's `inputs_json` to assert `inputs.input === 'hello'`. | `completed` | ≤ 5s |
+| Dispatch Input Mapping — unset variant | `conformance-dispatch-input-mapping-no-default` | RFC 0022 §A / HVMAP-1a-null — parent variant that DECLARES `parentName` but OMITS its `defaultValue`. The dispatch's `inputMapping` projects an unset parent variable; per §A normative bullet, child `inputs.childGreeting` MUST surface as `undefined` (NOT `null`, NOT omitted). Reuses `conformance-dispatch-input-mapping-child`. | `completed` | ≤ 30s |
+| subWorkflow Input Mapping — unset variant | `conformance-subworkflow-input-mapping-no-default` | RFC 0022 §B / HVMAP-2-unset — parent variant that DECLARES `currentPrdId` but OMITS its `defaultValue`. Per §B, the unset projection MUST surface as `undefined` (NOT `null`) — distinct from the child's own `defaultValue` fold. Reuses `conformance-subworkflow-input-mapping-child`. | `completed` | ≤ 10s |
+| Dispatch Per-Worker Mapping Override | `conformance-dispatch-per-worker-override` | RFC 0022 §A / HVMAP-1c-override — parent with BOTH a default `inputMapping` (`{ input: 'defaultX' }`) AND `perWorkerInputMappings.child-b: { input: 'sharedVar' }`. Verifies `effectiveInputMapping` precedence per §A: child-a receives the default, child-b receives the override. Reuses `conformance-dispatch-cross-worker-handoff-child-a` + `-child-b`. | `completed` | ≤ 30s |
+| Dispatch deterministic-fail child | `conformance-dispatch-deterministic-fail-child` | RFC 0022 §B / HVMAP-1b-failed — child workflow that ALWAYS terminates `failed` via `core.fail`. Used by `conformance-dispatch-output-mapping` to verify the parent's `outputMapping` is SKIPPED when the child fails terminally. | `failed` | ≤ 5s |
+| Dispatch cancellable child | `conformance-dispatch-cancellable-child` | RFC 0022 §B / HVMAP-1b-cancelled — child workflow with a long `core.delay` so the test cancels it externally via `POST /v1/runs/{childRunId}/cancel`. Verifies the parent's `outputMapping` is SKIPPED when the child terminates `cancelled`. | `cancelled` | ≤ 60s |
 | Agent Memory Round-Trip | `conformance-agent-memory-roundtrip` | Phase 3 — `MemoryAdapter.list/get` write → read | `completed` | ≤ 15s |
 | Agent Memory Cross-Tenant | `conformance-agent-memory-cross-tenant` | Phase 3 / CTI-1 — cross-tenant probe MUST return `[]` / `null` | `completed` | ≤ 10s |
 | Agent Memory Redaction | `conformance-agent-memory-redaction` | Phase 3 / SR-1 — BYOK plaintext surfaces as `[REDACTED:<id>]` on read | `completed` | ≤ 15s |

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@openwop/openwop-conformance",
-  "version": "1.2.0",
+  "version": "1.3.0",
   "description": "Production-ready black-box conformance suite for OpenWOP v1.0 compliant servers.",
   "repository": {
     "type": "git",

package/schemas/capabilities.schema.json

@@ -210,6 +210,17 @@
       "uniqueItems": true,
       "description": "Optional v1 host-advertised opaque capability ids that NodeModules may declare in `NodeModule.requires`. Naming convention: dotted, domain-scoped (`chat.sendPrompt`, `canvas.write`, `secrets.byok`). Provider value shapes are documented per-capability alongside consumers, NOT in the protocol package — the protocol owns the *check*, not domain provider contracts. A client that submits a workflow whose nodes declare a `requires` entry SHOULD first verify the host advertises that capability; a host that lacks a capability MUST refuse to dispatch nodes that declare it, terminating the run with `RunSnapshot.error.code = 'capability_not_provided'`. See `capabilities.md` §\"Runtime capabilities\"."
     },
+    "providerUsage": {
+      "type": "object",
+      "description": "RFC 0026 (Draft). Hosts that emit `provider.usage` events after every LLM provider invocation per RFC 0026 §B. The event carries per-call token counts in the durable event log; cost rollup remains in `RunSnapshot.metrics.openwopCost`. Old hosts ignore.",
+      "properties": {
+        "supported": { "type": "boolean", "description": "Host emits one `provider.usage` event per LLM provider call." },
+        "costEstimates": { "type": "boolean", "description": "When true, the host includes `costEstimateUsd` on `provider.usage` events using its internal rate table. When false/absent, only token counts are emitted." },
+        "currency": { "type": "string", "pattern": "^[A-Z]{3}$", "description": "Default ISO 4217 currency for `costEstimateUsd` values. When absent, USD is assumed." }
+      },
+      "required": ["supported"],
+      "additionalProperties": false
+    },
     "aiProviders": {
       "type": "object",
       "description": "Optional v1 companion to `secrets`. Advertises which AI providers the host's AI-proxy can route to and which permit BYOK.",
@@ -357,6 +368,11 @@
               "type": "integer",
               "minimum": 0,
               "description": "Effective cap on reasoning trace length when verbosity is `summary`. Default 512 tokens."
+            },
+            "streaming": {
+              "type": "boolean",
+              "default": false,
+              "description": "RFC 0024. When `true`, the host MAY emit `agent.reasoning.delta` events while a reasoning block is still open, in addition to the final `agent.reasoned`. Hosts that omit / `false` this flag emit only the final `agent.reasoned`. Consumers MUST tolerate both modes."
             }
           },
           "additionalProperties": false

package/schemas/core-conformance-mock-agent-config.schema.json

@@ -31,6 +31,11 @@
               "type": "integer",
               "minimum": 0,
               "description": "Optional reasoning-token count for observability/cost attribution."
+            },
+            "streamChunks": {
+              "type": "array",
+              "items": { "type": "string" },
+              "description": "RFC 0024. When present, the host MUST emit one `agent.reasoning.delta` event per array entry (in order; `sequence` starts at 0 and increments by 1), THEN emit the closing `agent.reasoned` event whose `reasoning` field equals the concatenation of all `streamChunks`. When absent, only the closing `agent.reasoned` event fires (the pre-RFC-0024 behavior). Gated at conformance test time on `capabilities.agents.reasoning.streaming: true`."
             }
           }
         }

package/schemas/run-event-payloads.schema.json

@@ -2,7 +2,7 @@
   "$schema": "https://json-schema.org/draft/2020-12/schema",
   "$id": "https://openwop.dev/spec/v1/run-event-payloads.schema.json",
   "title": "RunEventPayloads",
-  "description": "Per-RunEventType payload schemas. The base RunEventDoc shape (run-event.schema.json) leaves `payload` permissive for forward-compat. This schema defines the canonical payload contract for each known RunEventType. Consumers MAY pin strict payload validation via `$defs.<typeId>` and `ajv.validate(schema.$defs[event.type], event.payload)`. Unknown event types MUST be tolerated (no $defs match → fold best-effort).\n\n48 variants from `run-event.schema.json#$defs.RunEventType` are covered, grouped into ~20 shape families with shared $defs. Naming convention: camelCase keys mirror dotted RunEventType names (e.g., `run.started` → `runStarted`).",
+  "description": "Per-RunEventType payload schemas. The base RunEventDoc shape (run-event.schema.json) leaves `payload` permissive for forward-compat. This schema defines the canonical payload contract for each known RunEventType. Consumers MAY pin strict payload validation via `$defs.<typeId>` and `ajv.validate(schema.$defs[event.type], event.payload)`. Unknown event types MUST be tolerated (no $defs match → fold best-effort).\n\n50 variants from `run-event.schema.json#$defs.RunEventType` are covered, grouped into ~20 shape families with shared $defs. Naming convention: camelCase keys mirror dotted RunEventType names (e.g., `run.started` → `runStarted`).",
   "type": "object",
   "$defs": {
     "_typeIndex": {
@@ -49,6 +49,8 @@
         "lease.handed-off":          { "$ref": "#/$defs/leaseHandedOff" },
         "replay.diverged":           { "$ref": "#/$defs/replayDiverged" },
         "agent.reasoned":            { "$ref": "#/$defs/agentReasoned" },
+        "agent.reasoning.delta":     { "$ref": "#/$defs/agentReasoningDelta" },
+        "provider.usage":            { "$ref": "#/$defs/providerUsage" },
         "agent.toolCalled":          { "$ref": "#/$defs/agentToolCalled" },
         "agent.toolReturned":        { "$ref": "#/$defs/agentToolReturned" },
         "agent.handoff":             { "$ref": "#/$defs/agentHandoff" },
@@ -564,6 +566,38 @@
       "additionalProperties": true
     },
 
+    "agentReasoningDelta": {
+      "type": "object",
+      "description": "RFC 0024. Incremental reasoning chunk for live-streaming UX. Emitted while a reasoning block is still open, BEFORE the corresponding `agent.reasoned` finalization. Consumers concatenate `delta` strings in arrival order to reconstruct the in-progress trace; the closing `agent.reasoned` event carries the FULL authoritative `reasoning`. Gated on `capabilities.agents.reasoning.streaming: true`. NOTE: `additionalProperties: true` mirrors the Phase-1 multi-agent-shift carve-out applied to the sibling `agentReasoned` schema — a deliberate forward-compat exception per RFC 0024 §Compatibility, not a precedent generalizable to other event payloads.",
+      "required": ["agentId", "delta", "sequence"],
+      "properties": {
+        "agentId":   { "type": "string", "minLength": 3, "maxLength": 256, "description": "AgentRef.agentId of the reasoning agent. MUST match the eventual closing `agent.reasoned`." },
+        "delta":     { "type": "string", "description": "New reasoning content since the previous delta event in this block (or since block open, if `sequence` is 0)." },
+        "sequence":  { "type": "integer", "minimum": 0, "description": "Monotonically-increasing index within the current reasoning block. Starts at 0 for the first delta in a block; resets at each new block open. Consumers MAY use this to detect dropped events." },
+        "verbosity": { "type": "string", "enum": ["summary", "full", "off"], "description": "Verbosity mode the host resolved for this block. SHOULD match the verbosity reported on the closing `agent.reasoned`." }
+      },
+      "additionalProperties": true
+    },
+
+    "providerUsage": {
+      "type": "object",
+      "description": "RFC 0026. Per-call usage record emitted after every LLM provider invocation. Durably persisted in the run event log; consumed by replay, webhook subscribers, billing reconciliation. The OTel `openwop.cost.*` attribute group (per `observability.md §\"Cost attribution attributes\"`) is the observability sibling — this event type is the durable record. Replay determinism: `inputTokens` + `outputTokens` MUST replay identically; `costEstimateUsd` MAY be omitted on replay. The payload MUST NOT carry credentialRefs, hashed credential identifiers, or prompt/response substrings per `SECURITY/threat-model-secret-leakage.md §SR-1` (enforced by SECURITY invariant `provider-usage-no-credential-leak`).",
+      "required": ["provider", "model", "inputTokens", "outputTokens"],
+      "properties": {
+        "provider":        { "type": "string", "minLength": 1, "description": "Canonical provider id (lowercase ASCII, e.g. \"anthropic\", \"openai\", \"google\"). Same value as the `openwop.cost.provider` OTel attribute." },
+        "model":           { "type": "string", "minLength": 1, "description": "Provider-stamped model id as the model expects it. Same value used in the LLM cache-key recipe per `replay.md §A`." },
+        "inputTokens":     { "type": "integer", "minimum": 0, "description": "Input/prompt tokens billed for this call. Matches the provider response's input-token count verbatim." },
+        "outputTokens":    { "type": "integer", "minimum": 0, "description": "Output/completion tokens billed for this call. Matches the provider response's output-token count verbatim." },
+        "totalTokens":     { "type": "integer", "minimum": 0, "description": "Convenience sum (inputTokens + outputTokens). Consumers MAY compute themselves; emitters MAY include for readability." },
+        "costEstimateUsd": { "type": "number", "minimum": 0, "description": "ADVISORY estimate in USD computed by the host's static rate table. MUST NOT be used for billing — real billing is external. Hosts SHOULD omit when no rate is known rather than emit 0." },
+        "currency":        { "type": "string", "pattern": "^[A-Z]{3}$", "description": "ISO 4217 code when `costEstimateUsd` is non-USD; the field name stays `costEstimateUsd` for back-compat but `currency` overrides the implied denomination." },
+        "cacheHit":        { "type": "boolean", "description": "True iff this call was served from the LLM response cache per `replay.md §\"LLM cache-key recipe\"`. When true, inputTokens/outputTokens reflect the ORIGINAL call's billed values; the cached invocation incurred zero new provider cost." },
+        "nodeId":          { "type": "string", "description": "The node id that initiated the provider call. Required for per-node cost attribution dashboards." },
+        "traceId":         { "type": "string", "description": "OTel trace id linking this event to the matching `openwop.cost.*` span. Lets observability backends correlate event-log entries with traces." }
+      },
+      "additionalProperties": false
+    },
+
     "agentToolCalled": {
       "type": "object",
       "description": "Multi-Agent Shift Phase 1. Emitted when an agent invokes a tool. Pairs with `agent.toolReturned` via shared `callId`.",

package/schemas/run-event.schema.json

@@ -102,6 +102,8 @@
         "lease.handed-off",
         "replay.diverged",
         "agent.reasoned",
+        "agent.reasoning.delta",
+        "provider.usage",
         "agent.toolCalled",
         "agent.toolReturned",
         "agent.handoff",

package/src/lib/driver.ts

@@ -78,6 +78,21 @@ class OpenWOPDriver {
     return this.request('POST', path, { ...init, body });
   }
 
+  /** PUT helper. The body is JSON-stringified by default; pass a string
+   *  Content-Type header for raw-body PUTs (e.g. tarball uploads).
+   *  Production hosts that accept tarball PUTs on /v1/packs/* expect
+   *  `Content-Type: application/octet-stream`; callers MUST set the
+   *  header explicitly when uploading non-JSON. */
+  put(path: string, body: unknown, init: OpenWOPRequestInit = {}): Promise<OpenWOPResponse> {
+    return this.request('PUT', path, { ...init, body });
+  }
+
+  /** DELETE alias for the canonical name. Keeps the call-site shorter
+   *  for scenarios that delete via `driver.del(...)`. */
+  del(path: string, init: OpenWOPRequestInit = {}): Promise<OpenWOPResponse> {
+    return this.request('DELETE', path, init);
+  }
+
   delete(path: string, init: OpenWOPRequestInit = {}): Promise<OpenWOPResponse> {
     return this.request('DELETE', path, init);
   }