docs(ri-050): forge fail-closed docs + TASKS status catch-up (#1275)

Rebased onto next post-#1278. README + skills/mosaic-forge/SKILL.md
document the FORGE_NO_EXECUTOR fail-closed rule + --simulate (fast-follow
to RI-2-001, now merged). TASKS.md rows updated to measured state:
RI-0-001/RI-1-001/RI-1-002/RI-2-001 done (merged with reviews+CI evidence);
RI-2-002/RI-3-001/RI-4-001/RI-5-001 in-progress (PRs open, approved,
merge-queued); RI-3-002/RI-V-001 not-started.

.woodpecker/ci-image.yml

@@ -22,9 +22,9 @@ steps:
     image: gcr.io/kaniko-project/executor:debug
     environment:
       REGISTRY_USER:
-        from_secret: gitea_username
+        from_secret: REGISTRY_USERNAME
       REGISTRY_PASS:
-        from_secret: gitea_password
+        from_secret: REGISTRY_PASSWORD
       CI_COMMIT_BRANCH: ${CI_COMMIT_BRANCH}
       CI_COMMIT_TAG: ${CI_COMMIT_TAG}
       CI_COMMIT_SHA: ${CI_COMMIT_SHA}

.woodpecker/ci.yml

@@ -30,6 +30,19 @@ steps:
       # the baked pnpm store.
       - pnpm install --frozen-lockfile --prefer-offline
 
+  # ---------------------------------------------------------------------------
+  # The steps below (sanitization, upgrade-guard, typecheck, lint, format,
+  # test) are the COMPLETE mandatory verification set. SDLC-D-034 mirrors them
+  # one-for-one in the canonical terminal verification command — root
+  # `pnpm verify:release` (scripts/verify-release.mjs) — which the publish
+  # pipeline (.woodpecker/publish.yml `verify` step) runs before ANY publish
+  # effect. These lines stay direct (not routed through the runner) because the
+  # #1017 test-enumeration guard audits framework tool paths through THIS
+  # surface; scripts/verify-release.test.mjs enforces that the runner's stage
+  # table keeps matching these commands exactly, so the two cannot drift.
+  # ---------------------------------------------------------------------------
+
+  # Canonical verify:release stage `sanitization`.
   # Blocking gate: public framework package must contain no operator-specific
   # personal data or private $HOME defaults. Runs early (no node_modules needed).
   sanitization:
@@ -47,6 +60,7 @@ steps:
       # with everything it guards; this direct line keeps one instrument running.
       - bash packages/mosaic/framework/tools/quality/scripts/check-test-enumeration.sh
 
+  # Canonical verify:release stage `upgrade-guard`.
   # Blocking gate (#791): a framework upgrade must never write or delete an
   # operator-owned path. The HARD GATE proves an unanticipated operator sentinel
   # survives a keep-mode reseed byte-identical (with rsync present AND absent —
@@ -68,6 +82,8 @@ steps:
       - bash packages/mosaic/framework/tools/quality/scripts/test-upgrade-durable-snapshot.sh
       - bash packages/mosaic/framework/tools/quality/scripts/test-install-migration.sh
 
+  # Canonical verify:release stage `typecheck` — the same `pnpm typecheck`
+  # invocation (which runs the checkout preflight first, then turbo).
   typecheck:
     image: *node_image
     commands:
@@ -78,7 +94,8 @@ steps:
       - sanitization
       - upgrade-guard
 
-  # lint, format, and test are independent — run in parallel after typecheck
+  # lint, format, and test are independent — run in parallel after typecheck.
+  # Each runs exactly its canonical verify:release stage command.
   lint:
     image: *node_image
     commands:
@@ -95,6 +112,12 @@ steps:
     depends_on:
       - typecheck
 
+  # Canonical verify:release stage `test` — the `pnpm test` line below is the
+  # shared command; everything else in this step is PIPELINE-LEVEL
+  # prerequisite the canonical command expects its caller to provide (SDLC-D-034):
+  # the ci-postgres service + pg_isready wait + db:migrate (postgres path),
+  # `apk add openssl`, and the pinned pi install. None of those can move into
+  # the runner (it must also work locally on the PGlite path with no database).
   test:
     image: *node_image
     environment:

.woodpecker/publish.yml

@@ -1,5 +1,19 @@
 # Build, publish npm packages, and push Docker images
 # Runs on main for stable publishes and on next for integration-line prereleases/images
+#
+# SDLC-D-034 publish gate: every publish effect (publish-npm, publish-next-npm,
+# and every image build/push step) depends DIRECTLY on the `verify` step below.
+# `verify` (a) asserts the provider's commit identity matches the actual
+# checkout (CI_COMMIT_SHA == git rev-parse HEAD, fail closed on mismatch or
+# emptiness) and (b) runs the canonical terminal verification command
+# (`pnpm verify:release`), which mirrors the PR CI pipeline's complete
+# mandatory set (sanitization, upgrade-guard, preflight+typecheck, lint,
+# format:check, test, build) — see scripts/verify-release.mjs. A missing,
+# failed, skipped, cancelled, or inconclusive verification therefore skips the
+# dependent publish effects (fail closed). Path-filtered short-circuits may
+# skip publish EFFECTS (e.g. docs-only merges) but never bypass `verify` for a
+# publish that does run: `verify` itself carries no path filter.
+# scripts/verify-release.test.mjs enforces this DAG invariant at checkout time.
 
 variables:
   # Pre-baked CI base (see .woodpecker/ci-image.yml): node:24-alpine +
@@ -48,6 +62,45 @@ steps:
       # Resolve from the baked pnpm store instead of a cold network fetch.
       - pnpm install --frozen-lockfile --prefer-offline
 
+  # SDLC-D-034 exact-commit publish gate. No `when`/path filter on purpose: it
+  # runs for every event this pipeline serves so no publish effect can ever
+  # start without it. Fails closed on commit-identity mismatch (or either SHA
+  # being empty) and on any incomplete verification.
+  verify:
+    image: *node_image
+    commands:
+      - *enable_pnpm
+      # (a) Commit identity: the provider's claimed SHA must equal the actual
+      # checkout HEAD — verification of anything else must never authorize a
+      # publish of this commit.
+      - |
+        if [ -z "$CI_COMMIT_SHA" ]; then
+          echo "[verify] FATAL: CI_COMMIT_SHA is empty — cannot certify commit identity" >&2
+          exit 1
+        fi
+        CHECKOUT_SHA="$(git rev-parse HEAD 2>/dev/null || true)"
+        if [ -z "$CHECKOUT_SHA" ]; then
+          echo "[verify] FATAL: git rev-parse HEAD returned nothing — cannot certify commit identity" >&2
+          exit 1
+        fi
+        if [ "$CI_COMMIT_SHA" != "$CHECKOUT_SHA" ]; then
+          echo "[verify] FATAL: provider commit ($CI_COMMIT_SHA) != checkout HEAD ($CHECKOUT_SHA)" >&2
+          exit 1
+        fi
+        echo "[verify] commit identity confirmed: $CHECKOUT_SHA"
+      # (b) Canonical terminal verification. Caller-provided prerequisites the
+      # runner expects (see .woodpecker/ci.yml comments): bash/rsync for the
+      # guard stages, openssl + the pinned pi binary for the test stage. git is
+      # baked into ci-base but re-asserted here so the identity check above can
+      # never silently depend on a stale baked image. DATABASE_URL is
+      # deliberately NOT set: the canonical command must hold on the PGlite
+      # path too and never sets or requires a database itself.
+      - apk add --no-cache bash rsync openssl git
+      - npm install -g @earendil-works/[email protected]
+      - pnpm verify:release
+    depends_on:
+      - install
+
   build:
     image: *node_image
     commands:
@@ -55,6 +108,7 @@ steps:
       - pnpm build
     depends_on:
       - install
+      - verify
 
   publish-npm:
     image: *node_image
@@ -114,6 +168,7 @@ steps:
         exit 1
     depends_on:
       - build
+      - verify
 
   publish-next-npm:
     image: *node_image
@@ -192,6 +247,7 @@ steps:
         echo "[publish-next] @mosaicstack/mosaic@next resolves to $RESOLVED_VERSION"
     depends_on:
       - build
+      - verify
 
   # TODO: Uncomment when ready to publish to npmjs.org
   # publish-npmjs:
@@ -205,6 +261,7 @@ steps:
   #     - bash scripts/publish-npmjs.sh
   #   depends_on:
   #     - build
+  #     - verify
   #   when:
   #     - event: [tag]
 
@@ -213,9 +270,9 @@ steps:
     when: *image_build_when
     environment:
       REGISTRY_USER:
-        from_secret: gitea_username
+        from_secret: REGISTRY_USERNAME
       REGISTRY_PASS:
-        from_secret: gitea_password
+        from_secret: REGISTRY_PASSWORD
       CI_COMMIT_BRANCH: ${CI_COMMIT_BRANCH}
       CI_COMMIT_TAG: ${CI_COMMIT_TAG}
       CI_COMMIT_SHA: ${CI_COMMIT_SHA}
@@ -242,15 +299,16 @@ steps:
         /kaniko/executor --context . --dockerfile docker/gateway.Dockerfile $DESTINATIONS
     depends_on:
       - build
+      - verify
 
   build-appservice:
     image: gcr.io/kaniko-project/executor:debug
     when: *main_image_build_when
     environment:
       REGISTRY_USER:
-        from_secret: gitea_username
+        from_secret: REGISTRY_USERNAME
       REGISTRY_PASS:
-        from_secret: gitea_password
+        from_secret: REGISTRY_PASSWORD
       CI_COMMIT_BRANCH: ${CI_COMMIT_BRANCH}
       CI_COMMIT_TAG: ${CI_COMMIT_TAG}
       CI_COMMIT_SHA: ${CI_COMMIT_SHA}
@@ -268,15 +326,16 @@ steps:
         /kaniko/executor --context . --dockerfile docker/appservice.Dockerfile $DESTINATIONS
     depends_on:
       - build
+      - verify
 
   build-web:
     image: gcr.io/kaniko-project/executor:debug
     when: *main_image_build_when
     environment:
       REGISTRY_USER:
-        from_secret: gitea_username
+        from_secret: REGISTRY_USERNAME
       REGISTRY_PASS:
-        from_secret: gitea_password
+        from_secret: REGISTRY_PASSWORD
       CI_COMMIT_BRANCH: ${CI_COMMIT_BRANCH}
       CI_COMMIT_TAG: ${CI_COMMIT_TAG}
       CI_COMMIT_SHA: ${CI_COMMIT_SHA}
@@ -294,3 +353,4 @@ steps:
         /kaniko/executor --context . --dockerfile docker/web.Dockerfile $DESTINATIONS
     depends_on:
       - build
+      - verify

README.md

@@ -138,9 +138,9 @@ mosaic brain tasks
 mosaic brain conversations
 
 # Agent forge pipeline
-mosaic forge run
+mosaic forge run [--simulate]        # fails closed (FORGE_NO_EXECUTOR) with no executor wired; --simulate for typed simulated runs
 mosaic forge status
-mosaic forge resume
+mosaic forge resume [--simulate]     # same fail-closed rule as forge run
 mosaic forge personas
 
 # Structured logging

apps/gateway/src/__tests__/cross-user-isolation.test.ts

@@ -190,7 +190,13 @@ beforeEach((ctx) => {
 });
 
 afterAll(async () => {
-  if (!handle) return;
+  // Cleanup only when the fixture actually installed rows. `handle` is set
+  // before the first query (createDb connects lazily), so on an unreachable
+  // database `handle` is truthy while nothing was inserted — cleanup must
+  // honor `dbAvailable` or the skip path fails the file with ECONNREFUSED in
+  // afterAll (caught live by the publish pipeline's no-DATABASE_URL verify
+  // step, pipeline 2486).
+  if (!handle || !dbAvailable) return;
   const db = handle.db;
 
   // Delete in dependency order (FK constraints)

docs/PRD.md

@@ -1368,3 +1368,38 @@ All work is **alpha** (< 0.1.0) until Jason approves 0.1.0 beta release.
 10. ASSUMPTION: **Conversations and messages get their own PG tables** (not stored in brain's entity model). They follow a chat-specific schema with proper foreign keys to users and projects. Rationale: Chat has different access patterns (streaming, pagination, search) than brain entities.
 
 11. RESOLVED: **Pi handles all target LLM providers natively.** Anthropic, OpenAI/Codex, Z.ai, Ollama, LM Studio, and llama.cpp are all supported via Pi's built-in providers or `models.json` configuration with `openai-completions` API type. No custom provider adapters needed in @mosaicstack/agent — only configuration management.
+
+---
+
+## Release Integrity Workstream (RI, #1275)
+
+### Problem and objective
+
+At `next` 476db12b (review of 2026-08-17), publication from `next` is not bound to the full verification pipeline for the same commit: the publish pipeline's publish steps depend on `build` only, while ordinary push CI excludes `next`. Public Forge/MACP paths contain false-success placeholders: a stub executor that reports `completed` with exit zero, planning/remediation gates that execute literal `true`, a review gate that echoes an approving verdict, and a gate runner that treats empty commands and unimplemented CI-provider gates as passing. Shipping UI surfaces can render a failed fetch as an empty, healthy collection.
+
+Objective: for alpha 0.0.50, the release cannot publish, report, or display work state that the repository has not actually verified. Decisions SDLC-D-033 through SDLC-D-038 (Jason, 2026-08-17) scope this floor; full decision text and required-behavior lists live in jarvis-brain `docs/plans/2026-08-16_mosaic-stack-sdlc-protocol.md` and `data/decisions/mosaic-stack-sdlc-protocol.json`. This section restates only the normative requirements.
+
+### Normative requirements
+
+1. **RI-N1 Exact-commit publication verification (SDLC-D-034).** One canonical terminal verification command performs self-contained re-verification in the publish pipeline against the job's checked-out commit before any external publication effect. The command contains or invokes the complete mandatory verification set (semantic parity with the PR merge gate, including sanitization, upgrade-guard, typecheck, lint, format check, tests, and build); CI and publication do not maintain separate semantic checklists. Every publish step depends on the verification step in the executable pipeline DAG. Provider commit identity and `git rev-parse HEAD` must identify the same commit. Missing, skipped, cancelled, stale, or inconclusive checks fail closed. Documentation-only runs may skip publication but cannot bypass verification when a publication effect will occur. A negative control must prove that a broken check blocks every publish step.
+
+2. **RI-N2 Fail-closed Forge/MACP with explicit simulation (SDLC-D-035).** Simulation requires explicit caller intent (e.g. `--simulate`) and produces a distinct typed `simulated` state that can never satisfy dependencies, acceptance criteria, gates, merge, or release. Normal execution exits nonzero with a typed capability failure when a required executor, reviewer, command, or CI provider is absent — no stub completion, no literal-`true` gates, no synthetic approvals, no empty-command passes. A manual gate with no automation enters a waiting state; it does not pass. Positive tests prove explicit simulation still works; negative controls prove simulation and every missing-provider case cannot advance lifecycle state.
+
+3. **RI-N3 One transitional PRD authority (SDLC-D-036).** `@mosaicstack/prdy` structured storage under `docs/prdy/`, driven by `mosaic mission --plan`, is the authoritative PRD representation for the alpha. `mosaic prdy` either routes through the same application service or operates only as an explicit, named Markdown import/export adapter; `docs/PRD.md` is not a peer authority. `mission --plan` must persist the mission↔PRD linkage (mission id/version, PRD id/version, selected requirements). Markdown output is a generated view carrying source identity; editing it cannot mutate authority silently. Import is explicit, validated, and conflict-aware (proposed successor, never overwrite). Structural validity is separate from approval.
+
+4. **RI-N4 One quality-rails evaluator (SDLC-D-037).** The TypeScript quality-rails package is the sole authoritative evaluator. A complete probe inventory maps every current TypeScript and shell check to one canonical check with disposition (preserve/strengthen/retire, each named). Effective shell enforcement probes are absorbed before their independent paths retire; expected-file presence alone is not parity. The evaluator returns typed results (`passed`/`failed`/`blocked`/`error`/`not-applicable`) with check version, subject, and reason; missing implementation, missing input, unknown check, process error, timeout, or malformed output can never become `passed` or an unqualified skip. Check definitions and policy are versioned and digested. Shell commands become thin adapters with no separate verdict logic. The canonical terminal verification command (RI-N1) invokes this evaluator rather than duplicating its logic. Contract, parity, and negative-control tests are required, plus independent review of probe equivalence.
+
+5. **RI-N5 Consequence-aware stale UI (SDLC-D-038).** Mission Control distinguishes typed freshness states (`current`, `stale`, `partial`, `unknown`, `unavailable`) rather than inferring from empty arrays or null. A failed fetch never renders as an empty healthy collection. Last-known data may display for situational awareness only with source identity, version, and age visibly labeled; any derived completion/assurance/release verdict whose inputs are stale becomes `unknown`; all state-changing actions are disabled until fresh state loads and is revalidated. With no verified snapshot, surfaces show an explicit unavailable state. Cache corruption, cross-workspace data, schema mismatch, and version regression invalidate the snapshot. Tests cover the failure matrix (network, auth, malformed, partial, corruption, stale age, schema mismatch, recovery, stale-action rejection) with negative controls proving no case yields a current green verdict or enabled mutation.
+
+### Acceptance criteria
+
+- AC-RI-1: A push to `next` that fails any mandatory verification step publishes nothing (no npm package, no image), demonstrated by a checked-in negative control and by pipeline evidence on a real `next` publish run where the verification step is green and every publish step depends on it.
+- AC-RI-2: With no executor/reviewer/CI provider wired, Forge and MACP normal runs exit nonzero with typed capability failures; with `--simulate`, runs complete but every result is typed `simulated` and cannot satisfy any gate, dependency, or completion state — proven by unit tests including negative controls.
+- AC-RI-3: A PRD created or revised through either `mosaic mission --plan` or `mosaic prdy` resolves to one authority under `docs/prdy/` with stable identities and versions; the mission↔PRD linkage survives restart; a Markdown export is labeled as generated and cannot silently become a second writer; divergent legacy content blocks baseline claims until explicitly resolved — proven by contract tests.
+- AC-RI-4: `quality-rails check` through any entry point (TS CLI, framework shell adapter) returns the same typed verdict for the same subject; the probe inventory names every legacy check's disposition; a deliberately broken probe fails closed — proven by contract/parity/negative-control tests and independent review of probe equivalence.
+- AC-RI-5: No shipping surface renders a failed fetch as an empty healthy state; stale/partial/unavailable states are typed, labeled, and mutation-disabled — proven by the failure-matrix tests.
+- AC-RI-6: All cards merged to `next` via squash PR with terminal-green CI; release evidence for 0.0.50 records commit, verification run, and published artifacts.
+
+### Out of scope
+
+The canonical dispatcher/control-plane vertical slice (work graph, execution attempts, fenced leases, typed check-in, independent verifier dispatch) is decided post-alpha (SDLC-D-033, option B). Multi-pipeline verification certificates (SDLC-D-034 option B) are post-alpha. Full AF-1..AF-4 objective matrices and Mission Control portfolio surfaces are post-alpha.

docs/release-integrity/TASKS.md

@@ -0,0 +1,42 @@
+# Tasks — Release Integrity Workstream (RI-050, #1275)
+
+> Single-writer: the RI-050 orchestrator (jarvis, dragon-lin) only. Workers read but never modify.
+>
+> **Mission:** alpha 0.0.50 release-integrity floor (decisions SDLC-D-033..038).
+> **PRD:** [docs/PRD.md § Release Integrity Workstream](../PRD.md#release-integrity-workstream-ri-1275)
+> **Issue:** #1275 (remains open until RI-V-001 closes)
+> **Base branch:** `next` (all cards branch from `origin/next`, squash-merge via PR)
+>
+> **Execution note:** the `agent` column uses `pi-glm-5.3` — outside the pipeline-cron model
+> table on purpose. This workstream is executed by jarvis on dragon-lin with local pi workers
+> (`pi --model zai/glm-5.3:high`); pipeline crons must not auto-claim these rows.
+>
+> **Status values:** `not-started` | `in-progress` | `done` | `blocked` | `failed` | `needs-qa`
+> `done` requires: repo quality gates green, independent review recorded, terminal-green CI on
+> the PR head, squash merge to `next`, and acceptance evidence in notes.
+
+| id       | status      | description                                                                                                                                                                                                                                                                      | issue | agent      | repo              | branch                            | depends_on                                                 | estimate | notes                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           |
+| -------- | ----------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----- | ---------- | ----------------- | --------------------------------- | ---------------------------------------------------------- | -------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
+| RI-0-001 | done        | Bootstrap: issue #1275, PRD section, this DAG, scratchpad (docs only)                                                                                                                                                                                                            | #1275 | pi-glm-5.3 | mosaicstack/stack | docs/ri-050-mission-bootstrap     | —                                                          | 6K       | PR #1276 (head 758659dd): docs-only, CI green (2475). Review requested from fargo. Merges first (no publish run).                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               |
+| RI-1-001 | done        | RI-N1: canonical terminal verification command + publish-pipeline exact-commit gate (every publish step depends on verify; commit identity check; fail closed)                                                                                                                   | #1275 | pi-glm-5.3 | mosaicstack/stack | feat/ri-050-publish-gate          | RI-0-001                                                   | 25K      | PR #1277 (head 46784c8d): CI GREEN at head after serialized retry (pipeline 2476, 2026-08-18) - earlier red was CI-agent contention (web SPA timeouts under concurrent pipelines), not code. Review requested from fargo at pinned head (comms 20260818T021025Z).                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               |
+| RI-1-002 | done        | RI-N1 negative control: checked-in tests proving a broken mandatory check blocks every publish step and that DAG edges cannot be bypassed                                                                                                                                        | #1275 | pi-glm-5.3 | mosaicstack/stack | test/ri-050-publish-gate-negative | RI-1-001                                                   | 12K      |                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 |
+| RI-2-001 | done        | RI-N2 (Forge): remove stub-executor false success; `--simulate` typed `simulated` results that satisfy nothing; literal-`true` gates and echo-review replaced with real gates or typed waiting-for-authority                                                                     | #1275 | pi-glm-5.3 | mosaicstack/stack | fix/ri-050-forge-fail-closed      | RI-0-001                                                   | 20K      | Independent review APPROVED 2026-08-17 (Gitea review 172 on PR #1278, head 99b8f6ea; reviewing seat fargo — recorded under shared host principal mos-dt-0, provenance correction posted by fred; wrapper gap filed by fred). Executed at head: forge tests 116/116, lint green, typecheck green after building macp dist (minimal-install artifact, not a defect), workspace typecheck 45/45, no external type consumers of the changed interfaces. CI red = known lane-wide fleet-test failure only, carries no information about this change (fred, log-content analysis, pipelines 2456-2458). Non-blocking finding: README L141-143 + skills/mosaic-forge/SKILL.md document bare forge run/resume, which now fails closed — fast-follow docs touch. Merge queued behind #1270. UPDATE 2026-08-18: #1270 merged; CI GREEN at head 4917df1f via serialized retry (pipeline 2477) - root cause of prior reds was CI-agent contention (web SPA timeouts under concurrent pipelines), superseding the fleet-test-failure theory. |
+| RI-2-002 | done        | RI-N2 (MACP): gate runner fails closed on empty commands, stub executors, and unimplemented CI-provider gates unless explicit simulate; typed capability failures                                                                                                                | #1275 | pi-glm-5.3 | mosaicstack/stack | fix/ri-050-macp-fail-closed       | RI-0-001                                                   | 15K      | PR #1293 (head 2097379e): CI green (pipeline 2465), independent review APPROVED (Gitea review 173, jarvis seat, 2026-08-17) - macp 109/109 verified at head. Merge queued behind #1276/#1277/#1278.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                             |
+| RI-3-001 | in-progress | RI-N4: complete probe inventory mapping every TS and shell quality-rail check to one canonical check with disposition (preserve/strengthen/retire, each named)                                                                                                                   | #1275 | pi-glm-5.3 | mosaicstack/stack | docs/ri-050-qr-probe-inventory    | RI-0-001                                                   | 12K      |                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 |
+| RI-3-002 | not-started | RI-N4: TS evaluator absorbs effective shell probes; typed results (passed/failed/blocked/error/not-applicable) with versioned digested check definitions; shell commands become thin adapters; contract/parity/negative-control tests                                            | #1275 | pi-glm-5.3 | mosaicstack/stack | feat/ri-050-qr-evaluator          | RI-3-001                                                   | 30K      |                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 |
+| RI-4-001 | in-progress | RI-N3: one PRD application service — `mission --plan` persists mission↔PRD linkage (ids/versions/selected requirements); `mosaic prdy` routes through the service or becomes a named import/export adapter; Markdown is a labeled generated view; explicit conflict-aware import | #1275 | pi-glm-5.3 | mosaicstack/stack | feat/ri-050-prd-authority         | RI-0-001                                                   | 35K      | PR #1294 (head 8d258e1d): CI green (pipeline 2466), independent review APPROVED (Gitea review 174, jarvis seat, 2026-08-17) - prdy 20/20 + command specs 9/9 at head. Merge queued behind #1276/#1277/#1278.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                    |
+| RI-5-001 | done        | RI-N5: typed freshness states (current/stale/partial/unknown/unavailable); no failed-fetch-renders-empty; stale derived verdicts → unknown; mutations disabled when stale; failure-matrix tests                                                                                  | #1275 | pi-glm-5.3 | mosaicstack/stack | feat/ri-050-web-stale-safety      | RI-0-001                                                   | 25K      |                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 |
+| RI-V-001 | not-started | Final verification + release evidence: all cards verified merged, negative controls demonstrated, real `next` publish run green on exact commit, evidence pack recorded                                                                                                          | #1275 | pi-glm-5.3 | mosaicstack/stack | docs/ri-050-release-evidence      | RI-1-002, RI-2-001, RI-2-002, RI-3-002, RI-4-001, RI-5-001 | 10K      |                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 |
+
+## Dispatch waves (max 2 parallel workers)
+
+1. RI-1-001 + RI-2-001
+2. RI-2-002 + RI-4-001
+3. RI-3-001 + RI-5-001
+4. RI-1-002 + RI-3-002
+5. RI-V-001
+
+## Budget
+
+Derived soft cap: 250K tokens (no explicit cap given). Projected total: 190K.
+Conservative mode (1 worker) above 70% projected; freeze above 90%.

docs/scratchpads/ri-050.md

@@ -0,0 +1,242 @@
+# Scratchpad — RI-050 orchestrator (jarvis, dragon-lin)
+
+Mission: alpha 0.0.50 release-integrity floor. Issue #1275. Base `next` @ 476db12b.
+Design SSOT: jarvis-brain `docs/plans/2026-08-16_mosaic-stack-sdlc-protocol.md` (SDLC-D-033..038).
+
+## Mode (Jason's directives)
+
+- Orchestrator: jarvis (this session, dragon-lin). NOT mos-claude; work stays on this host.
+- Workers: local pi headless — `pi --model zai/glm-5.3:high -p` in the card's worktree, tools read,bash,edit,write.
+- Delegation override of stack AGENTS.md `agent` column: rows carry `pi-glm-5.3` (outside cron table so no auto-claim).
+- Target branch: `next`. Cards branch from `origin/next`, squash-merge via PR.
+
+## Operational constraints (measured this session)
+
+- Main checkout at `/home/jwoltje/src/mosaic-stack` is a dirty diverged `main` (ahead 1139/behind 711) — NEVER touched. All work in `/home/jwoltje/src/mosaic-stack-worktrees/<branch>`.
+- Disk: /home 187G free. /tmp only 8.7G — keep pnpm stores/node_modules under /home.
+- `main` and `next` have DIVERGED; PRs target `next`.
+- Identity: pin `GITEA_LOGIN=mosaicstack-jarvis` for all wrapper ops. Issue #1275 verified authored by @jarvis.
+- `ci-queue-wait.sh` on this host is fail-open (board: fix #1032 not installed) — substitute SHA-status checks via `/commits/{sha}/status` and diff failing step names.
+- CI on PRs runs `pull_request` pipelines (any branch) incl. ci-postgres service. Push CI runs on main only; publish runs on push/tag to next + manual.
+- Wrapper gaps on this host per board (7 gaps; e.g. no pr-review-list, issue-assign broken, pr-merge makes no trailers): verify outcomes by reading back provider state, never trust rc alone.
+- Publish pipeline currently: install → build → publish-npm/publish-next-npm (+image). No verify. CI steps: install, sanitization, upgrade-guard, typecheck, lint, format, test, ci-postgres.
+
+## Budget
+
+Soft cap 250K. Projected 190K across 10 cards. Track per-card used vs estimate in TASKS.md notes.
+
+## Progress log
+
+- 2026-08-16 23:52 — Issue #1275 created (@jarvis verified).
+- 2026-08-16 23:5x — Bootstrap branch `docs/ri-050-mission-bootstrap` from origin/next@476db12b; PRD section + TASKS.md + this scratchpad written. RI-0-001 in-progress.
+
+## Wave 1 dispatched (2026-08-17 00:35)
+
+- RI-1-001 worker: pi glm-5.3:high, pid 2322125, worktree ri-1-001, log /var/tmp/ri-050/ri-1-001-run.log
+- RI-2-001 worker: pi glm-5.3:high, pid 2322126, worktree ri-2-001, log /var/tmp/ri-050/ri-2-001-run.log
+- Gotcha recorded: pi has no -f flag (that's pi-do.sh); pass brief as positional message. First launch died "Unknown option: -f" — relaunched.
+- CI lane: PR #1276 (bootstrap) fails `test` at base like every next PR — fred's green #1270 unblocks (comms sent 2026-08-17T05:21Z, `comms/20260817T052148Z__from-jarvis__650fe8.md`). Merge gate for all RI PRs queues behind #1270.
+- Live RI-N1 evidence posted to #1275 (comment 22915): pipeline 2439 publish-next-npm SUCCESS beside build-gateway FAILURE.
+
+---
+
+# HANDOFF — RI-050 continuation (written 2026-08-17 ~08:45 UTC, jarvis/dragon-lin)
+
+You are taking over the alpha 0.0.50 release-integrity workstream in place. Everything you
+need is on the remote. Read this whole file, then `docs/release-integrity/TASKS.md` (same
+branch), then the PRD section (`docs/PRD.md` § Release Integrity Workstream, same branch).
+
+## Identity / mode
+
+- Orchestrator identity: `jarvis` (dragon-lin). You continue as the RI-050 orchestrator under
+  whatever identity Jason gives you — if you are NOT jarvis, say so in comms and PR bodies.
+- Jason's standing directives for this mission: work happens on THIS repo (mosaicstack/stack),
+  PRs target `next` (NOT main), workers are local pi headless sessions on
+  `zai/glm-5.3:high`. Do not hand this to mos-claude. Do not borrow other seats' lanes.
+- All wrapper ops: pin `GITEA_LOGIN=mosaicstack-jarvis` (issue #1275 was verified authored by
+  @jarvis; keep identity consistent or verify yours with issue-view and READ BACK user.login).
+- CI substitution rule (this host's ci-queue-wait.sh is fail-open; fix #1032 not installed):
+  judge CI by SHA-status via `/api/v1/repos/mosaicstack/stack/commits/{sha}/status` or the
+  woodpecker API (`pipeline-status.sh -r mosaicstack/stack -n N -f json`), and DIFF THE
+  FAILING STEP NAMES rather than trusting rc.
+
+## Mission state at handoff
+
+Mission: alpha 0.0.50 release-integrity floor. Issue #1275 (open, has live-evidence comment).
+Decisions SDLC-D-033..038 live in jarvis-brain
+`docs/plans/2026-08-16_mosaic-stack-sdlc-protocol.md` (normative text also mirrored in the
+PRD section on this branch, so this repo is self-sufficient).
+
+Base: `origin/next` @ 476db12b. NOTE: `main` and `next` have DIVERGED — never base on main.
+
+Branches (all pushed, all clean trees):
+
+- `docs/ri-050-mission-bootstrap` @ 5114faa2 → PR #1276 (open, mergeable) — bootstrap docs +
+  this scratchpad + TASKS.md DAG. STATUS: CI red on `test` only, which is the known lane-wide
+  failure (see blocker below); own prettier issue already fixed.
+- `feat/ri-050-publish-gate` @ 0aa5ed35 → PR #1277 (open, mergeable) — RI-1-001 COMPLETE
+  (worker reported success, orchestrator review PASSED: verify step asserts CI_COMMIT_SHA ==
+  git rev-parse HEAD then runs canonical `pnpm verify:release`; every publish/image step
+  depends_on verify directly, confirmed by parsing the DAG: publish-npm, publish-next-npm,
+  build-gateway/appservice/web all -> [build, verify]; invariant test
+  scripts/verify-release.test.mjs passes 7/7 locally with negative fixtures). CI: same known
+  lane-red `test` step only.
+- `fix/ri-050-forge-fail-closed` @ 99b8f6ea → PR #1278 (open, mergeable) — RI-2-001 worker
+  reported success (typed `FORGE_*` capability errors, --simulate typed simulated everywhere,
+  vacuous true/echo gates replaced, closed ForgeOutcome set, 116 tests green incl. 16 new).
+  ORCHESTRATOR REVIEW NOT YET DONE — your first job. Review the diff
+  (1391 insertions across forge src), check the fail-closed paths and that simulated
+  results cannot satisfy any consumer, run `pnpm --filter @mosaicstack/forge test`.
+
+## The one blocker
+
+Every `next` PR pipeline is red on ONE assertion:
+`packages/mosaic/framework/tools/fleet/test-start-agent-session.sh:103` ("host provides 'pi'
+in the system path"). Pre-existing at base; affects PRs #1276/#1277/#1278 identically.
+fred's PR #1270 ("unblocks every PR on next") is green and open — it is HIS to merge; do not
+merge it yourself. jarvis sent comms (`comms/20260817T052148Z__from-jarvis__650fe8.md` in
+jarvis-brain) asking merge timing; no reply yet as of handoff. Merge gates for ALL RI PRs
+queue behind #1270 landing. Until then: review/develop freely, merge nothing that needs the
+green gate (docs-only #1276 arguably could merge red-lane with Jason's explicit call — ask,
+don't assume).
+
+## Remaining DAG (docs/release-integrity/TASKS.md is canonical)
+
+Wave 2 (next): RI-2-002 MACP fail-closed (brief pattern: mirror RI-2-001 for
+packages/macp/src/gate-runner.ts — empty commands, stub executors, unimplemented CI-provider
+gates fail closed; explicit simulate) and RI-4-001 PRD authority (one PRD service;
+@mosaicstack/prdy docs/prdy authoritative via `mosaic mission --plan`; `mosaic prdy` routes
+or becomes named Markdown adapter; mission<->PRD linkage persists — see PRD RI-N3).
+Wave 3: RI-3-001 probe inventory (docs), RI-5-001 web stale-safety.
+Wave 4: RI-1-002 negative-control tests, RI-3-002 TS evaluator absorbs shell probes.
+Final: RI-V-001 evidence pack (real green next publish run post-gate + all cards verified).
+
+## Worker mechanics (measured, reuse)
+
+- Dispatch: create worktree `git -C /home/jwoltje/src/mosaic-stack worktree add
+/home/jwoltje/src/mosaic-stack-worktrees/<id> -b <branch> origin/next`, write a brief to
+  /var/tmp/ri-050/, then run from INSIDE the worktree:
+  `pi -p --no-session --model zai/glm-5.3:high --tools read,bash,edit,write "$(cat brief.md)"`
+  (pi has NO -f flag — pass the brief as a positional message; first dispatch died on that).
+- Briefs for 1-001/2-001 are at /var/tmp/ri-050/ on dragon-lin (may not survive; the
+  pattern is fully described above and in TASKS.md).
+- Briefs must carry: worktree path, branch, base, requirements, known base-red list (so the
+  worker doesn't chase it), gates to run, PR creation command with GITEA_LOGIN pin, "do NOT
+  merge, do NOT touch docs/TASKS.md", and the JSON report format.
+- Verify worker claims: read the PR, run their tests yourself, parse pipeline step names.
+
+## Do-not-touch
+
+- Main checkout at /home/jwoltje/src/mosaic-stack (dirty diverged main) — never touch.
+- fred's open PRs (#1270 and others) — review evidence welcome, merging his is not yours.
+- Other RI PRs' authors' lanes: #1277/#1278 are yours to gate and merge ONCE lane is green
+  and review is recorded.
+- Never `--no-verify`; never bypass the wrapper-fails-closed rule (wrapper failure ⇒
+  `blocked + report exact command + stop`).
+
+## Session-restore command sequence
+
+1. `git -C /home/jwoltje/src/mosaic-stack-worktrees/ri-050 fetch origin --prune`
+2. Read this file + `docs/release-integrity/TASKS.md` + PRD section.
+3. Check PR states (#1270, #1276, #1277, #1278) and lane CI (SHA-status per above).
+4. Review RI-2-001 (PR #1278) if not yet done; then dispatch wave 2.
+
+— jarvis, 2026-08-17
+
+---
+
+# CONTINUATION — fargo (sb-it-1-dt)
+
+Orchestrator seat is now **fargo** on sb-it-1-dt (Jason, 2026-08-17): Claude seat, worktree discipline
+per fred's ruling (`~/agent-work/<slug>`, create → work → commit → push → remove as one act; the
+helper's `/src` refusal is a web1 convention, does not bind here). fred supports; lane rulings are
+his. Workers remain local pi `zai/glm-5.3:high` + limited Claude per Jason.
+
+## 2026-08-17 — RI-2-001 independent review DONE
+
+- **PR #1278 APPROVED** (Gitea review 172, pinned to head 99b8f6ea). Executed evidence, not read-only:
+  forge suite 116/116 at head (matches PR claim), forge lint green, forge typecheck green after
+  building `@mosaicstack/macp` dist (TS2307 on bare `pnpm install --frozen-lockfile` is a
+  minimal-install build-order artifact — the macp import is type-only, vitest passes unbuilt; CI
+  installs build workspace deps, hence green there), **workspace typecheck 45/45 at head**,
+  consumer sweep: no external type consumers of RunManifest/StageStatus/ForgeTaskResult/
+  TaskExecutor; only importer of the package is packages/mosaic via registerForgeCommand
+  (smoke test asserts registration/help only — cannot break). Digest gate (shaggy's) before==after
+  with both-arm reactivity controls.
+- CI red on #1276/#1277/#1278: lane-wide `test` failure only
+  (test-start-agent-session.sh:103, fred's guard mis-wired; #1270 unwires it). Fred measured log
+  content: one real byte-identical failure per pipeline (2456/2457/2458); 13 of ~14 `FAIL` grep
+  hits are passing fail-loud test NAMES. **The red carries no information about the RI changes.**
+- Non-blocking finding: README L141-143 + skills/mosaic-forge/SKILL.md document bare
+  `mosaic forge run`/`resume`, which now exits 1 FORGE_NO_EXECUTOR — fast-follow docs touch.
+- **Identity incident, ruled on by fred:** review 172 recorded under shared host principal
+  mos-dt-0, not fargo. Mechanism (measured, wrapper source): pr-review.sh resolves its acting login
+  from the tea login list only; no fargo tea login on this host → silent host-default fallback;
+  MOSAIC_GIT_IDENTITY is only read in detect-platform.sh get_gitea_token's fallback arm, never
+  reached. Exact-id read-back verifies against the writing token, so it passed while attribution
+  was wrong — durable-provenance machinery proves the write, not the seat. Fred's ruling: review
+  172 stands (substance/verdict/pin correct; label wrong); NO re-approval (one approval,
+  annotated, is the stronger record); fred posts the provenance correction under @fred with
+  --login fred-ms (hard-fail path); no fargo tea login ever (freeze + Jason's to authorize);
+  tooling gap filed by fred. Also explains (does not reopen) #1228's mos-dt-0 attribution.
+- Merge gate: all RI PRs queue behind fred's green #1270 (Jason's call).
+
+## Next
+
+1. Wave 2 dispatch: RI-2-002 (MACP fail-closed, mirror RI-2-001 pattern for
+   packages/macp/src/gate-runner.ts) + RI-4-001 (PRD authority). Two parallel workers max.
+2. Docs fast-follow (README + mosaic-forge skill) — fold into #1276 or a tiny docs card.
+3. RI-V-001 evidence at the end.
+
+— fargo, 2026-08-17
+
+---
+
+# RESUMPTION + DAILY-HANDOFF PROTOCOL (Jason, 2026-08-17)
+
+Orchestrator seat is back with **jarvis** (dragon-lin). Expect daily handoff between jarvis
+and fargo. Protocol (both seats, every handoff):
+
+1. **This file is the shared mission log.** Append a dated section per session: state
+   measured, actions taken, PR/review states, next actions. Never rewrite prior sections.
+2. **TASKS.md stays current within one session** — status, PR number in notes, review
+   evidence. Stale rows are handoff debt.
+3. **Cross-review rule (SDLC-D-011 in practice):** the reviewing seat must differ from the
+   producing seat. jarvis reviews fargo-dispatched PRs, fargo reviews jarvis-dispatched
+   PRs. Producers are always pi workers; dispatching seats verify before push; the other
+   seat records the Gitea review.
+4. Handoff = append here + push + (optional) issue #1275 comment if a decision changed.
+
+## RESUMED — jarvis/dragon-lin, 2026-08-17 (afternoon)
+
+- Measured: next = 8199261c (#1270 merged — lane unblocked for new PRs). #1293/#1294
+  (fargo, wave 2) CI-green, mergeable, no recorded reviews. #1276/#1277/#1278 still based
+  on 476db12b with stale red CI → need rebase onto 8199261c. #1278 review pinned to old
+  head 99b8f6ea by @mos-dt-0 (fargo's, mis-attributed per his note) — rebase will dismiss
+  it; re-approval must come from fargo/fred (author is @jarvis, cannot self-approve).
+- Live evidence #2: push pipeline 2462 (the #1270 merge itself) ran publish-next-npm
+  SUCCESS beside build-gateway FAILURE again.
+- Plan: rebase the three original branches; independently review #1293/#1294; merge order
+  once green+reviewed: #1276 (docs) → #1277 (publish gate) → #1278/#1293/#1294 (code).
+  After #1277 merges, watch the next push pipeline prove the verify gate live.
+- fargo's non-RI PRs (#1291/#1296/#1297/#1281) stay strictly his lane.
+
+## jarvis session 2026-08-17 (evening) — reviews, rebases, merge plan
+
+- Rebased #1276/#1277/#1278 onto 8199261c (heads 59e2c460 / 46784c8d / 4917df1f);
+  invariant tests 7/7 and forge 116/116 re-run green at new heads. #1270 touched
+  test-enumeration-exclusions.txt + package.json, NOT ci.yml — no semantic overlap with
+  #1277's ci.yml changes (checked, was a real concern).
+- Independent reviews recorded: #1293 APPROVED (review 173; macp 109/109; fail-closed paths
+  + aggregate state machine verified), #1294 APPROVED (review 174; prdy 20/20 + command
+  specs 9/9; single-writer + linkage persistence + labeled export + conflict-aware import
+  verified). Note: 19 unrelated mosaic suites fail on bare minimal install (known workspace
+  build-order artifact, documented by fargo) — not this change.
+- Measured: `next` has NO branch protection (API: only main listed). Cross-seat review
+  discipline is protocol-enforced, not Gitea-enforced. Flagged to fargo for Jason: direct
+  pushes to next trigger ungated publishes; protection is Jason's call (#1231 adjacent).
+- Merge order planned: #1276 (docs-only — no publish run) -> #1277 (first gated publish)
+  -> #1278 -> #1293 -> #1294. Sent fargo review requests with pinned head SHAs
+  (comms/20260818T011932Z__from-jarvis__a9c02b.md). Not merging #1293/#1294 before my three
+  clear fargo's review — order optimality beats speed; every pre-#1277 merge publishes ungated.
+- CI on the three rebased heads: pending at time of this entry.

eslint.config.mjs

@@ -34,6 +34,7 @@ export default tseslint.config(
             'packages/storage/vitest.config.ts',
             'packages/mosaic/vitest.config.ts',
             'packages/mosaic/__tests__/*.ts',
+            'packages/forge/__tests__/*.ts',
             'tools/federation-harness/*.ts',
           ],
         },

package.json

@@ -9,6 +9,7 @@
     "preflight": "node scripts/preflight.mjs",
     "clean:generated": "node scripts/clean-generated.mjs",
     "typecheck": "pnpm preflight && turbo run typecheck",
+    "verify:release": "node scripts/verify-release.mjs",
     "test:checkout": "node --test scripts/*.test.mjs",
     "test": "pnpm test:checkout && turbo run test && pnpm run test:installer",
     "test:installer": "bash tools/install-next-lane.test.sh",

packages/forge/PLAN.md

@@ -539,3 +539,43 @@ Not every brief needs full Board of Directors review. The classification system
 ### Backward compatibility
 
 Existing briefs without a `class` field are auto-classified. The default (no matching keywords) is `strategic`, so all existing runs get the full pipeline unless keywords trigger `technical`.
+
+---
+
+## Fail-Closed Execution & Explicit Simulation (SDLC-D-035)
+
+**Added:** 2026-08-17
+
+Forge fails closed when a required capability is missing. It never runs a
+pipeline with a stub executor and reports success.
+
+### Normal mode (default)
+
+- No task executor wired → the CLI exits nonzero with the typed capability
+  error `FORGE_NO_EXECUTOR`. No run is created.
+- A stage whose gate is approval-based (board approval, planning approvals,
+  remediation re-review, discovery/analysis attestations) records a typed
+  `waiting-for-authority` stage result and raises `FORGE_AUTHORITY_REQUIRED`.
+  It never passes vacuously.
+- A stage whose gate requires an unwired provider (AI reviewer, CI pipeline)
+  records a typed `blocked` stage result and raises `FORGE_NO_REVIEWER` /
+  `FORGE_NO_CI_PIPELINE`. The synthetic echo-review approval in `06-review`
+  and all vacuous `true` gates were removed.
+
+### Explicit simulation (`--simulate`)
+
+Opts into stub/synthetic execution. Every stage result, every gate result, and
+the run manifest carry the distinct typed status `simulated` (manifest also
+records `mode: "simulated"`). `simulated` is a non-satisfying outcome:
+`isSatisfyingOutcome()` and all completion/gate consumers treat only `passed`
+as satisfying. The CLI exits 0 for a simulated run only because the caller
+explicitly passed `--simulate`, and prints a loud SIMULATED banner.
+
+### Typed outcome model
+
+Every gate/task outcome is one of the closed set
+`passed | failed | blocked | error | waiting-for-authority | simulated |
+not-applicable`, with the reason recorded on the stage status and each gate
+result in `manifest.json`. Missing implementations, missing gate evidence,
+unknown stages, process errors, and timeouts map to fail-closed members —
+never to `passed`.

packages/forge/__tests__/fail-closed.test.ts

@@ -0,0 +1,319 @@
+import fs from 'node:fs';
+import os from 'node:os';
+import path from 'node:path';
+import { describe, it, expect, beforeEach, afterEach } from 'vitest';
+
+import { generateBoardTasks } from '../src/board-tasks.js';
+import { STAGE_SPECS } from '../src/constants.js';
+import { ForgeCapabilityError } from '../src/errors.js';
+import {
+  evaluateStageGates,
+  gateLabel,
+  isCommandGate,
+  isSatisfyingOutcome,
+} from '../src/outcomes.js';
+import { loadManifest, runPipeline } from '../src/pipeline-runner.js';
+import type { ForgeTask, ForgeTaskResult, TaskExecutor } from '../src/types.js';
+
+/**
+ * Mock real executor that returns typed results.
+ *
+ * Command gates are "verified" by the mock so normal-mode runs can pass
+ * mechanically gated stages; authority/provider gates are never reported
+ * because they have no mechanical implementation.
+ */
+function createTypedExecutor(options?: {
+  failStage?: string;
+  gateOutcomes?: Record<string, 'passed' | 'failed' | 'simulated' | 'error' | 'blocked'>;
+}): TaskExecutor & { submittedTasks: ForgeTask[] } {
+  const submittedTasks: ForgeTask[] = [];
+  return {
+    submittedTasks,
+    async submitTask(task: ForgeTask) {
+      submittedTasks.push(task);
+    },
+    async waitForCompletion(taskId: string): Promise<ForgeTaskResult> {
+      const task = submittedTasks.find((t) => t.id === taskId);
+      const stageName = task?.metadata?.['stageName'] as string | undefined;
+
+      if (options?.failStage && stageName === options.failStage) {
+        return {
+          task_id: taskId,
+          outcome: 'failed',
+          reason: 'mock task failure',
+          completed_at: new Date().toISOString(),
+          exit_code: 1,
+          gate_results: [],
+        };
+      }
+
+      const gateResults = (task?.qualityGates ?? [])
+        .filter((gate) => isCommandGate(gate))
+        .map((gate) => {
+          const label = gateLabel(gate);
+          const outcome = options?.gateOutcomes?.[label] ?? 'passed';
+          return {
+            gate: label,
+            outcome,
+            reason: outcome === 'passed' ? 'mock verified' : `mock gate outcome: ${outcome}`,
+          };
+        });
+
+      return {
+        task_id: taskId,
+        outcome: 'passed',
+        reason: 'mock verified',
+        completed_at: new Date().toISOString(),
+        exit_code: 0,
+        gate_results: gateResults,
+      };
+    },
+    async getTaskStatus() {
+      return 'completed' as const;
+    },
+  };
+}
+
+describe('fail-closed: no executor wired', () => {
+  let tmpDir: string;
+  let briefPath: string;
+
+  beforeEach(() => {
+    tmpDir = fs.mkdtempSync(path.join(os.tmpdir(), 'forge-failclosed-'));
+    briefPath = path.join(tmpDir, 'brief.md');
+    fs.writeFileSync(briefPath, '# Fix bug\n\nA bugfix for lint cleanup.');
+  });
+
+  afterEach(() => {
+    fs.rmSync(tmpDir, { recursive: true, force: true });
+  });
+
+  it('throws a typed FORGE_NO_EXECUTOR capability error without --simulate', async () => {
+    await expect(
+      runPipeline(briefPath, tmpDir, {
+        // no executor, no simulate — must fail closed, never run with a stub
+        stages: ['00-intake'],
+      }),
+    ).rejects.toMatchObject({
+      name: 'ForgeCapabilityError',
+      code: 'FORGE_NO_EXECUTOR',
+      capability: 'task-executor',
+    });
+  });
+
+  it('does not create a run directory when failing closed on a missing executor', async () => {
+    try {
+      await runPipeline(briefPath, tmpDir, { stages: ['00-intake'] });
+    } catch {
+      // expected
+    }
+    expect(fs.existsSync(path.join(tmpDir, '.forge', 'runs'))).toBe(false);
+  });
+
+  it('completes with every result typed simulated when simulate is set', async () => {
+    const result = await runPipeline(briefPath, tmpDir, {
+      simulate: true,
+      stages: ['00-intake', '00b-discovery', '02-planning-1', '06-review'],
+    });
+
+    expect(result.manifest.mode).toBe('simulated');
+    expect(result.manifest.status).toBe('simulated');
+
+    for (const stage of result.stages) {
+      const stageStatus = result.manifest.stages[stage];
+      expect(stageStatus?.status, `stage ${stage}`).toBe('simulated');
+      expect(stageStatus?.status, `stage ${stage}`).not.toBe('passed');
+      expect(stageStatus?.reason, `stage ${stage}`).toBeTruthy();
+      for (const gateResult of stageStatus?.gateResults ?? []) {
+        expect(gateResult.outcome, `gate ${gateResult.gate} of ${stage}`).toBe('simulated');
+        expect(gateResult.outcome, `gate ${gateResult.gate} of ${stage}`).not.toBe('passed');
+      }
+    }
+
+    // The persisted manifest agrees.
+    const persisted = loadManifest(result.runDir);
+    expect(persisted.mode).toBe('simulated');
+    expect(persisted.status).toBe('simulated');
+    expect(persisted.stages['02-planning-1']?.status).toBe('simulated');
+  });
+});
+
+describe('fail-closed: typed outcome model', () => {
+  it('only passed satisfies the gate/dependency predicate', () => {
+    expect(isSatisfyingOutcome('passed')).toBe(true);
+    expect(isSatisfyingOutcome('failed')).toBe(false);
+    expect(isSatisfyingOutcome('blocked')).toBe(false);
+    expect(isSatisfyingOutcome('error')).toBe(false);
+    expect(isSatisfyingOutcome('waiting-for-authority')).toBe(false);
+    expect(isSatisfyingOutcome('simulated')).toBe(false);
+    expect(isSatisfyingOutcome('not-applicable')).toBe(false);
+  });
+
+  it('a simulated gate result cannot satisfy the stage gate evaluation', () => {
+    const evaluation = evaluateStageGates('05-coding', STAGE_SPECS['05-coding']!.qualityGates, {
+      task_id: 'FORGE-x-05',
+      outcome: 'passed',
+      reason: 'executor claims success',
+      completed_at: new Date().toISOString(),
+      exit_code: 0,
+      gate_results: [{ gate: 'pnpm lint', outcome: 'simulated', reason: 'simulated gate' }],
+    });
+    expect(isSatisfyingOutcome(evaluation.outcome)).toBe(false);
+    expect(evaluation.outcome).toBe('error');
+  });
+
+  it('a simulated task outcome cannot satisfy evaluation in normal mode', () => {
+    const evaluation = evaluateStageGates('00-intake', [], {
+      task_id: 'FORGE-x-00',
+      outcome: 'simulated',
+      reason: 'executor reported simulated',
+      completed_at: new Date().toISOString(),
+      exit_code: 0,
+      gate_results: [],
+    });
+    expect(isSatisfyingOutcome(evaluation.outcome)).toBe(false);
+  });
+
+  it('a missing gate result blocks the stage instead of passing vacuously', () => {
+    const evaluation = evaluateStageGates('05-coding', STAGE_SPECS['05-coding']!.qualityGates, {
+      task_id: 'FORGE-x-05',
+      outcome: 'passed',
+      reason: 'executor claims success',
+      completed_at: new Date().toISOString(),
+      exit_code: 0,
+      gate_results: [],
+    });
+    expect(evaluation.outcome).toBe('blocked');
+  });
+});
+
+describe('fail-closed: authority and provider gates', () => {
+  let tmpDir: string;
+  let briefPath: string;
+
+  beforeEach(() => {
+    tmpDir = fs.mkdtempSync(path.join(os.tmpdir(), 'forge-authority-'));
+    briefPath = path.join(tmpDir, 'brief.md');
+    fs.writeFileSync(briefPath, '# Fix bug\n\nA bugfix for lint cleanup.');
+  });
+
+  afterEach(() => {
+    fs.rmSync(tmpDir, { recursive: true, force: true });
+  });
+
+  it.each(['02-planning-1', '03-planning-2', '04-planning-3', '07-remediate'])(
+    'planning/remediation stage %s yields waiting-for-authority (not passed) in normal mode',
+    async (stage) => {
+      const executor = createTypedExecutor();
+      let runDir: string | undefined;
+
+      try {
+        await runPipeline(briefPath, tmpDir, {
+          executor,
+          stages: [stage as string],
+        });
+        expect.unreachable('runPipeline should have failed closed');
+      } catch (err) {
+        expect(err).toBeInstanceOf(ForgeCapabilityError);
+        expect((err as ForgeCapabilityError).code).toBe('FORGE_AUTHORITY_REQUIRED');
+        runDir = path.join(tmpDir, '.forge', 'runs');
+      }
+
+      const runIds = fs.readdirSync(runDir!);
+      expect(runIds).toHaveLength(1);
+      const manifest = loadManifest(path.join(runDir!, runIds[0]!));
+      expect(manifest.stages[stage]?.status).toBe('waiting-for-authority');
+      expect(manifest.stages[stage]?.status).not.toBe('passed');
+      expect(manifest.status).toBe('waiting-for-authority');
+    },
+  );
+
+  it('review stage fails closed with a typed FORGE_NO_REVIEWER error in normal mode', async () => {
+    const executor = createTypedExecutor();
+
+    try {
+      await runPipeline(briefPath, tmpDir, {
+        executor,
+        stages: ['06-review'],
+      });
+      expect.unreachable('runPipeline should have failed closed');
+    } catch (err) {
+      expect(err).toBeInstanceOf(ForgeCapabilityError);
+      expect((err as ForgeCapabilityError).code).toBe('FORGE_NO_REVIEWER');
+      expect((err as ForgeCapabilityError).capability).toBe('reviewer');
+    }
+
+    const runsDir = path.join(tmpDir, '.forge', 'runs');
+    const runIds = fs.readdirSync(runsDir);
+    const manifest = loadManifest(path.join(runsDir, runIds[0]!));
+    expect(manifest.stages['06-review']?.status).toBe('blocked');
+    expect(manifest.stages['06-review']?.status).not.toBe('passed');
+    expect(manifest.status).toBe('failed');
+  });
+
+  it('review stage produces simulated results under --simulate', async () => {
+    const result = await runPipeline(briefPath, tmpDir, {
+      simulate: true,
+      stages: ['06-review'],
+    });
+
+    expect(result.manifest.mode).toBe('simulated');
+    expect(result.manifest.stages['06-review']?.status).toBe('simulated');
+    for (const gateResult of result.manifest.stages['06-review']?.gateResults ?? []) {
+      expect(gateResult.outcome).toBe('simulated');
+    }
+  });
+
+  it('deploy stage fails closed without a wired ci-pipeline provider in normal mode', async () => {
+    const executor = createTypedExecutor();
+
+    await expect(
+      runPipeline(briefPath, tmpDir, {
+        executor,
+        stages: ['09-deploy'],
+      }),
+    ).rejects.toMatchObject({
+      name: 'ForgeCapabilityError',
+      code: 'FORGE_NO_CI_PIPELINE',
+    });
+  });
+});
+
+describe('fail-closed: no vacuous gate commands remain', () => {
+  it('stage constants contain no echo/synthetic-approval, vacuous true, or empty gate commands', () => {
+    for (const [stageName, spec] of Object.entries(STAGE_SPECS)) {
+      for (const gate of spec.qualityGates) {
+        const serialized = JSON.stringify(gate);
+        // The echo-review synthetic approval must be gone.
+        expect(serialized, `stage ${stageName} gate ${serialized}`).not.toContain('echo');
+        expect(serialized, `stage ${stageName} gate ${serialized}`).not.toMatch(/"verdict"\s*:/);
+        expect(serialized, `stage ${stageName} gate ${serialized}`).not.toMatch(
+          /"summary"\s*:\s*"review-pass"/,
+        );
+        // No vacuous literal `true` gate.
+        expect(gate, `stage ${stageName}`).not.toBe('true');
+        // Command gates must carry a real, non-empty command.
+        if (isCommandGate(gate)) {
+          const command = typeof gate === 'string' ? gate : gate.command;
+          expect(command.trim().length, `stage ${stageName} gate ${serialized}`).toBeGreaterThan(0);
+        }
+      }
+    }
+  });
+
+  it('board tasks contain no vacuous true gates', () => {
+    const tmpDir = fs.mkdtempSync(path.join(os.tmpdir(), 'forge-board-gates-'));
+    try {
+      const tasks = generateBoardTasks('# Brief', [], tmpDir, 'BOARD-TEST');
+      for (const task of tasks) {
+        for (const gate of task.qualityGates) {
+          expect(gate, `task ${task.id}`).not.toBe('true');
+          const serialized = JSON.stringify(gate);
+          expect(serialized, `task ${task.id} gate ${serialized}`).not.toContain('echo');
+        }
+      }
+    } finally {
+      fs.rmSync(tmpDir, { recursive: true, force: true });
+    }
+  });
+});

packages/forge/__tests__/pipeline-runner.test.ts

@@ -12,10 +12,10 @@ import {
   resumePipeline,
   getPipelineStatus,
 } from '../src/pipeline-runner.js';
-import type { ForgeTask, RunManifest, TaskExecutor } from '../src/types.js';
-import type { TaskResult } from '@mosaicstack/macp';
+import type { ForgeTask, ForgeTaskResult, RunManifest, TaskExecutor } from '../src/types.js';
+import { gateLabel, isCommandGate } from '../src/outcomes.js';
 
-/** Mock TaskExecutor that records submitted tasks and returns success. */
+/** Mock TaskExecutor that records submitted tasks and returns typed results. */
 function createMockExecutor(options?: {
   failStage?: string;
 }): TaskExecutor & { submittedTasks: ForgeTask[] } {
@@ -25,7 +25,7 @@ function createMockExecutor(options?: {
     async submitTask(task: ForgeTask) {
       submittedTasks.push(task);
     },
-    async waitForCompletion(taskId: string): Promise<TaskResult> {
+    async waitForCompletion(taskId: string): Promise<ForgeTaskResult> {
       const failStage = options?.failStage;
       const task = submittedTasks.find((t) => t.id === taskId);
       const stageName = task?.metadata?.['stageName'] as string | undefined;
@@ -33,7 +33,8 @@ function createMockExecutor(options?: {
       if (failStage && stageName === failStage) {
         return {
           task_id: taskId,
-          status: 'failed',
+          outcome: 'failed',
+          reason: 'mock task failure',
           completed_at: new Date().toISOString(),
           exit_code: 1,
           gate_results: [],
@@ -41,10 +42,17 @@ function createMockExecutor(options?: {
       }
       return {
         task_id: taskId,
-        status: 'completed',
+        outcome: 'passed',
+        reason: 'mock verified',
         completed_at: new Date().toISOString(),
         exit_code: 0,
-        gate_results: [],
+        gate_results: (task?.qualityGates ?? [])
+          .filter((gate) => isCommandGate(gate))
+          .map((gate) => ({
+            gate: gateLabel(gate),
+            outcome: 'passed' as const,
+            reason: 'mock verified',
+          })),
       };
     },
     async getTaskStatus() {
@@ -156,12 +164,13 @@ describe('runPipeline', () => {
     const executor = createMockExecutor();
     const result = await runPipeline(briefPath, tmpDir, {
       executor,
-      stages: ['00-intake', '00b-discovery'],
+      stages: ['00-intake', '05-coding'],
     });
 
     expect(result.runId).toMatch(/^\d{8}-\d{6}$/);
-    expect(result.stages).toEqual(['00-intake', '00b-discovery']);
+    expect(result.stages).toEqual(['00-intake', '05-coding']);
     expect(result.manifest.status).toBe('completed');
+    expect(result.manifest.mode).toBe('normal');
     expect(executor.submittedTasks).toHaveLength(2);
   });
 
@@ -180,12 +189,17 @@ describe('runPipeline', () => {
     const executor = createMockExecutor();
     const result = await runPipeline(briefPath, tmpDir, {
       executor,
-      stages: ['00-intake', '00b-discovery'],
+      stages: ['00-intake', '05-coding'],
     });
 
     const manifest = loadManifest(result.runDir);
     expect(manifest.stages['00-intake']?.status).toBe('passed');
-    expect(manifest.stages['00b-discovery']?.status).toBe('passed');
+    expect(manifest.stages['05-coding']?.status).toBe('passed');
+    expect(manifest.stages['05-coding']?.gateResults?.map((g) => g.outcome)).toEqual([
+      'passed',
+      'passed',
+      'passed',
+    ]);
   });
 
   it('respects CLI class override', async () => {
@@ -215,7 +229,7 @@ describe('runPipeline', () => {
     const executor = createMockExecutor();
     await runPipeline(briefPath, tmpDir, {
       executor,
-      stages: ['00-intake', '00b-discovery', '02-planning-1'],
+      stages: ['00-intake', '05-coding', '08-test'],
     });
 
     expect(executor.submittedTasks[0]!.dependsOn).toBeUndefined();
@@ -224,14 +238,14 @@ describe('runPipeline', () => {
   });
 
   it('handles stage failure', async () => {
-    const executor = createMockExecutor({ failStage: '00b-discovery' });
+    const executor = createMockExecutor({ failStage: '05-coding' });
 
     await expect(
       runPipeline(briefPath, tmpDir, {
         executor,
-        stages: ['00-intake', '00b-discovery'],
+        stages: ['00-intake', '05-coding'],
       }),
-    ).rejects.toThrow('Stage 00b-discovery failed');
+    ).rejects.toThrow('Stage 05-coding failed');
   });
 
   it('marks manifest as failed on stage failure', async () => {
@@ -270,30 +284,143 @@ describe('resumePipeline', () => {
     fs.rmSync(tmpDir, { recursive: true, force: true });
   });
 
-  it('resumes from first incomplete stage', async () => {
-    // First run fails on discovery
-    const executor1 = createMockExecutor({ failStage: '00b-discovery' });
-    let runDir: string;
+  it('resumes from first incomplete stage and fails closed at the next provider gate', async () => {
+    // Simulate a run whose authority stages were approved out-of-band
+    // (recorded as passed) and whose coding stage failed mechanically.
+    const runId = '20260101-000000';
+    const runDir = path.join(tmpDir, '.forge', 'runs', runId);
+    fs.mkdirSync(runDir, { recursive: true });
+    const passed = { status: 'passed' as const, startedAt: '2026-01-01T00:00:00Z' };
+    saveManifest(runDir, {
+      runId,
+      brief: briefPath,
+      codebase: tmpDir,
+      briefClass: 'hotfix',
+      classSource: 'frontmatter',
+      forceBoard: false,
+      mode: 'normal',
+      createdAt: '2026-01-01T00:00:00Z',
+      updatedAt: '2026-01-01T00:00:00Z',
+      currentStage: '05-coding',
+      status: 'failed',
+      stages: {
+        '00-intake': passed,
+        '00b-discovery': passed,
+        '02-planning-1': passed,
+        '03-planning-2': passed,
+        '04-planning-3': passed,
+        '05-coding': { status: 'failed', reason: 'gate failed' },
+      },
+    });
 
-    try {
-      await runPipeline(briefPath, tmpDir, {
-        executor: executor1,
-        stages: ['00-intake', '00b-discovery', '02-planning-1'],
-      });
-    } catch {
-      // expected
+    // Resume re-runs 05-coding (the first non-passed stage), then fails
+    // closed at 06-review because no reviewer provider is wired.
+    const executor = createMockExecutor();
+    await expect(resumePipeline(runDir, executor)).rejects.toMatchObject({
+      name: 'ForgeCapabilityError',
+      code: 'FORGE_NO_REVIEWER',
+    });
+
+    const manifest = loadManifest(runDir);
+    expect(manifest.stages['05-coding']?.status).toBe('passed');
+    expect(manifest.stages['06-review']?.status).toBe('blocked');
+    expect(manifest.status).toBe('failed');
+  });
+
+  it('resumes to completion as simulated under explicit simulate', async () => {
+    const runId = '20260101-000003';
+    const runDir = path.join(tmpDir, '.forge', 'runs', runId);
+    fs.mkdirSync(runDir, { recursive: true });
+    const passed = { status: 'passed' as const, startedAt: '2026-01-01T00:00:00Z' };
+    saveManifest(runDir, {
+      runId,
+      brief: briefPath,
+      codebase: tmpDir,
+      briefClass: 'hotfix',
+      classSource: 'frontmatter',
+      forceBoard: false,
+      mode: 'normal',
+      createdAt: '2026-01-01T00:00:00Z',
+      updatedAt: '2026-01-01T00:00:00Z',
+      currentStage: '05-coding',
+      status: 'failed',
+      stages: {
+        '00-intake': passed,
+        '00b-discovery': passed,
+        '02-planning-1': passed,
+        '03-planning-2': passed,
+        '04-planning-3': passed,
+        '05-coding': { status: 'failed', reason: 'gate failed' },
+      },
+    });
+
+    const result = await resumePipeline(runDir, undefined, { simulate: true });
+
+    expect(result.manifest.status).toBe('simulated');
+    expect(result.manifest.mode).toBe('simulated');
+    expect(result.stages[0]).toBe('05-coding');
+    for (const stage of result.stages) {
+      expect(result.manifest.stages[stage]?.status).toBe('simulated');
     }
+  });
 
-    const runsDir = path.join(tmpDir, '.forge', 'runs');
-    runDir = path.join(runsDir, fs.readdirSync(runsDir)[0]!);
+  it('fails closed on resume when the next stage needs authority sign-off', async () => {
+    const runId = '20260101-000001';
+    const runDir = path.join(tmpDir, '.forge', 'runs', runId);
+    fs.mkdirSync(runDir, { recursive: true });
+    saveManifest(runDir, {
+      runId,
+      brief: briefPath,
+      codebase: tmpDir,
+      briefClass: 'hotfix',
+      classSource: 'frontmatter',
+      forceBoard: false,
+      mode: 'normal',
+      createdAt: '2026-01-01T00:00:00Z',
+      updatedAt: '2026-01-01T00:00:00Z',
+      currentStage: '00-intake',
+      status: 'in_progress',
+      stages: {
+        '00-intake': { status: 'passed' },
+      },
+    });
 
-    // Resume should pick up from 00b-discovery
-    const executor2 = createMockExecutor();
-    const result = await resumePipeline(runDir, executor2);
+    const executor = createMockExecutor();
+    await expect(resumePipeline(runDir, executor)).rejects.toMatchObject({
+      name: 'ForgeCapabilityError',
+      code: 'FORGE_AUTHORITY_REQUIRED',
+    });
 
-    expect(result.manifest.status).toBe('completed');
-    // Should have re-run from 00b-discovery onward
-    expect(result.stages[0]).toBe('00b-discovery');
+    const manifest = loadManifest(runDir);
+    expect(manifest.stages['00b-discovery']?.status).toBe('waiting-for-authority');
+    expect(manifest.status).toBe('waiting-for-authority');
+  });
+
+  it('fails closed on resume without an executor or --simulate', async () => {
+    const runId = '20260101-000002';
+    const runDir = path.join(tmpDir, '.forge', 'runs', runId);
+    fs.mkdirSync(runDir, { recursive: true });
+    saveManifest(runDir, {
+      runId,
+      brief: briefPath,
+      codebase: tmpDir,
+      briefClass: 'hotfix',
+      classSource: 'frontmatter',
+      forceBoard: false,
+      mode: 'normal',
+      createdAt: '2026-01-01T00:00:00Z',
+      updatedAt: '2026-01-01T00:00:00Z',
+      currentStage: '00-intake',
+      status: 'in_progress',
+      stages: {
+        '00-intake': { status: 'passed' },
+      },
+    });
+
+    await expect(resumePipeline(runDir)).rejects.toMatchObject({
+      name: 'ForgeCapabilityError',
+      code: 'FORGE_NO_EXECUTOR',
+    });
   });
 });
 

packages/forge/src/board-tasks.ts

@@ -95,7 +95,14 @@ export function generateBoardTasks(
       briefPath,
       resultPath: resultRelPath,
       timeoutSeconds: 120,
-      qualityGates: ['true'],
+      qualityGates: [
+        {
+          kind: 'authority',
+          capability: 'board-approval',
+          reason:
+            'persona evaluation is judged by board synthesis (authority review); no mechanical gate exists',
+        },
+      ],
       metadata: {
         personaName: persona.name,
         personaSlug: persona.slug,
@@ -121,7 +128,13 @@ export function generateBoardTasks(
     timeoutSeconds: 120,
     dependsOn: personaTaskIds,
     dependsOnPolicy: 'all_terminal',
-    qualityGates: ['true'],
+    qualityGates: [
+      {
+        kind: 'authority',
+        capability: 'board-approval',
+        reason: 'board synthesis is an authority decision; no mechanical gate exists',
+      },
+    ],
     metadata: {
       resultOutputPath: synthesisResult,
       inputResultPaths: personaResultPaths,

packages/forge/src/cli.spec.ts

@@ -1,7 +1,11 @@
+import fs from 'node:fs';
+import os from 'node:os';
+import path from 'node:path';
 import { Command } from 'commander';
-import { describe, expect, it } from 'vitest';
+import { describe, expect, it, vi, beforeEach, afterEach } from 'vitest';
 
 import { registerForgeCommand } from './cli.js';
+import { loadManifest } from './pipeline-runner.js';
 
 describe('registerForgeCommand', () => {
   it('registers a "forge" command on the parent program', () => {
@@ -55,3 +59,94 @@ describe('registerForgeCommand', () => {
     }).not.toThrow();
   });
 });
+
+describe('forge run fail-closed behavior (SDLC-D-035)', () => {
+  let tmpDir: string;
+  let briefPath: string;
+  let errSpy: ReturnType<typeof vi.spyOn>;
+  let logSpy: ReturnType<typeof vi.spyOn>;
+  let prevExitCode: string | number | null | undefined;
+
+  const parse = (args: string[]) => {
+    const program = new Command();
+    registerForgeCommand(program);
+    return program.parseAsync(['forge', ...args], { from: 'user' });
+  };
+
+  beforeEach(() => {
+    tmpDir = fs.mkdtempSync(path.join(os.tmpdir(), 'forge-cli-failclosed-'));
+    briefPath = path.join(tmpDir, 'brief.md');
+    fs.writeFileSync(briefPath, '# Fix bug\n\nA bugfix for lint cleanup.');
+    errSpy = vi.spyOn(console, 'error').mockImplementation(() => {});
+    logSpy = vi.spyOn(console, 'log').mockImplementation(() => {});
+    prevExitCode = process.exitCode;
+  });
+
+  afterEach(() => {
+    errSpy.mockRestore();
+    logSpy.mockRestore();
+    process.exitCode = prevExitCode;
+    fs.rmSync(tmpDir, { recursive: true, force: true });
+  });
+
+  it('exits nonzero with a typed FORGE_NO_EXECUTOR error when no executor is wired and --simulate is absent', async () => {
+    await parse(['run', '--brief', briefPath, '--codebase', tmpDir]);
+
+    expect(process.exitCode).toBe(1);
+    const errText = errSpy.mock.calls.map((c) => c.join(' ')).join('\n');
+    expect(errText).toContain('FORGE_NO_EXECUTOR');
+    // It must never run the pipeline with a stub and report success.
+    expect(fs.existsSync(path.join(tmpDir, '.forge', 'runs'))).toBe(false);
+  });
+
+  it('completes with typed simulated results and exit 0 under explicit --simulate', async () => {
+    await parse(['run', '--brief', briefPath, '--codebase', tmpDir, '--simulate']);
+
+    expect(process.exitCode).toBeUndefined();
+
+    // Loud simulated-mode summary.
+    const logText = logSpy.mock.calls.map((c) => c.join(' ')).join('\n');
+    expect(logText).toContain('SIMULATED');
+
+    // Manifest records the mode and simulated per-result statuses.
+    const runsDir = path.join(tmpDir, '.forge', 'runs');
+    const runIds = fs.readdirSync(runsDir);
+    expect(runIds).toHaveLength(1);
+    const manifest = loadManifest(path.join(runsDir, runIds[0]!));
+    expect(manifest.mode).toBe('simulated');
+    expect(manifest.status).toBe('simulated');
+    for (const stageStatus of Object.values(manifest.stages)) {
+      expect(stageStatus?.status).toBe('simulated');
+      for (const gateResult of stageStatus?.gateResults ?? []) {
+        expect(gateResult.outcome).toBe('simulated');
+      }
+    }
+  });
+
+  it('resume exits nonzero with a typed FORGE_NO_EXECUTOR error without --simulate', async () => {
+    const runDir = path.join(tmpDir, '.forge', 'runs', '20260101-000000');
+    fs.mkdirSync(runDir, { recursive: true });
+    fs.writeFileSync(
+      path.join(runDir, 'manifest.json'),
+      JSON.stringify({
+        runId: '20260101-000000',
+        brief: briefPath,
+        codebase: tmpDir,
+        briefClass: 'hotfix',
+        classSource: 'frontmatter',
+        forceBoard: false,
+        createdAt: '2026-01-01T00:00:00Z',
+        updatedAt: '2026-01-01T00:00:00Z',
+        currentStage: '00-intake',
+        status: 'in_progress',
+        stages: { '00-intake': { status: 'passed' } },
+      }),
+    );
+
+    await parse(['resume', '20260101-000000', '--project', tmpDir]);
+
+    expect(process.exitCode).toBe(1);
+    const errText = errSpy.mock.calls.map((c) => c.join(' ')).join('\n');
+    expect(errText).toContain('FORGE_NO_EXECUTOR');
+  });
+});

packages/forge/src/cli.ts

@@ -5,37 +5,47 @@ import type { Command } from 'commander';
 
 import { classifyBrief } from './brief-classifier.js';
 import { STAGE_LABELS, STAGE_SEQUENCE } from './constants.js';
+import { ForgeCapabilityError } from './errors.js';
 import { getEffectivePersonas, loadBoardPersonas } from './persona-loader.js';
 import { generateRunId, getPipelineStatus, loadManifest, runPipeline } from './pipeline-runner.js';
-import type { PipelineOptions, RunManifest, TaskExecutor } from './types.js';
-
-// ---------------------------------------------------------------------------
-// Stub executor — used when no real executor is wired at CLI invocation time.
-// ---------------------------------------------------------------------------
-
-const stubExecutor: TaskExecutor = {
-  async submitTask(task) {
-    console.log(`  [forge] stage submitted: ${task.id} (${task.title})`);
-  },
-  async waitForCompletion(taskId, _timeoutMs) {
-    console.log(`  [forge] stage complete:  ${taskId}`);
-    return {
-      task_id: taskId,
-      status: 'completed' as const,
-      completed_at: new Date().toISOString(),
-      exit_code: 0,
-      gate_results: [],
-    };
-  },
-  async getTaskStatus(_taskId) {
-    return 'completed' as const;
-  },
-};
+import { createSimulatedExecutor } from './simulated-executor.js';
+import type { PipelineOptions, RunManifest, RunMode } from './types.js';
 
 // ---------------------------------------------------------------------------
 // Helpers
 // ---------------------------------------------------------------------------
 
+/** Resolve a run's effective mode, defaulting legacy manifests to normal. */
+function runModeOf(manifest: RunManifest): RunMode {
+  return manifest.mode ?? 'normal';
+}
+
+/** Print a loud banner so a simulated run can never be misread as verified. */
+function printSimulatedBanner(): void {
+  console.log('');
+  console.log('[forge] ===============================================================');
+  console.log('[forge] MODE: SIMULATED — no stage or gate was really executed.');
+  console.log('[forge] All results are synthetic and MUST NOT be read as verified');
+  console.log('[forge] success. Wire a real executor/providers and re-run to verify.');
+  console.log('[forge] ===============================================================');
+}
+
+/** Print a typed error line for fail-closed capability errors. */
+function printCapabilityError(err: ForgeCapabilityError): void {
+  console.error(`[forge] error ${err.code}: ${err.message}`);
+  console.error(`[forge] missing capability: ${err.capability}`);
+}
+
+/** Handle a pipeline error uniformly: typed capability errors get their code. */
+function handlePipelineError(err: unknown): void {
+  if (err instanceof ForgeCapabilityError) {
+    printCapabilityError(err);
+  } else {
+    console.error(`[forge] pipeline failed: ${err instanceof Error ? err.message : String(err)}`);
+  }
+  process.exitCode = 1;
+}
+
 function formatDuration(startedAt?: string, completedAt?: string): string {
   if (!startedAt || !completedAt) return '-';
   const ms = new Date(completedAt).getTime() - new Date(startedAt).getTime();
@@ -44,19 +54,24 @@ function formatDuration(startedAt?: string, completedAt?: string): string {
 }
 
 function printManifestTable(manifest: RunManifest): void {
+  const mode = runModeOf(manifest);
   console.log(`\nRun ID : ${manifest.runId}`);
   console.log(`Status : ${manifest.status}`);
+  console.log(`Mode   : ${mode}`);
+  if (mode === 'simulated') {
+    console.log('WARNING: SIMULATED RUN — results are synthetic, not verified success.');
+  }
   console.log(`Brief  : ${manifest.brief}`);
   console.log(`Class  : ${manifest.briefClass} (${manifest.classSource})`);
   console.log(`Updated: ${manifest.updatedAt}`);
   console.log('');
-  console.log('Stage'.padEnd(22) + 'Status'.padEnd(14) + 'Duration');
-  console.log('-'.repeat(50));
+  console.log('Stage'.padEnd(22) + 'Status'.padEnd(24) + 'Duration');
+  console.log('-'.repeat(60));
   for (const stage of STAGE_SEQUENCE) {
     const s = manifest.stages[stage];
     if (!s) continue;
     const label = (STAGE_LABELS[stage] ?? stage).padEnd(22);
-    const status = s.status.padEnd(14);
+    const status = s.status.padEnd(24);
     const dur = formatDuration(s.startedAt, s.completedAt);
     console.log(`${label}${status}${dur}`);
   }
@@ -90,23 +105,58 @@ function listRecentRuns(projectRoot?: string): void {
   }
 
   console.log('\nRecent runs:');
-  console.log('Run ID'.padEnd(22) + 'Status'.padEnd(14) + 'Brief');
-  console.log('-'.repeat(70));
+  console.log('Run ID'.padEnd(22) + 'Status'.padEnd(24) + 'Mode'.padEnd(12) + 'Brief');
+  console.log('-'.repeat(80));
 
   for (const runId of entries) {
     const runDir = path.join(runsDir, runId);
     try {
       const manifest = loadManifest(runDir);
-      const status = manifest.status.padEnd(14);
+      const status = manifest.status.padEnd(24);
+      const mode = runModeOf(manifest).padEnd(12);
       const brief = path.basename(manifest.brief);
-      console.log(`${runId.padEnd(22)}${status}${brief}`);
+      console.log(`${runId.padEnd(22)}${status}${mode}${brief}`);
     } catch {
-      console.log(`${runId.padEnd(22)}${'(unreadable)'.padEnd(14)}`);
+      console.log(`${runId.padEnd(22)}${'(unreadable)'.padEnd(24)}`);
     }
   }
   console.log('');
 }
 
+/**
+ * Apply the exit-code policy for a finished pipeline run (SDLC-D-035):
+ *
+ * - exit 0 only for a verified `completed` normal run, or for an overall
+ *   `simulated` run when the caller explicitly passed --simulate;
+ * - anything else exits nonzero so it can never be read as success.
+ */
+function applyRunExitPolicy(result: { manifest: RunManifest; runDir: string }, simulate: boolean) {
+  const { manifest } = result;
+
+  if (runModeOf(manifest) === 'simulated') {
+    if (!simulate || manifest.status !== 'simulated') {
+      console.error(
+        '[forge] error FORGE_MODE_MISMATCH: run reports simulated results without an explicit, ' +
+          'consistent --simulate request; refusing to report success.',
+      );
+      process.exitCode = 1;
+      return;
+    }
+    printSimulatedBanner();
+    console.log(`[forge] run directory: ${result.runDir}`);
+    return; // exit 0 — the caller explicitly opted into simulation
+  }
+
+  if (manifest.status !== 'completed') {
+    console.error(`[forge] run did not complete: terminal status '${manifest.status}'`);
+    process.exitCode = 1;
+    return;
+  }
+
+  console.log(`[forge] pipeline complete (mode: normal): ${manifest.runId}`);
+  console.log(`[forge] run directory: ${result.runDir}`);
+}
+
 // ---------------------------------------------------------------------------
 // Register function
 // ---------------------------------------------------------------------------
@@ -129,6 +179,11 @@ export function registerForgeCommand(parent: Command): void {
     .option('--config <path>', 'Path to forge config file (.forge/config.yaml)')
     .option('--codebase <path>', 'Codebase root to pass to the pipeline', process.cwd())
     .option('--dry-run', 'Print planned stages without executing', false)
+    .option(
+      '--simulate',
+      'Simulate execution without real providers (every result is typed simulated, never verified)',
+      false,
+    )
     .action(
       async (opts: {
         brief: string;
@@ -137,6 +192,7 @@ export function registerForgeCommand(parent: Command): void {
         config?: string;
         codebase: string;
         dryRun: boolean;
+        simulate: boolean;
       }) => {
         const briefPath = path.resolve(opts.brief);
 
@@ -149,14 +205,22 @@ export function registerForgeCommand(parent: Command): void {
         const briefContent = fs.readFileSync(briefPath, 'utf-8');
         const briefClass = classifyBrief(briefContent);
         const projectRoot = opts.codebase;
+        // A real executor is never wired at CLI invocation time today, so the
+        // only executor we may construct is the explicitly-requested simulated
+        // one. Normal mode fails closed with FORGE_NO_EXECUTOR.
+        const executor = opts.simulate ? createSimulatedExecutor() : undefined;
 
         if (opts.resume) {
           const runId = opts.runId ?? generateRunId();
           const runDir = resolveRunDir(runId, projectRoot);
           console.log(`[forge] resuming run: ${runId}`);
-          const { resumePipeline } = await import('./pipeline-runner.js');
-          const result = await resumePipeline(runDir, stubExecutor);
-          console.log(`[forge] pipeline complete: ${result.runId}`);
+          try {
+            const { resumePipeline } = await import('./pipeline-runner.js');
+            const result = await resumePipeline(runDir, executor, { simulate: opts.simulate });
+            applyRunExitPolicy(result, opts.simulate);
+          } catch (err) {
+            handlePipelineError(err);
+          }
           return;
         }
 
@@ -164,7 +228,8 @@ export function registerForgeCommand(parent: Command): void {
           briefClass,
           codebase: projectRoot,
           dryRun: opts.dryRun,
-          executor: stubExecutor,
+          executor,
+          simulate: opts.simulate,
         };
 
         if (opts.dryRun) {
@@ -180,16 +245,15 @@ export function registerForgeCommand(parent: Command): void {
 
         console.log(`[forge] starting pipeline for brief: ${briefPath}`);
         console.log(`[forge] classified as: ${briefClass}`);
+        if (opts.simulate) {
+          console.log('[forge] mode: SIMULATED (explicit --simulate)');
+        }
 
         try {
           const result = await runPipeline(briefPath, projectRoot, pipelineOptions);
-          console.log(`[forge] pipeline complete: ${result.runId}`);
-          console.log(`[forge] run directory: ${result.runDir}`);
+          applyRunExitPolicy(result, opts.simulate);
         } catch (err) {
-          console.error(
-            `[forge] pipeline failed: ${err instanceof Error ? err.message : String(err)}`,
-          );
-          process.exitCode = 1;
+          handlePipelineError(err);
         }
       },
     );
@@ -224,7 +288,12 @@ export function registerForgeCommand(parent: Command): void {
     .command('resume <runId>')
     .description('Resume a stopped or failed pipeline run')
     .option('--project <path>', 'Project root (defaults to cwd)', process.cwd())
-    .action(async (runId: string, opts: { project: string }) => {
+    .option(
+      '--simulate',
+      'Simulate execution without real providers (every result is typed simulated, never verified)',
+      false,
+    )
+    .action(async (runId: string, opts: { project: string; simulate: boolean }) => {
       const runDir = resolveRunDir(runId, opts.project);
 
       if (!fs.existsSync(runDir)) {
@@ -234,15 +303,20 @@ export function registerForgeCommand(parent: Command): void {
       }
 
       console.log(`[forge] resuming run: ${runId}`);
+      if (opts.simulate) {
+        console.log('[forge] mode: SIMULATED (explicit --simulate)');
+      }
+
+      // No real executor is wired at CLI invocation time; only the explicitly
+      // requested simulated executor may be constructed (fail closed otherwise).
+      const executor = opts.simulate ? createSimulatedExecutor() : undefined;
 
       try {
         const { resumePipeline } = await import('./pipeline-runner.js');
-        const result = await resumePipeline(runDir, stubExecutor);
-        console.log(`[forge] pipeline complete: ${result.runId}`);
-        console.log(`[forge] run directory: ${result.runDir}`);
+        const result = await resumePipeline(runDir, executor, { simulate: opts.simulate });
+        applyRunExitPolicy(result, opts.simulate);
       } catch (err) {
-        console.error(`[forge] resume failed: ${err instanceof Error ? err.message : String(err)}`);
-        process.exitCode = 1;
+        handlePipelineError(err);
       }
     });
 

packages/forge/src/constants.ts

@@ -9,7 +9,16 @@ export const PACKAGE_ROOT = path.resolve(path.dirname(fileURLToPath(import.meta.
 /** Pipeline asset directory (stages, agents, rails, gates, templates). */
 export const PIPELINE_DIR = path.join(PACKAGE_ROOT, 'pipeline');
 
-/** Stage specifications — defines every pipeline stage. */
+/** Stage specifications — defines every pipeline stage.
+ *\n * Gate semantics (SDLC-D-035): every gate is one of
+ * - a real command string / GateEntry a mechanical runner can execute,
+ * - an `authority` gate (human/board sign-off; produces waiting-for-authority),
+ * - a `provider` gate (requires a wired provider such as a reviewer or CI pipeline).
+ *
+ * Vacuous gates (`true`, echo'd synthetic approvals, placeholder ci-pipeline
+ * commands) are forbidden: a stage whose gate has no real implementation
+ * fails closed instead of passing.
+ */
 export const STAGE_SPECS: Record<string, StageSpec> = {
   '00-intake': {
     number: '00',
@@ -27,7 +36,13 @@ export const STAGE_SPECS: Record<string, StageSpec> = {
     type: 'research',
     gate: 'discovery-complete',
     promptFile: '00b-discovery.md',
-    qualityGates: ['true'],
+    qualityGates: [
+      {
+        kind: 'authority',
+        capability: 'discovery-complete',
+        reason: 'discovery completion is attested by an authority; no mechanical check exists',
+      },
+    ],
   },
   '01-board': {
     number: '01',
@@ -36,7 +51,13 @@ export const STAGE_SPECS: Record<string, StageSpec> = {
     type: 'review',
     gate: 'board-approval',
     promptFile: '01-board.md',
-    qualityGates: [{ type: 'ci-pipeline', command: 'board-approval (via board-tasks)' }],
+    qualityGates: [
+      {
+        kind: 'authority',
+        capability: 'board-approval',
+        reason: 'board approval is a board/human decision; no mechanical gate exists',
+      },
+    ],
   },
   '01b-brief-analyzer': {
     number: '01b',
@@ -45,7 +66,13 @@ export const STAGE_SPECS: Record<string, StageSpec> = {
     type: 'research',
     gate: 'brief-analysis-complete',
     promptFile: '01-board.md',
-    qualityGates: ['true'],
+    qualityGates: [
+      {
+        kind: 'authority',
+        capability: 'brief-analysis-complete',
+        reason: 'brief analysis completion is attested by an authority; no mechanical check exists',
+      },
+    ],
   },
   '02-planning-1': {
     number: '02',
@@ -54,7 +81,13 @@ export const STAGE_SPECS: Record<string, StageSpec> = {
     type: 'research',
     gate: 'architecture-approval',
     promptFile: '02-planning-1-architecture.md',
-    qualityGates: ['true'],
+    qualityGates: [
+      {
+        kind: 'authority',
+        capability: 'architecture-approval',
+        reason: 'ADR approval requires authority sign-off; no mechanical check exists',
+      },
+    ],
   },
   '03-planning-2': {
     number: '03',
@@ -63,7 +96,14 @@ export const STAGE_SPECS: Record<string, StageSpec> = {
     type: 'research',
     gate: 'implementation-approval',
     promptFile: '03-planning-2-implementation.md',
-    qualityGates: ['true'],
+    qualityGates: [
+      {
+        kind: 'authority',
+        capability: 'implementation-approval',
+        reason:
+          'implementation spec approval requires authority sign-off; no mechanical check exists',
+      },
+    ],
   },
   '04-planning-3': {
     number: '04',
@@ -72,7 +112,14 @@ export const STAGE_SPECS: Record<string, StageSpec> = {
     type: 'research',
     gate: 'decomposition-approval',
     promptFile: '04-planning-3-decomposition.md',
-    qualityGates: ['true'],
+    qualityGates: [
+      {
+        kind: 'authority',
+        capability: 'decomposition-approval',
+        reason:
+          'task decomposition approval requires authority sign-off; no mechanical check exists',
+      },
+    ],
   },
   '05-coding': {
     number: '05',
@@ -92,9 +139,10 @@ export const STAGE_SPECS: Record<string, StageSpec> = {
     promptFile: '06-review.md',
     qualityGates: [
       {
-        type: 'ai-review',
-        command:
-          'echo \'{"summary":"review-pass","verdict":"approve","findings":[],"stats":{"blockers":0,"should_fix":0,"suggestions":0}}\'',
+        kind: 'provider',
+        capability: 'reviewer',
+        reason:
+          'review verdicts require a wired reviewer provider; synthetic approvals are not permitted',
       },
     ],
   },
@@ -105,7 +153,13 @@ export const STAGE_SPECS: Record<string, StageSpec> = {
     type: 'coding',
     gate: 're-review',
     promptFile: '07-remediate.md',
-    qualityGates: ['true'],
+    qualityGates: [
+      {
+        kind: 'authority',
+        capability: 're-review',
+        reason: 'remediation re-review is an approval-based gate; no mechanical check exists',
+      },
+    ],
   },
   '08-test': {
     number: '08',
@@ -123,7 +177,13 @@ export const STAGE_SPECS: Record<string, StageSpec> = {
     type: 'deploy',
     gate: 'deploy-verification',
     promptFile: '09-deploy.md',
-    qualityGates: [{ type: 'ci-pipeline', command: 'deploy-verification' }],
+    qualityGates: [
+      {
+        kind: 'provider',
+        capability: 'ci-pipeline',
+        reason: 'deploy verification requires a wired CI pipeline provider',
+      },
+    ],
   },
 };
 

packages/forge/src/errors.ts

@@ -0,0 +1,46 @@
+/**
+ * Typed fail-closed capability errors (SDLC-D-035).
+ *
+ * A Forge run must fail closed when a required capability (executor, reviewer
+ * provider, CI pipeline, authority sign-off) is missing. These typed errors
+ * name the missing capability so callers can distinguish "not wired" from
+ * ordinary execution failures.
+ */
+
+/** Closed set of typed Forge capability error codes. */
+export const FORGE_ERROR_CODES = [
+  'FORGE_NO_EXECUTOR',
+  'FORGE_NO_REVIEWER',
+  'FORGE_NO_CI_PIPELINE',
+  'FORGE_NO_PROVIDER',
+  'FORGE_AUTHORITY_REQUIRED',
+] as const;
+
+export type ForgeErrorCode = (typeof FORGE_ERROR_CODES)[number];
+
+/** Raised when a required capability is missing and the pipeline must fail closed. */
+export class ForgeCapabilityError extends Error {
+  /** Typed error code from the closed FORGE_ERROR_CODES set. */
+  readonly code: ForgeErrorCode;
+  /** The missing capability, e.g. `task-executor`, `reviewer`, `board-approval`. */
+  readonly capability: string;
+
+  constructor(code: ForgeErrorCode, capability: string, message: string) {
+    super(message);
+    this.name = 'ForgeCapabilityError';
+    this.code = code;
+    this.capability = capability;
+  }
+}
+
+/** Map a provider gate capability to its typed error code. */
+export function providerErrorCode(capability: string): ForgeErrorCode {
+  switch (capability) {
+    case 'reviewer':
+      return 'FORGE_NO_REVIEWER';
+    case 'ci-pipeline':
+      return 'FORGE_NO_CI_PIPELINE';
+    default:
+      return 'FORGE_NO_PROVIDER';
+  }
+}

packages/forge/src/index.ts

@@ -5,6 +5,13 @@ export type {
   StageSpec,
   BriefClass,
   ClassSource,
+  ForgeOutcome,
+  AuthorityGate,
+  ProviderGate,
+  ForgeGate,
+  ForgeGateResult,
+  ForgeTaskResult,
+  RunMode,
   StageStatus,
   RunManifest,
   ForgeTaskStatus,
@@ -81,5 +88,24 @@ export {
   getPipelineStatus,
 } from './pipeline-runner.js';
 
+// Fail-closed errors and typed outcome model (SDLC-D-035)
+export { FORGE_ERROR_CODES, ForgeCapabilityError, providerErrorCode } from './errors.js';
+export type { ForgeErrorCode } from './errors.js';
+export {
+  isSatisfyingOutcome,
+  isCapabilityGate,
+  isCommandGate,
+  gateLabel,
+  uniformGateResults,
+  simulatedGateResults,
+  waitingGateResults,
+  blockedGateResults,
+  evaluateStageGates,
+} from './outcomes.js';
+export type { StageEvaluation } from './outcomes.js';
+
+// Simulated executor (explicit --simulate only)
+export { createSimulatedExecutor } from './simulated-executor.js';
+
 // CLI
 export { registerForgeCommand } from './cli.js';

packages/forge/src/outcomes.ts

@@ -0,0 +1,147 @@
+import type { GateEntry } from '@mosaicstack/macp';
+
+import type {
+  AuthorityGate,
+  ForgeGate,
+  ForgeGateResult,
+  ForgeOutcome,
+  ForgeTaskResult,
+  ProviderGate,
+} from './types.js';
+
+/**
+ * Gate and dependency satisfaction predicate (SDLC-D-035).
+ *
+ * ONLY a verified `passed` outcome satisfies. Every other member of the closed
+ * outcome set — including `simulated` — is non-satisfying, so a simulated or
+ * authority-blocked result can never be read as success-by-verification.
+ */
+export function isSatisfyingOutcome(outcome: ForgeOutcome): boolean {
+  return outcome === 'passed';
+}
+
+/** Whether a gate is an authority or provider gate (capability-based, command-less). */
+export function isCapabilityGate(gate: ForgeGate): gate is AuthorityGate | ProviderGate {
+  if (typeof gate !== 'object' || gate === null) return false;
+  const kind = (gate as Record<string, unknown>)['kind'];
+  return kind === 'authority' || kind === 'provider';
+}
+
+/** Whether a gate definition carries a real command a mechanical runner can execute. */
+export function isCommandGate(gate: ForgeGate): gate is string | GateEntry {
+  if (typeof gate === 'string') {
+    return gate.trim().length > 0;
+  }
+  if (isCapabilityGate(gate)) {
+    // Authority and provider gates are satisfied by a capability, not a command.
+    return false;
+  }
+  return typeof gate.command === 'string' && gate.command.trim().length > 0;
+}
+
+/** Typed label identifying a gate in results and logs. */
+export function gateLabel(gate: ForgeGate): string {
+  if (typeof gate === 'string') return gate;
+  if (isCapabilityGate(gate)) return `${gate.kind}:${gate.capability}`;
+  return gate.command || gate.type || 'unnamed-gate';
+}
+
+/** Reason string stamped on every simulated gate result. */
+export const SIMULATED_GATE_REASON =
+  'simulated execution (--simulate): gate was not evaluated by a real implementation';
+
+/** Build typed gate results with a uniform outcome for a stage's declared gates. */
+export function uniformGateResults(
+  gates: ForgeGate[],
+  outcome: ForgeOutcome,
+  reason: string,
+): ForgeGateResult[] {
+  return gates.map((gate) => ({ gate: gateLabel(gate), outcome, reason }));
+}
+
+/** Typed simulated gate results — used exclusively in `--simulate` runs. */
+export function simulatedGateResults(gates: ForgeGate[]): ForgeGateResult[] {
+  return uniformGateResults(gates, 'simulated', SIMULATED_GATE_REASON);
+}
+
+/** Typed waiting-for-authority gate results for approval-based stages. */
+export function waitingGateResults(gates: ForgeGate[], reason: string): ForgeGateResult[] {
+  return uniformGateResults(gates, 'waiting-for-authority', reason);
+}
+
+/** Typed blocked gate results for stages whose provider capability is not wired. */
+export function blockedGateResults(gates: ForgeGate[], reason: string): ForgeGateResult[] {
+  return uniformGateResults(gates, 'blocked', reason);
+}
+
+/** Outcome of evaluating a completed stage in normal mode. */
+export interface StageEvaluation {
+  outcome: ForgeOutcome;
+  reason: string;
+  gateResults: ForgeGateResult[];
+}
+
+/**
+ * Evaluate a stage's declared gates against the executor's typed result.
+ *
+ * Fail-closed mapping:
+ * - a `simulated` task or gate outcome in normal mode maps to `error`
+ * - a missing gate result for a required command gate maps to `blocked`
+ * - a non-passing task outcome propagates as the stage outcome
+ * - only verified `passed` task and gate outcomes yield a `passed` stage
+ */
+export function evaluateStageGates(
+  stageName: string,
+  gates: ForgeGate[],
+  result: ForgeTaskResult,
+): StageEvaluation {
+  const gateResults = result.gate_results ?? [];
+
+  if (result.outcome === 'simulated') {
+    return {
+      outcome: 'error',
+      reason: `executor reported a simulated outcome for stage '${stageName}' in normal mode — refusing to treat simulated results as verified`,
+      gateResults,
+    };
+  }
+
+  if (!isSatisfyingOutcome(result.outcome)) {
+    return {
+      outcome: result.outcome,
+      reason: `task outcome is '${result.outcome}': ${result.reason}`,
+      gateResults,
+    };
+  }
+
+  for (const gate of gates) {
+    // Authority and provider gates are pre-flighted before execution; they have
+    // no mechanical result to verify here.
+    if (!isCommandGate(gate)) continue;
+
+    const label = gateLabel(gate);
+    const gateResult = gateResults.find((r) => r.gate === label);
+    if (!gateResult) {
+      return {
+        outcome: 'blocked',
+        reason: `no gate result was reported for required gate '${label}' (stage '${stageName}')`,
+        gateResults,
+      };
+    }
+    if (!isSatisfyingOutcome(gateResult.outcome)) {
+      return {
+        outcome: gateResult.outcome === 'simulated' ? 'error' : gateResult.outcome,
+        reason: `gate '${label}' outcome is '${gateResult.outcome}': ${gateResult.reason}`,
+        gateResults,
+      };
+    }
+  }
+
+  return {
+    outcome: 'passed',
+    reason:
+      gates.length === 0
+        ? "stage declares no gates; task outcome 'passed' accepted"
+        : 'all declared gates verified passed',
+    gateResults,
+  };
+}

packages/forge/src/pipeline-runner.ts

@@ -1,18 +1,33 @@
 import fs from 'node:fs';
 import path from 'node:path';
 
-import { STAGE_SEQUENCE } from './constants.js';
+import { STAGE_SEQUENCE, STAGE_SPECS } from './constants.js';
 import { determineBriefClass, stagesForClass } from './brief-classifier.js';
+import { ForgeCapabilityError, providerErrorCode } from './errors.js';
+import {
+  blockedGateResults,
+  evaluateStageGates,
+  isCapabilityGate,
+  simulatedGateResults,
+  waitingGateResults,
+} from './outcomes.js';
 import { mapStageToTask } from './stage-adapter.js';
+import { createSimulatedExecutor } from './simulated-executor.js';
 import type {
   ForgeTask,
+  ForgeTaskResult,
   PipelineOptions,
   PipelineResult,
   RunManifest,
+  RunMode,
   StageStatus,
   TaskExecutor,
 } from './types.js';
 
+/** Reason stamped on stages that complete under explicit simulation. */
+const SIMULATED_STAGE_REASON =
+  'simulated execution (--simulate): stage was not executed by a real executor';
+
 /**
  * Generate a timestamp-based run ID.
  */
@@ -47,6 +62,7 @@ function createManifest(opts: {
   briefClass: RunManifest['briefClass'];
   classSource: RunManifest['classSource'];
   forceBoard: boolean;
+  mode: RunMode;
   runDir: string;
 }): RunManifest {
   const ts = nowISO();
@@ -57,6 +73,7 @@ function createManifest(opts: {
     briefClass: opts.briefClass,
     classSource: opts.classSource,
     forceBoard: opts.forceBoard,
+    mode: opts.mode,
     createdAt: ts,
     updatedAt: ts,
     currentStage: '',
@@ -108,20 +125,199 @@ export function selectStages(stages?: string[], skipTo?: string): string[] {
   return selected.slice(skipIndex);
 }
 
+/**
+ * Fail closed when the required executor capability is missing (SDLC-D-035).
+ */
+function requireExecutor(executor: TaskExecutor | undefined, simulate: boolean): TaskExecutor {
+  if (executor) return executor;
+  if (simulate) return createSimulatedExecutor({ log: false });
+  throw new ForgeCapabilityError(
+    'FORGE_NO_EXECUTOR',
+    'task-executor',
+    'no task executor is wired; refusing to run the pipeline with a stub executor (fail closed). ' +
+      'Pass --simulate to opt into explicitly simulated execution.',
+  );
+}
+
+/**
+ * Pre-flight a stage's gates in normal mode (fail closed, SDLC-D-035).
+ *
+ * - authority gates: record a typed `waiting-for-authority` stage result and
+ *   raise FORGE_AUTHORITY_REQUIRED — approval-based gates never pass vacuously.
+ * - provider gates: record a typed `blocked` stage result and raise the typed
+ *   capability error for the missing provider.
+ *
+ * Returns the stage status to record when the pre-flight blocks, or undefined
+ * when the stage may proceed.
+ */
+function preflightStageGates(
+  stageName: string,
+  manifest: RunManifest,
+): { status: StageStatus; error: ForgeCapabilityError } | undefined {
+  const spec = STAGE_SPECS[stageName];
+  if (!spec) throw new Error(`Unknown Forge stage: ${stageName}`);
+
+  for (const gate of spec.qualityGates) {
+    if (!isCapabilityGate(gate)) continue;
+
+    const startedAt = manifest.stages[stageName]?.startedAt;
+    const completedAt = nowISO();
+
+    if (gate.kind === 'authority') {
+      const reason = `gate '${gate.capability}' requires authority sign-off; no mechanical implementation exists (${gate.reason})`;
+      return {
+        status: {
+          status: 'waiting-for-authority',
+          reason,
+          startedAt,
+          completedAt,
+          gateResults: waitingGateResults(spec.qualityGates, reason),
+        },
+        error: new ForgeCapabilityError(
+          'FORGE_AUTHORITY_REQUIRED',
+          gate.capability,
+          `stage '${stageName}' is blocked on authority gate '${gate.capability}': ${gate.reason}. ` +
+            'The pipeline fails closed instead of passing vacuously. Record the approval out-of-band ' +
+            'or run with --simulate for explicitly simulated execution.',
+        ),
+      };
+    }
+
+    const reason = `gate '${gate.capability}' requires provider '${gate.capability}' and none is wired (${gate.reason})`;
+    return {
+      status: {
+        status: 'blocked',
+        reason,
+        startedAt,
+        completedAt,
+        gateResults: blockedGateResults(spec.qualityGates, reason),
+      },
+      error: new ForgeCapabilityError(
+        providerErrorCode(gate.capability),
+        gate.capability,
+        `stage '${stageName}' requires provider '${gate.capability}' which is not wired: ${gate.reason}. ` +
+          'The pipeline fails closed instead of passing vacuously.',
+      ),
+    };
+  }
+
+  return undefined;
+}
+
+/**
+ * Execute the given stage tasks sequentially, updating the manifest.
+ *
+ * Normal mode requires a real executor and evaluates every declared command
+ * gate through the typed outcome model; any non-verified result fails closed.
+ * Simulate mode types every stage and gate result as `simulated`.
+ */
+async function executeStages(opts: {
+  manifest: RunManifest;
+  runDir: string;
+  tasks: ForgeTask[];
+  stageNames: string[];
+  executor: TaskExecutor;
+  simulate: boolean;
+}): Promise<void> {
+  const { manifest, runDir, tasks, stageNames, executor, simulate } = opts;
+
+  for (let i = 0; i < tasks.length; i++) {
+    const task = tasks[i]!;
+    const stageName = stageNames[i]!;
+    const spec = STAGE_SPECS[stageName];
+    if (!spec) throw new Error(`Unknown Forge stage: ${stageName}`);
+
+    // Update manifest: stage in progress
+    manifest.currentStage = stageName;
+    manifest.stages[stageName] = {
+      status: 'in_progress',
+      startedAt: nowISO(),
+    };
+    saveManifest(runDir, manifest);
+
+    // Fail-closed pre-flight (normal mode only): authority/provider gates have
+    // no mechanical implementation and must never pass vacuously.
+    if (!simulate) {
+      const blocked = preflightStageGates(stageName, manifest);
+      if (blocked) {
+        manifest.stages[stageName] = blocked.status;
+        manifest.status =
+          blocked.status.status === 'waiting-for-authority' ? 'waiting-for-authority' : 'failed';
+        saveManifest(runDir, manifest);
+        throw blocked.error;
+      }
+    }
+
+    let result: ForgeTaskResult;
+    try {
+      await executor.submitTask(task);
+      result = await executor.waitForCompletion(task.id, task.timeoutSeconds * 1000);
+    } catch (error) {
+      // Process errors (including timeouts) map to the fail-closed `error` outcome.
+      const reason = error instanceof Error ? error.message : String(error);
+      manifest.stages[stageName] = {
+        status: 'error',
+        reason: `executor error: ${reason}`,
+        startedAt: manifest.stages[stageName]?.startedAt,
+        completedAt: nowISO(),
+        gateResults: [],
+      };
+      manifest.status = 'failed';
+      saveManifest(runDir, manifest);
+      throw error instanceof Error ? error : new Error(reason);
+    }
+
+    if (simulate) {
+      manifest.stages[stageName] = {
+        status: 'simulated',
+        reason: SIMULATED_STAGE_REASON,
+        startedAt: manifest.stages[stageName]?.startedAt,
+        completedAt: nowISO(),
+        gateResults: simulatedGateResults(spec.qualityGates),
+      };
+      saveManifest(runDir, manifest);
+      continue;
+    }
+
+    const evaluation = evaluateStageGates(stageName, spec.qualityGates, result);
+    manifest.stages[stageName] = {
+      status: evaluation.outcome,
+      reason: evaluation.reason,
+      startedAt: manifest.stages[stageName]?.startedAt,
+      completedAt: nowISO(),
+      gateResults: evaluation.gateResults,
+    };
+
+    if (evaluation.outcome !== 'passed') {
+      manifest.status =
+        evaluation.outcome === 'waiting-for-authority' ? 'waiting-for-authority' : 'failed';
+      saveManifest(runDir, manifest);
+      throw new Error(`Stage ${stageName} ${evaluation.outcome}: ${evaluation.reason}`);
+    }
+
+    saveManifest(runDir, manifest);
+  }
+}
+
 /**
  * Run the Forge pipeline.
  *
- * 1. Classify the brief
- * 2. Generate a run ID and create run directory
- * 3. Map stages to tasks and submit to TaskExecutor
- * 4. Track manifest with stage statuses
- * 5. Return pipeline result
+ * 1. Fail closed unless a real executor is wired or simulation is explicit
+ * 2. Classify the brief
+ * 3. Generate a run ID and create run directory
+ * 4. Map stages to tasks and submit to TaskExecutor
+ * 5. Track manifest with typed stage outcomes
+ * 6. Return pipeline result
  */
 export async function runPipeline(
   briefPath: string,
   projectRoot: string,
   options: PipelineOptions,
 ): Promise<PipelineResult> {
+  const simulate = options.simulate ?? false;
+  const executor = requireExecutor(options.executor, simulate);
+  const mode: RunMode = simulate ? 'simulated' : 'normal';
+
   const resolvedRoot = path.resolve(projectRoot);
   const resolvedBrief = path.resolve(briefPath);
   const briefContent = fs.readFileSync(resolvedBrief, 'utf-8');
@@ -146,6 +342,7 @@ export async function runPipeline(
     briefClass,
     classSource,
     forceBoard: options.forceBoard ?? false,
+    mode,
     runDir,
   });
 
@@ -172,54 +369,10 @@ export async function runPipeline(
   }
 
   // Execute stages
-  const { executor } = options;
-  for (let i = 0; i < tasks.length; i++) {
-    const task = tasks[i]!;
-    const stageName = selectedStages[i]!;
+  await executeStages({ manifest, runDir, tasks, stageNames: selectedStages, executor, simulate });
 
-    // Update manifest: stage in progress
-    manifest.currentStage = stageName;
-    manifest.stages[stageName] = {
-      status: 'in_progress',
-      startedAt: nowISO(),
-    };
-    saveManifest(runDir, manifest);
-
-    try {
-      await executor.submitTask(task);
-      const result = await executor.waitForCompletion(task.id, task.timeoutSeconds * 1000);
-
-      // Update manifest: stage completed or failed
-      const stageStatus: StageStatus = {
-        status: result.status === 'completed' ? 'passed' : 'failed',
-        startedAt: manifest.stages[stageName]!.startedAt,
-        completedAt: nowISO(),
-      };
-      manifest.stages[stageName] = stageStatus;
-
-      if (result.status !== 'completed') {
-        manifest.status = 'failed';
-        saveManifest(runDir, manifest);
-        throw new Error(`Stage ${stageName} failed with status: ${result.status}`);
-      }
-
-      saveManifest(runDir, manifest);
-    } catch (error) {
-      if (!manifest.stages[stageName]?.completedAt) {
-        manifest.stages[stageName] = {
-          status: 'failed',
-          startedAt: manifest.stages[stageName]?.startedAt,
-          completedAt: nowISO(),
-        };
-      }
-      manifest.status = 'failed';
-      saveManifest(runDir, manifest);
-      throw error;
-    }
-  }
-
-  // All stages passed
-  manifest.status = 'completed';
+  // All stages reached a terminal state for this mode
+  manifest.status = simulate ? 'simulated' : 'completed';
   saveManifest(runDir, manifest);
 
   return {
@@ -234,22 +387,30 @@ export async function runPipeline(
 }
 
 /**
- * Resume a pipeline from the last incomplete stage.
+ * Resume a pipeline from the last non-passed stage.
  */
 export async function resumePipeline(
   runDir: string,
-  executor: TaskExecutor,
+  executor?: TaskExecutor,
+  options?: { simulate?: boolean },
 ): Promise<PipelineResult> {
+  const simulate = options?.simulate ?? false;
+  const wiredExecutor = requireExecutor(executor, simulate);
+  const mode: RunMode = simulate ? 'simulated' : 'normal';
+
   const manifest = loadManifest(runDir);
   const resolvedRoot = path.dirname(path.dirname(path.dirname(runDir))); // .forge/runs/{id} → project root
 
   const briefContent = fs.readFileSync(manifest.brief, 'utf-8');
   const allStages = stagesForClass(manifest.briefClass, manifest.forceBoard);
 
-  // Find first non-passed stage
+  manifest.mode = mode;
+
+  // Find first non-satisfying stage (only a verified `passed` counts as done;
+  // simulated and waiting-for-authority stages are re-run).
   const resumeFrom = allStages.find((s) => manifest.stages[s]?.status !== 'passed');
   if (!resumeFrom) {
-    manifest.status = 'completed';
+    manifest.status = mode === 'simulated' ? 'simulated' : 'completed';
     saveManifest(runDir, manifest);
     return {
       runId: manifest.runId,
@@ -284,49 +445,16 @@ export async function resumePipeline(
     tasks.push(task);
   }
 
-  for (let i = 0; i < tasks.length; i++) {
-    const task = tasks[i]!;
-    const stageName = remainingStages[i]!;
+  await executeStages({
+    manifest,
+    runDir,
+    tasks,
+    stageNames: remainingStages,
+    executor: wiredExecutor,
+    simulate,
+  });
 
-    manifest.currentStage = stageName;
-    manifest.stages[stageName] = {
-      status: 'in_progress',
-      startedAt: nowISO(),
-    };
-    saveManifest(runDir, manifest);
-
-    try {
-      await executor.submitTask(task);
-      const result = await executor.waitForCompletion(task.id, task.timeoutSeconds * 1000);
-
-      manifest.stages[stageName] = {
-        status: result.status === 'completed' ? 'passed' : 'failed',
-        startedAt: manifest.stages[stageName]!.startedAt,
-        completedAt: nowISO(),
-      };
-
-      if (result.status !== 'completed') {
-        manifest.status = 'failed';
-        saveManifest(runDir, manifest);
-        throw new Error(`Stage ${stageName} failed with status: ${result.status}`);
-      }
-
-      saveManifest(runDir, manifest);
-    } catch (error) {
-      if (!manifest.stages[stageName]?.completedAt) {
-        manifest.stages[stageName] = {
-          status: 'failed',
-          startedAt: manifest.stages[stageName]?.startedAt,
-          completedAt: nowISO(),
-        };
-      }
-      manifest.status = 'failed';
-      saveManifest(runDir, manifest);
-      throw error;
-    }
-  }
-
-  manifest.status = 'completed';
+  manifest.status = simulate ? 'simulated' : 'completed';
   saveManifest(runDir, manifest);
 
   return {

packages/forge/src/simulated-executor.ts

@@ -0,0 +1,32 @@
+import type { ForgeTask, ForgeTaskResult, TaskExecutor } from './types.js';
+
+/**
+ * Simulated executor — used ONLY when the caller explicitly passes --simulate.
+ *
+ * It submits no real work and returns typed `simulated` results so a simulated
+ * run can never be confused with a verified one. In normal mode (no --simulate)
+ * the CLI refuses to run at all with FORGE_NO_EXECUTOR instead of wiring this
+ * stub (fail closed, SDLC-D-035).
+ */
+export function createSimulatedExecutor(options?: { log?: boolean }): TaskExecutor {
+  const log = options?.log ?? true;
+  return {
+    async submitTask(task: ForgeTask) {
+      if (log) console.log(`  [forge:simulated] stage submitted: ${task.id} (${task.title})`);
+    },
+    async waitForCompletion(taskId: string): Promise<ForgeTaskResult> {
+      if (log) console.log(`  [forge:simulated] stage complete:  ${taskId}`);
+      return {
+        task_id: taskId,
+        outcome: 'simulated',
+        reason: 'no executor wired; simulated execution requested via --simulate',
+        completed_at: new Date().toISOString(),
+        exit_code: 0,
+        gate_results: [],
+      };
+    },
+    async getTaskStatus() {
+      return 'completed' as const;
+    },
+  };
+}

packages/forge/src/types.ts

@@ -1,4 +1,4 @@
-import type { GateEntry, TaskResult } from '@mosaicstack/macp';
+import type { GateEntry } from '@mosaicstack/macp';
 
 /** Stage dispatch mode. */
 export type StageDispatch = 'exec' | 'yolo' | 'pi';
@@ -6,6 +6,58 @@ export type StageDispatch = 'exec' | 'yolo' | 'pi';
 /** Stage type — determines agent selection and gate requirements. */
 export type StageType = 'research' | 'review' | 'coding' | 'deploy';
 
+/**
+ * Typed outcome for every gate and stage evaluation — closed set (SDLC-D-035).
+ *
+ * Only `passed` means "verified by a real implementation". `simulated` is
+ * produced exclusively in explicit `--simulate` runs and is never satisfying.
+ */
+export type ForgeOutcome =
+  | 'passed'
+  | 'failed'
+  | 'blocked'
+  | 'error'
+  | 'waiting-for-authority'
+  | 'simulated'
+  | 'not-applicable';
+
+/** A gate that requires authority (human/board) sign-off; no mechanical command can satisfy it. */
+export interface AuthorityGate {
+  kind: 'authority';
+  capability: string;
+  reason: string;
+}
+
+/** A gate that requires a wired provider (e.g. an AI reviewer, CI pipeline) to evaluate. */
+export interface ProviderGate {
+  kind: 'provider';
+  capability: string;
+  reason: string;
+}
+
+/** Forge quality gate: a real command, an authority sign-off, or a provider-backed check. */
+export type ForgeGate = string | GateEntry | AuthorityGate | ProviderGate;
+
+/** Typed result of evaluating a single quality gate. */
+export interface ForgeGateResult {
+  gate: string;
+  outcome: ForgeOutcome;
+  reason: string;
+  exitCode?: number;
+  output?: string;
+  timedOut?: boolean;
+}
+
+/** Typed result of a task/stage execution returned by a TaskExecutor. */
+export interface ForgeTaskResult {
+  task_id: string;
+  outcome: ForgeOutcome;
+  reason: string;
+  completed_at: string;
+  exit_code: number;
+  gate_results: ForgeGateResult[];
+}
+
 /** Stage specification — defines a single pipeline stage. */
 export interface StageSpec {
   number: string;
@@ -14,7 +66,7 @@ export interface StageSpec {
   type: StageType;
   gate: string;
   promptFile: string;
-  qualityGates: (string | GateEntry)[];
+  qualityGates: ForgeGate[];
 }
 
 /** Brief classification. */
@@ -25,11 +77,18 @@ export type ClassSource = 'cli' | 'frontmatter' | 'auto';
 
 /** Per-stage status within a run manifest. */
 export interface StageStatus {
-  status: 'pending' | 'in_progress' | 'passed' | 'failed';
+  status: 'pending' | 'in_progress' | ForgeOutcome;
+  /** Why the stage reached its current (terminal) outcome, when applicable. */
+  reason?: string;
   startedAt?: string;
   completedAt?: string;
+  /** Typed per-gate results recorded alongside the stage outcome. */
+  gateResults?: ForgeGateResult[];
 }
 
+/** Execution mode of a run. */
+export type RunMode = 'normal' | 'simulated';
+
 /** Run manifest — persisted to disk as manifest.json. */
 export interface RunManifest {
   runId: string;
@@ -38,10 +97,23 @@ export interface RunManifest {
   briefClass: BriefClass;
   classSource: ClassSource;
   forceBoard: boolean;
+  /**
+   * Execution mode. `simulated` runs stub execution; their results are typed
+   * `simulated` and must never be read as verified success. Optional because
+   * manifests written before this field existed default to `normal`.
+   */
+  mode?: RunMode;
   createdAt: string;
   updatedAt: string;
   currentStage: string;
-  status: 'in_progress' | 'completed' | 'failed' | 'interrupted' | 'rejected';
+  status:
+    | 'in_progress'
+    | 'completed'
+    | 'failed'
+    | 'interrupted'
+    | 'rejected'
+    | 'simulated'
+    | 'waiting-for-authority';
   stages: Record<string, StageStatus>;
 }
 
@@ -65,7 +137,7 @@ export interface ForgeTask {
   briefPath: string;
   resultPath: string;
   timeoutSeconds: number;
-  qualityGates: (string | GateEntry)[];
+  qualityGates: ForgeGate[];
   worktree?: string;
   command?: string;
   dependsOn?: string[];
@@ -76,7 +148,7 @@ export interface ForgeTask {
 /** Abstract task executor — decouples from packages/coord. */
 export interface TaskExecutor {
   submitTask(task: ForgeTask): Promise<void>;
-  waitForCompletion(taskId: string, timeoutMs: number): Promise<TaskResult>;
+  waitForCompletion(taskId: string, timeoutMs: number): Promise<ForgeTaskResult>;
   getTaskStatus(taskId: string): Promise<ForgeTaskStatus>;
 }
 
@@ -122,7 +194,16 @@ export interface PipelineOptions {
   stages?: string[];
   skipTo?: string;
   dryRun?: boolean;
-  executor: TaskExecutor;
+  /**
+   * Real task executor. Required in normal mode: the pipeline fails closed
+   * with FORGE_NO_EXECUTOR when it is absent.
+   */
+  executor?: TaskExecutor;
+  /**
+   * Explicit opt-in to simulated execution. Every stage and gate result is
+   * typed `simulated` and is never satisfying.
+   */
+  simulate?: boolean;
 }
 
 /** Pipeline run result. */

packages/mosaic/framework/fleet/README.md

@@ -12,6 +12,33 @@ The default tmux socket is `mosaic-fleet` so fleet commands do not touch the
 default tmux server. The roster is the desired-state authority; generated environment files are
 rebuildable projections, never a second source of configuration.
 
+## Brain-home split (fleet state vs framework templates)
+
+When a mosaic-brain clone is present, fleet **state** resolves from the brain
+home while framework templates and dispatch state stay in the config home
+(three-tree model, canon `docs/STRUCTURE-CANON.md` §2):
+
+| Path                                                                            | Without brain (legacy)                | With brain                     |
+| ------------------------------------------------------------------------------- | ------------------------------------- | ------------------------------ |
+| `fleet/agents/<seat>.env.*`                                                     | `~/.config/mosaic/fleet/agents/`      | `~/.mosaic/fleet/agents/`      |
+| `fleet/roles.local/` (overrides)                                                | `~/.config/mosaic/fleet/roles.local/` | `~/.mosaic/fleet/roles.local/` |
+| `fleet/profiles/` (working copies)                                              | `~/.config/mosaic/fleet/profiles/`    | `~/.mosaic/fleet/profiles/`    |
+| `fleet/roster.yaml`, `fleet/roles/` (baseline), `fleet/run/`, `fleet/services/` | `~/.config/mosaic/fleet/…`            | unchanged (config home)        |
+
+Activation (`packages/mosaic/src/fleet/brain-home.ts`, mirrored in
+`tools/fleet/start-agent-session.sh`):
+
+1. `MOSAIC_BRAIN_HOME` env var — explicit, always wins.
+2. Canonical `~/.mosaic` — adopted only when `MOSAIC_HOME` is the default
+   `~/.config/mosaic` AND `~/.mosaic/fleet/agents` exists. Custom
+   `--mosaic-home` values (tests, sandboxes, canaries) never adopt, keeping
+   them hermetic.
+3. Otherwise the config home (legacy single-tree behavior).
+
+Seat env dirs under a brain are subject to the same privacy boundary (0700
+dirs, 0600 files); `.env.generated` files are structure-valuable and tracked
+in the brain repo, hand-maintained `.env`/`.env.local` stay ignored and private.
+
 ## Examples
 
 - `examples/minimal.yaml` starts one local canary slot.

packages/mosaic/framework/tools/_scripts/mosaic-doctor

@@ -255,6 +255,68 @@ fleet_declared_transport() {
   printf '%s\n' "${declared:-tmux}"
 }
 
+# Brain-home fleet-state resolution (#1298; canon STRUCTURE-CANON §2).
+#
+# Seat launch envs, roles.local overrides, and profile working copies resolve
+# from the brain home when one is active; roster, baseline roles, run/, and
+# services stay under MOSAIC_HOME. This check surfaces which tree fleet state
+# resolves from and the drift a launch would otherwise hit at runtime:
+#
+# - a stale MOSAIC_BRAIN_HOME pointing at a directory with no fleet/agents is a
+#   misconfiguration the resolver honors (explicit wins) — warn, don't pass;
+# - a symlinked brain or agents dir defeats the managed-directory boundary;
+# - a group/world-readable agents dir violates the 0700 projection boundary;
+# - env files left in the config-home tree while a brain is active are split
+#   state — the write path rejects NEW split writes, but nothing would ever
+#   tell the operator the old files are stranded.
+resolve_brain_home() {
+  local explicit="${MOSAIC_BRAIN_HOME:-}"
+  if [[ -n "$(printf '%s' "$explicit" | tr -d '[:space:]')" ]]; then
+    printf '%s' "$explicit"
+    return
+  fi
+  if [[ "$(cd "$MOSAIC_HOME" 2>/dev/null && pwd -P)" == "$HOME/.config/mosaic" \
+     && -d "$HOME/.mosaic/fleet/agents" ]]; then
+    printf '%s' "$HOME/.mosaic"
+    return
+  fi
+  printf '%s' "$MOSAIC_HOME"
+}
+
+check_brain_home() {
+  local brain agents mode
+  brain="$(resolve_brain_home)"
+
+  if [[ "$brain" == "$MOSAIC_HOME" ]]; then
+    pass "Fleet state home: $MOSAIC_HOME (legacy single-tree; no brain adopted)"
+    return
+  fi
+
+  agents="$brain/fleet/agents"
+  if [[ ! -d "$agents" ]]; then
+    warn "Brain home '$brain' has no fleet/agents — seat envs will not resolve from it. Point MOSAIC_BRAIN_HOME at a brain carrying fleet/agents, or unset it."
+    return
+  fi
+  if [[ -L "$brain" || -L "$agents" ]]; then
+    warn "Brain fleet-state path resolves through a symlink ($brain) — the managed-directory boundary requires regular directories."
+    return
+  fi
+
+  mode="$(stat -c '%a' -- "$agents" 2>/dev/null)" || mode=""
+  if [[ -n "$mode" ]] && (( (8#$mode & 8#077) != 0 )); then
+    warn "Brain agents dir '$agents' is group/world-accessible (mode $mode) — the projection boundary requires 0700."
+    return
+  fi
+
+  if [[ -d "$MOSAIC_HOME/fleet/agents" ]] \
+     && ls "$MOSAIC_HOME/fleet/agents/"*.env* >/dev/null 2>&1; then
+    warn "Fleet env files exist in BOTH trees — brain '$brain' is active but '$MOSAIC_HOME/fleet/agents' still carries env files (split state). Migrate them (mosaic fleet regen) and remove the config-home copies."
+    return
+  fi
+
+  pass "Fleet state home: $brain (brain active); roster + templates: $MOSAIC_HOME"
+}
+
 check_fleet_transport() {
   local transport
   transport="$(fleet_declared_transport)"
@@ -273,6 +335,8 @@ check_fleet_transport() {
 
 check_fleet_transport
 
+check_brain_home
+
 # Legacy migration surfaces should no longer contain symlink trees.
 legacy_paths=(
   "$HOME/.claude/agent-guides"

packages/mosaic/framework/tools/_scripts/test-brain-home-check.sh

@@ -0,0 +1,108 @@
+#!/usr/bin/env bash
+# Covers the brain-home fleet-state check in `mosaic-doctor` (#1298 follow-up).
+#
+# The functions are extracted from the shipped script rather than copied here
+# (same discipline as test-fleet-transport-check.sh): a test that carries its
+# own copy of the logic keeps passing after the shipped copy changes.
+# Extraction is by exact function header and a closing brace in column one.
+set -euo pipefail
+
+SCRIPT_DIR=$(cd -- "$(dirname "$0")" && pwd)
+DOCTOR="$SCRIPT_DIR/mosaic-doctor"
+
+fail() {
+  echo "FAIL: $*" >&2
+  exit 1
+}
+
+[ -f "$DOCTOR" ] || fail "missing mosaic-doctor at $DOCTOR"
+
+extract_function() {
+  local name="$1"
+  local extracted
+  extracted=$(sed -n "/^${name}() {/,/^}/p" "$DOCTOR")
+  [ -n "$extracted" ] || fail "could not extract ${name}() from mosaic-doctor — script reshaped?"
+  printf '%s\n' "$extracted"
+}
+
+for fn in resolve_brain_home check_brain_home; do
+  extract_function "$fn" >/dev/null
+done
+
+warn_count=0
+warn() { warn_count=$((warn_count + 1)); echo "[WARN] $*"; }
+pass() { echo "[OK] $*"; return 0; }
+
+eval "$(extract_function resolve_brain_home)"
+eval "$(extract_function check_brain_home)"
+
+ROOT=$(mktemp -d)
+trap 'rm -rf "$ROOT"' EXIT
+
+run_case() {
+  # label, expect (ok|warn), then env assignments as arguments.
+  # The check runs under `env` in a subshell, so its warn() also prints a
+  # sentinel the parent counts — a subshell counter would never be visible.
+  local label="$1" expect="$2"
+  shift 2
+  local out warns
+  out=$(env "$@" bash -c "warn() { echo \"[WARN] \$*\"; }; pass() { echo \"[OK] \$*\"; return 0; }; $(extract_function resolve_brain_home); $(extract_function check_brain_home); check_brain_home" 2>&1)
+  warns=$(printf '%s\n' "$out" | grep -c '^\[WARN\]' || true)
+  if [[ "$expect" == ok && "$warns" -eq 0 ]]; then
+    echo "ok - $label"
+  elif [[ "$expect" == warn && "$warns" -gt 0 ]]; then
+    echo "ok - $label (warned)"
+  else
+    echo "output: $out" >&2
+    fail "$label: expected $expect (warns=$warns)"
+  fi
+}
+
+# ── legacy: no brain, custom home never adopts ─────────────────────────────
+mkdir -p "$ROOT/legacy-mosaic/fleet/agents"
+run_case "custom home without brain stays legacy" ok \
+  MOSAIC_HOME="$ROOT/legacy-mosaic" HOME="$ROOT"
+
+# ── healthy brain at the default config home ───────────────────────────────
+mkdir -p "$ROOT/home/.config/mosaic" "$ROOT/home/.mosaic/fleet/agents"
+chmod 700 "$ROOT/home/.mosaic/fleet/agents"
+run_case "default home adopts healthy brain" ok \
+  MOSAIC_HOME="$ROOT/home/.config/mosaic" HOME="$ROOT/home"
+
+# ── explicit MOSAIC_BRAIN_HOME to a brain without fleet/agents → warn ──────
+mkdir -p "$ROOT/brain-noagents/fleet" "$ROOT/config"
+run_case "explicit brain without agents warns" warn \
+  MOSAIC_HOME="$ROOT/config" HOME="$ROOT" MOSAIC_BRAIN_HOME="$ROOT/brain-noagents"
+
+# ── explicit MOSAIC_BRAIN_HOME to a healthy brain → ok ─────────────────────
+mkdir -p "$ROOT/brain-ok/fleet/agents" "$ROOT/config2"
+chmod 700 "$ROOT/brain-ok/fleet/agents"
+run_case "explicit healthy brain passes" ok \
+  MOSAIC_HOME="$ROOT/config2" HOME="$ROOT" MOSAIC_BRAIN_HOME="$ROOT/brain-ok"
+
+# ── group-readable agents dir → warn (0700 boundary) ───────────────────────
+mkdir -p "$ROOT/brain-loose/fleet/agents" "$ROOT/config3"
+chmod 750 "$ROOT/brain-loose/fleet/agents"
+run_case "group-readable brain agents warns" warn \
+  MOSAIC_HOME="$ROOT/config3" HOME="$ROOT" MOSAIC_BRAIN_HOME="$ROOT/brain-loose"
+
+# ── symlinked agents dir → warn (managed-directory boundary) ───────────────
+mkdir -p "$ROOT/brain-link/real-agents" "$ROOT/brain-link/fleet" "$ROOT/config4"
+ln -s "$ROOT/brain-link/real-agents" "$ROOT/brain-link/fleet/agents"
+run_case "symlinked brain agents warns" warn \
+  MOSAIC_HOME="$ROOT/config4" HOME="$ROOT" MOSAIC_BRAIN_HOME="$ROOT/brain-link"
+
+# ── split state: envs in BOTH trees → warn ─────────────────────────────────
+mkdir -p "$ROOT/brain-split/fleet/agents" "$ROOT/config5/fleet/agents"
+chmod 700 "$ROOT/brain-split/fleet/agents" "$ROOT/config5/fleet/agents"
+touch "$ROOT/config5/fleet/agents/coder0.env.generated"
+run_case "env files in both trees warns (split state)" warn \
+  MOSAIC_HOME="$ROOT/config5" HOME="$ROOT" MOSAIC_BRAIN_HOME="$ROOT/brain-split"
+
+# ── config-home agents dir WITHOUT env files alongside a brain → ok ────────
+mkdir -p "$ROOT/brain-clean/fleet/agents" "$ROOT/config6/fleet/agents"
+chmod 700 "$ROOT/brain-clean/fleet/agents" "$ROOT/config6/fleet/agents"
+run_case "empty config-home agents dir alongside brain passes" ok \
+  MOSAIC_HOME="$ROOT/config6" HOME="$ROOT" MOSAIC_BRAIN_HOME="$ROOT/brain-clean"
+
+echo "ok - mosaic-doctor brain-home check"

packages/mosaic/framework/tools/fleet/start-agent-session.sh

@@ -80,6 +80,26 @@ safe_path "$MOSAIC_HOME" || fail_env unsafe-path MOSAIC_HOME "$MOSAIC_HOME"
 
 FLEET_DIR="$MOSAIC_HOME/fleet"
 AGENT_ENV_DIR="$FLEET_DIR/agents"
+
+# Brain-home split (canon docs/STRUCTURE-CANON.md §2): seat launch envs live
+# under the brain home's fleet/agents when a brain is active; roster, roles
+# baseline, and runtime state (fleet/run) stay under MOSAIC_HOME.
+# Resolution mirrors packages/mosaic/src/fleet/brain-home.ts:
+#   1. MOSAIC_BRAIN_HOME env (explicit, always wins)
+#   2. ~/.mosaic — adopted only when MOSAIC_HOME is the default config home AND
+#      ~/.mosaic/fleet/agents exists
+#   3. MOSAIC_HOME (legacy single-tree)
+BRAIN_HOME="${MOSAIC_BRAIN_HOME:-}"
+if [ -z "$BRAIN_HOME" ]; then
+  BRAIN_HOME="$MOSAIC_HOME"
+  if [ "$(cd "$MOSAIC_HOME" 2>/dev/null && pwd -P)" = "$HOME/.config/mosaic" ] \
+     && [ -d "$HOME/.mosaic/fleet/agents" ]; then
+    BRAIN_HOME="$HOME/.mosaic"
+  fi
+fi
+if [ "$BRAIN_HOME" != "$MOSAIC_HOME" ]; then
+  AGENT_ENV_DIR="$BRAIN_HOME/fleet/agents"
+fi
 assert_managed_directory "$MOSAIC_HOME"
 assert_managed_directory "$FLEET_DIR"
 assert_private_directory "$AGENT_ENV_DIR"

packages/mosaic/framework/tools/fleet/test-start-agent-session.sh

@@ -167,6 +167,54 @@ if echo "$valid_args" | grep -qF 'bash -c'; then
   fail "launcher constructed a shell command payload"
 fi
 
+# ── Brain-home split (canon §2) ─────────────────────────────────────────
+# When MOSAIC_HOME is the default config home under $HOME and the host carries
+# $HOME/.mosaic/fleet/agents, seat envs resolve from the brain tree; the config
+# home still owns fleet/run (holder-owner) and remains a managed boundary.
+: > "$TMUX_CALLS"
+HOME_BRAIN="$ROOT/brain-home"
+CONFIG_HOME="$HOME_BRAIN/.config/mosaic"
+BRAIN="$HOME_BRAIN/.mosaic"
+mkdir -p "$CONFIG_HOME/fleet/run" "$BRAIN/fleet/agents" "$HOME_BRAIN/work"
+chmod 700 "$CONFIG_HOME" "$CONFIG_HOME/fleet" "$CONFIG_HOME/fleet/run" \
+  "$BRAIN/fleet/agents" "$HOME_BRAIN/work"
+printf '123e4567-e89b-12d3-a456-426614174000\n' > "$CONFIG_HOME/fleet/run/holder-owner"
+chmod 600 "$CONFIG_HOME/fleet/run/holder-owner"
+cat > "$BRAIN/fleet/agents/coder-brain.env.generated" <<EOF
+MOSAIC_AGENT_NAME=coder-brain
+MOSAIC_AGENT_CLASS=code
+MOSAIC_AGENT_RUNTIME=pi
+MOSAIC_AGENT_MODEL=openai-codex/gpt-5.6-sol
+MOSAIC_AGENT_REASONING=high
+MOSAIC_AGENT_TOOL_POLICY=code
+MOSAIC_AGENT_WORKDIR=$HOME_BRAIN/work
+MOSAIC_TMUX_SOCKET=mosaic-test
+EOF
+chmod 600 "$BRAIN/fleet/agents/coder-brain.env.generated"
+install_pane_binaries "$HOME_BRAIN"
+HOME="$HOME_BRAIN" PATH="$FAKE_BIN:$PATH" MOSAIC_TEST_TMUX_CALLS="$TMUX_CALLS" \
+  MOSAIC_TEST_PANE_PID=$$ MOSAIC_TEST_HOME="$HOME_BRAIN" \
+  MOSAIC_TEST_FLEET_OWNER=123e4567-e89b-12d3-a456-426614174000 \
+  MOSAIC_HOME="$CONFIG_HOME" "$START" coder-brain
+brain_args=$(tr '\0' '\n' < "$TMUX_CALLS")
+echo "$brain_args" | grep -qF new-session || fail "brain-home generated projection did not reach tmux"
+echo "$brain_args" | grep -qF 'coder-brain' || fail "brain-home agent env was not the launch source"
+[ -f "$BRAIN/fleet/agents/coder-brain.env.generated" ] || fail "brain generated env vanished"
+
+# Negative control: the SAME default-config-home shape but without
+# ~/.mosaic/fleet/agents — the config-home env tree is used directly (legacy).
+: > "$TMUX_CALLS"
+HOME_NOBRAIN="$ROOT/brainless-home"
+CONFIG_HOME_NOBRAIN="$HOME_NOBRAIN/.config/mosaic"
+write_generated "$CONFIG_HOME_NOBRAIN" "coder-legacy"
+install_pane_binaries "$HOME_NOBRAIN"
+HOME="$HOME_NOBRAIN" PATH="$FAKE_BIN:$PATH" MOSAIC_TEST_TMUX_CALLS="$TMUX_CALLS" \
+  MOSAIC_TEST_PANE_PID=$$ MOSAIC_TEST_HOME="$HOME_NOBRAIN" \
+  MOSAIC_TEST_FLEET_OWNER=123e4567-e89b-12d3-a456-426614174000 \
+  MOSAIC_HOME="$CONFIG_HOME_NOBRAIN" "$START" coder-legacy
+legacy_args=$(tr '\0' '\n' < "$TMUX_CALLS")
+echo "$legacy_args" | grep -qF new-session || fail "legacy single-tree launch regressed"
+
 # The pane must start through an absolute clean-environment boundary. Its
 # runtime command remains an argv vector, but no holder/session environment
 # control variable can pass through the pane command.

packages/mosaic/framework/tools/git/README.md

@@ -30,7 +30,7 @@ The Gitea API token is **never passed on a curl command line.** An `Authorizatio
 
 ### `--login` override
 
-Both `pr-review.sh` and `issue-comment.sh` accept an optional `--login <name>` flag that overrides the automatically detected Gitea login for that single invocation (as of #1280, `pr-create.sh`, `pr-merge.sh` and `issue-create.sh` accept it too, and it wins over `MOSAIC_GIT_IDENTITY` everywhere). The override selects **which credential the REST write, the `/user` identity lookup, and the read-back all use** — its token is resolved from the tea config for that login name (`get_gitea_token_for_login`), falling back to the repo host's credential when no login is named. The resolved login is **host- and port-bound**: the login's configured URL host **and effective port** (the scheme's default port — 80 for `http`, 443 for `https` — applies when a port is omitted, symmetrically on both sides) must match the repo remote's, so a login name shared across hosts (or an override configured for a different Gitea, including one on a different port of the same host) can never send one host's credential to another — a host or port mismatch fails closed rather than leaking a cross-host token. Resolving the acting identity and the read-back from the _same_ login that performs the write is essential: a write performed under an overridden login must be verified against that login's identity, not the host default's. Callers who need a different login than the host default should pass `--login <reviewer-login>`.
+Both `pr-review.sh` and `issue-comment.sh` accept an optional `--login <name>` flag that overrides the automatically detected Gitea login for that single invocation. The override selects **which credential the REST write, the `/user` identity lookup, and the read-back all use** — its token is resolved from the tea config for that login name (`get_gitea_token_for_login`), falling back to the repo host's credential when no login is named. The resolved login is **host- and port-bound**: the login's configured URL host **and effective port** (the scheme's default port — 80 for `http`, 443 for `https` — applies when a port is omitted, symmetrically on both sides) must match the repo remote's, so a login name shared across hosts (or an override configured for a different Gitea, including one on a different port of the same host) can never send one host's credential to another — a host or port mismatch fails closed rather than leaking a cross-host token. Resolving the acting identity and the read-back from the _same_ login that performs the write is essential: a write performed under an overridden login must be verified against that login's identity, not the host default's. Callers who need a different login than the host default should pass `--login <reviewer-login>`.
 
 As a durable successor to this mechanism, consider giving each reviewer/approver slot its own dedicated Gitea login credential, so that author≠reviewer holds at the credential level rather than relying on wrapper-level `--login` bookkeeping. This is a recommendation for future hardening, not something implemented by this flag.
 
@@ -58,36 +58,6 @@ token file present, both tools fall through to the existing shared-account path
 unchanged, so this feature is a no-op on any host that hasn't provisioned per-slot
 tokens.
 
-### Identity-first principal resolution in the wrappers (#1280)
-
-`resolve_gitea_principal()` (detect-platform.sh) gives the write wrappers —
-`pr-create.sh`, `pr-merge.sh`, `pr-review.sh`, `issue-create.sh`, `issue-comment.sh` —
-ONE precedence for choosing the acting principal:
-
-1. an explicit `--login <name>` (now accepted by all five; operator intent beats
-   environment), then
-2. the per-agent identity above (`MOSAIC_GIT_IDENTITY` env / worktree
-   `mosaic.gitIdentity`) when a per-slot token exists — the wrapper then writes via the
-   REST API with that identity's token and never consults `tea`, so the tea login list
-   cannot shadow the requested principal, then
-3. the tea login list — the LAST resort, never the first, because it enumerates
-   whatever logins the host happens to hold and knows nothing about which seat is
-   calling.
-
-A requested identity whose per-slot token is absent, or a `--login` whose token cannot
-resolve host-bound, **fails loud** (nonzero, naming the identity/login and the expected
-slot) instead of silently writing under whatever account `tea` has configured — that
-silent fallthrough is defect #1280 (reviews, comments, merges, PRs and issues filed
-under the wrong account). `pr-merge.sh --dry-run` reports the principal the merge would
-act as, resolved exactly as the real merge resolves it. ⚠ A **workstation-global**
-`mosaic.gitIdentity` shadows every seat on that host (a fresh clone with no local value
-resolves the global one) — set it per-worktree, not with `--global`.
-
-The resolver is covered by `test-gitea-principal-resolution.sh`; the happy-path
-ordering (identity arm REACHED, not sitting behind a tea failure) by
-`test-pr-create-identity-first.sh`; merge credential binding by
-`test-pr-merge-principal-resolution.sh`.
-
 ### Enabling it for a clone
 
 The framework installer syncs `git-credential-mosaic` to

packages/mosaic/framework/tools/git/detect-platform.sh

@@ -497,32 +497,6 @@ get_gitea_url_for_host() {
     return 1
 }
 
-# Map a Gitea host to the per-agent identity-token slot PREFIX ("gitea-usc" /
-# "gitea-mosaicstack") used by identity-first principal resolution
-# (MOSAIC_GIT_IDENTITY / git config mosaic.gitIdentity; #1280). Returns 1 for
-# hosts with no per-slot scheme — callers treat that as "identity does not
-# bind here" and fall through to existing behavior, never as an error. This is
-# the single source of truth for the slot layout: get_gitea_token and
-# resolve_gitea_principal both derive their slot paths from here, so the two
-# resolutions can never disagree about where an identity's credential lives.
-gitea_identity_slot_prefix() {
-    case "$1" in
-        git.uscllc.com)      echo "gitea-usc" ;;
-        git.mosaicstack.dev) echo "gitea-mosaicstack" ;;
-        *) return 1 ;;
-    esac
-}
-
-# Resolve the per-slot token FILE PATH for an identity on a host. Prints the
-# absolute path on success; returns 1 (no output) when the host has no per-slot
-# scheme. Prints a PATH only — never a token value.
-gitea_identity_token_slot() {
-    local identity="$1" host="$2" prefix
-    [[ -n "$identity" ]] || return 1
-    prefix=$(gitea_identity_slot_prefix "$host") || return 1
-    printf '%s\n' "$HOME/.config/mosaic/secrets/gitea-tokens/${prefix}-${identity}.token"
-}
-
 # Resolve a Gitea API token for the given host.
 # Priority: Mosaic credential loader → GITEA_TOKEN env → ~/.git-credentials
 get_gitea_token() {
@@ -543,8 +517,13 @@ get_gitea_token() {
         _ident_src="git config mosaic.gitIdentity"
     fi
     if [[ -n "$_ident" ]]; then
-        local _idtok=""
-        if _idtok="$(gitea_identity_token_slot "$_ident" "$host" 2>/dev/null)"; then
+        local _idpfx=""
+        case "$host" in
+            git.uscllc.com)      _idpfx=gitea-usc ;;
+            git.mosaicstack.dev) _idpfx=gitea-mosaicstack ;;
+        esac
+        if [[ -n "$_idpfx" ]]; then
+            local _idtok="$HOME/.config/mosaic/secrets/gitea-tokens/${_idpfx}-${_ident}.token"
             if [[ -r "$_idtok" ]]; then
                 cat "$_idtok"
                 return 0
@@ -1486,81 +1465,6 @@ raise SystemExit(1)
 PY
 }
 
-# resolve_gitea_principal — identity-first acting-principal resolution shared by
-# the git wrappers (#1280). The defect this fixes: wrappers resolved their
-# acting principal from tea's login list FIRST, and that list enumerates
-# whatever logins happen to be configured on the host — it knows nothing about
-# which seat is calling — so a wrapper invoked with MOSAIC_GIT_IDENTITY=fargo
-# still wrote under whichever account tea held (mos-dt-0), and the correct
-# identity-aware code sat behind arms that only ran when the tea path failed.
-# Precedence here is the contract:
-#   1. an explicit login override ($1, the wrapper's --login) — operator intent
-#      beats environment;
-#   2. MOSAIC_GIT_IDENTITY env, else per-worktree `git config mosaic.gitIdentity`
-#      (mirroring get_gitea_token exactly, so resolver and token resolution can
-#      never disagree) — binds only on hosts with a per-slot token scheme;
-#   3. the tea login list — LAST resort, never the first.
-#
-# Prints exactly one line, three tab-separated fields (machine-readable for
-# wrapper dispatch and tests):
-#   mode        "login" | "identity" | "default"
-#   principal   login name (login) | identity name (identity) | tea login or "" (default)
-#   source      "tea-login:<name>" | "identity-slot:<path>" | "tea-default" | "host-credential"
-#
-# Fails LOUD (nonzero, empty stdout, stderr diagnostic) when an explicit
-# override cannot be honored — a refusal is a good day; silently falling
-# through to whoever tea has configured is the exact defect this resolves:
-#   - login mode: no host-bound token for that tea login. The existence check
-#     runs the same tea-config lookup tea itself uses; the token VALUE is
-#     discarded (never printed, never used).
-#   - identity mode: no per-slot token file for that identity on a recognized
-#     host — the diagnostic names the identity, its source, and the expected
-#     slot path. An identity requested on a host with NO per-slot scheme does
-#     not bind (matching get_gitea_token's containment) and falls to default.
-#
-# NEVER prints a token value — principal names and slot paths only.
-# $1 = explicit login override ("" when absent), $2 = host (default: the
-# origin remote's host).
-resolve_gitea_principal() {
-    local login_override="${1:-}" host="${2:-}" ident ident_src slot login
-    [[ -n "$host" ]] || { host=$(get_remote_host) || return 1; }
-
-    if [[ -n "$login_override" ]]; then
-        get_gitea_token_for_login "$login_override" "$host" >/dev/null || {
-            echo "Error: --login '$login_override' has no host-matched token on host '$host' (tea config lookup); refusing to fall back to any other principal (#1280 identity-first resolution)." >&2
-            return 1
-        }
-        printf 'login\t%s\ttea-login:%s\n' "$login_override" "$login_override"
-        return 0
-    fi
-
-    ident="${MOSAIC_GIT_IDENTITY:-}"
-    ident_src="MOSAIC_GIT_IDENTITY"
-    if [[ -z "$ident" ]]; then
-        ident="$(git config --get mosaic.gitIdentity 2>/dev/null || true)"
-        ident_src="git config mosaic.gitIdentity"
-    fi
-    if [[ -n "$ident" ]] && slot="$(gitea_identity_token_slot "$ident" "$host" 2>/dev/null)"; then
-        if [[ -r "$slot" ]]; then
-            printf 'identity\t%s\tidentity-slot:%s\n' "$ident" "$slot"
-            return 0
-        fi
-        echo "Error: git identity '$ident' requested (via $ident_src) for host '$host', but no per-slot token at $slot (#1280 identity-first resolution)." >&2
-        echo "       Refusing to fall back to the tea login list or shared credentials. Provision the per-slot token, or unset the identity." >&2
-        return 1
-    fi
-
-    # No override requested: tea's login list is the LAST resort. Absence is
-    # not an error here — callers fall back to the host credential, exactly as
-    # they did before this resolver existed (preserved behavior).
-    if login=$(get_gitea_login_for_host "$host" 2>/dev/null); then
-        printf 'default\t%s\ttea-default\n' "$login"
-    else
-        printf 'default\t\thost-credential\n'
-    fi
-    return 0
-}
-
 # Resolve HTTPS basic auth credentials for a Gitea host from ~/.git-credentials.
 # Prints "username:password" for direct curl -u consumption. Callers must not log it.
 get_gitea_basic_auth() {

packages/mosaic/framework/tools/git/issue-comment.sh

@@ -76,36 +76,27 @@ fi
 detect_platform >/dev/null
 
 # Resolve and cache the Gitea REST endpoint + token for the current remote,
-# bound to a SPECIFIC acting principal ($1) selected identity-first (#1280):
-# an explicit --login wins, else MOSAIC_GIT_IDENTITY / git config
-# mosaic.gitIdentity binds the per-slot credential, else the tea login list
-# (last resort). Populates GITEA_API_ROOT (…/api/v1), GITEA_API_BASE
-# (…/api/v1/repos/<slug>), and GITEA_API_TOKEN.
+# bound to a SPECIFIC login identity ($1). Populates GITEA_API_ROOT (…/api/v1),
+# GITEA_API_BASE (…/api/v1/repos/<slug>), and GITEA_API_TOKEN.
 #
-# The token is resolved for the EFFECTIVE principal so that the single
-# credential used for the write ALSO drives the /user identity read and the
-# read-back — write token and read-back token are the same identity by
-# construction (this is the credential-ordering fix: a --login override is no
-# longer written under one credential and verified under a different default
-# one). When $2 is "identity" the principal ($1) is a requested git identity:
-# the token MUST resolve from that identity's per-slot token (get_gitea_token's
-# identity arm), failing closed rather than borrowing the tea default login —
-# the tea login list must never shadow a requested identity (#1280). When $2
-# is "explicit" the principal came from a caller-supplied --login: that exact
-# login's token MUST resolve, and we FAIL CLOSED rather than silently
-# downgrading the write to the host default identity. Otherwise the best-effort
-# default path applies (per-login token, else the host-scoped credential).
-# Returns non-zero (clear stderr) on any resolution failure.
+# The token is resolved for the EFFECTIVE login (the --login override when
+# given, otherwise the detected default) so that the single credential used for
+# the write ALSO drives the /user identity read and the read-back — write token
+# and read-back token are the same identity by construction (this is the
+# credential-ordering fix: a --login override is no longer written under one
+# credential and verified under a different default one). Falls back to the
+# host-scoped credential ONLY when NO --login override was supplied (the
+# best-effort default path). When $2 is "explicit" the login came from a
+# caller-supplied --login: that exact login's token MUST resolve, and we FAIL
+# CLOSED rather than silently downgrading the write to the host default
+# identity — otherwise a caller relying on a dedicated per-role credential would
+# be told the write succeeded as requested while it was attributed to the shared
+# default. Returns non-zero (clear stderr) on any resolution failure.
 gitea_resolve_api_for_login() {
     local effective_login="$1" override_explicit="${2:-}" host configured_url repo
 
     host=$(get_remote_host)
-    if [[ "$override_explicit" == "identity" ]]; then
-        GITEA_API_TOKEN=$(get_gitea_token "$host") || {
-            echo "Error: could not resolve the per-slot token for requested git identity '$effective_login' on host '$host'; refusing to fall back to the tea login list or shared credentials (comment write/read-back, #1280)." >&2
-            return 1
-        }
-    elif [[ -n "$override_explicit" ]]; then
+    if [[ -n "$override_explicit" ]]; then
         GITEA_API_TOKEN=$(get_gitea_token_for_login "$effective_login" "$host") || {
             echo "Error: could not resolve a host-matched Gitea token for --login '$effective_login' on host '$host'; refusing to fall back to the host default identity or a cross-host credential (comment write/read-back)" >&2
             return 1
@@ -327,31 +318,23 @@ if [[ "$PLATFORM" == "github" ]]; then
     gh issue comment "$ISSUE_NUMBER" --body "$COMMENT"
     echo "Added comment to GitHub issue #$ISSUE_NUMBER"
 elif [[ "$PLATFORM" == "gitea" ]]; then
-    # Resolve the acting principal identity-first (#1280): an explicit --login
-    # wins; otherwise MOSAIC_GIT_IDENTITY / git config mosaic.gitIdentity
-    # selects the principal when a per-slot token exists (fail-loud when it
-    # does not); the tea login list is the LAST resort — it knows nothing about
-    # which seat is calling, so resolving from it first wrote under whichever
-    # account tea had configured (the #1280 family).
-    principal_host=$(get_remote_host)
-    if ! principal_resolved="$(resolve_gitea_principal "$LOGIN_OVERRIDE" "$principal_host")"; then
-        # resolve_gitea_principal already printed the fail-loud diagnostic.
-        exit 1
-    fi
-    PRINCIPAL_MODE="$(printf '%s' "$principal_resolved" | cut -f1)"
-    PRINCIPAL_NAME="$(printf '%s' "$principal_resolved" | cut -f2)"
+    # Resolve the login this comment should be attributed to: the --login
+    # override when given, otherwise the detected default for this repo's host.
+    # A --login override always wins. Otherwise name this repo host's login only
+    # as a best effort: the login name merely selects a per-login token, and
+    # gitea_resolve_api_for_login falls back to the host credential
+    # (get_gitea_token) when no tea login is named, so the default credential
+    # still resolves even when the host tea has no matching login entry.
+    EFFECTIVE_LOGIN="$LOGIN_OVERRIDE"
+    [[ -n "$EFFECTIVE_LOGIN" ]] || EFFECTIVE_LOGIN=$(get_gitea_login 2>/dev/null || true)
 
-    # Bind the REST endpoint + token to the resolved principal, then derive the
+    # Bind the REST endpoint + token to the effective login, then derive the
     # acting identity from that SAME credential (GET /user). The write below and
     # its read-back both use this credential, so the write is verified against
-    # the identity that actually performed it.
-    if [[ "$PRINCIPAL_MODE" == "identity" ]]; then
-        gitea_resolve_api_for_login "$PRINCIPAL_NAME" identity || exit 1
-    elif [[ "$PRINCIPAL_MODE" == "login" ]]; then
-        gitea_resolve_api_for_login "$PRINCIPAL_NAME" explicit || exit 1
-    else
-        gitea_resolve_api_for_login "$PRINCIPAL_NAME" "" || exit 1
-    fi
+    # the identity that actually performed it. Passing "explicit" when --login
+    # was supplied forbids the host-default fallback: an unresolvable explicit
+    # override fails closed instead of writing under the default identity.
+    gitea_resolve_api_for_login "$EFFECTIVE_LOGIN" "${LOGIN_OVERRIDE:+explicit}" || exit 1
     ACTING_LOGIN=$(gitea_authenticated_login) || exit 1
 
     comment_id=$(gitea_create_comment_verified "$ISSUE_NUMBER" "$COMMENT" "$ACTING_LOGIN") || {

packages/mosaic/framework/tools/git/issue-create.sh

@@ -1,15 +1,6 @@
 #!/bin/bash
 # issue-create.sh - Create issues on Gitea or GitHub
-# Usage: issue-create.sh -t "Title" [-b "Body"] [-l "label1,label2"] [-m "milestone"] [--login <name>]
-#
-# Acting principal is resolved identity-first (#1280): an explicit --login
-# wins; otherwise MOSAIC_GIT_IDENTITY / per-worktree git config
-# mosaic.gitIdentity selects the principal when a per-slot token exists (and
-# the wrapper then creates the issue through the REST API with that identity's
-# token — tea is never invoked, so the tea login list cannot shadow the
-# requested principal); the tea login list is the LAST resort. A requested
-# identity with no per-slot token fails LOUD rather than writing under
-# whichever account tea happens to hold.
+# Usage: issue-create.sh -t "Title" [-b "Body"] [-l "label1,label2"] [-m "milestone"]
 
 set -e
 
@@ -25,14 +16,6 @@ INTERACTIVE=false
 
 # get_remote_host and get_gitea_token are provided by detect-platform.sh
 
-# Acting-principal mode set in the Gitea branch below (from
-# resolve_gitea_principal): "login" when --login was given, "identity" when a
-# git identity bound, "default" otherwise. PRINCIPAL_MODE=login makes the API
-# arm resolve the --login principal's token too, so an explicit --login keeps
-# winning even on the tea-FAILURE fallback arm.
-PRINCIPAL_MODE=""
-PRINCIPAL_NAME=""
-
 gitea_issue_create_api() {
     local host repo token url payload
     host=$(get_remote_host) || {
@@ -43,19 +26,10 @@ gitea_issue_create_api() {
         echo "Error: could not determine repo owner/name for API fallback" >&2
         return 1
     }
-    if [[ "$PRINCIPAL_MODE" == "login" ]]; then
-        token=$(get_gitea_token_for_login "$PRINCIPAL_NAME" "$host") || {
-            echo "Error: could not resolve a host-matched Gitea token for --login '$PRINCIPAL_NAME' on host '$host' (API path)" >&2
-            return 1
-        }
-    else
-        # Identity-first when MOSAIC_GIT_IDENTITY / git config mosaic.gitIdentity
-        # is set (per-slot token, fail-loud on absence); shared default otherwise.
-        token=$(get_gitea_token "$host") || {
-            echo "Error: Gitea token not found for API fallback (set GITEA_TOKEN or configure ~/.git-credentials)" >&2
-            return 1
-        }
-    fi
+    token=$(get_gitea_token "$host") || {
+        echo "Error: Gitea token not found for API fallback (set GITEA_TOKEN or configure ~/.git-credentials)" >&2
+        return 1
+    }
 
     if [[ -n "$LABELS" || -n "$MILESTONE" ]]; then
         echo "Warning: API fallback currently applies title/body only; labels/milestone require authenticated tea setup." >&2
@@ -93,7 +67,6 @@ Options:
     -b, --body BODY         Issue body/description
     -l, --labels LABELS     Comma-separated labels (e.g., "bug,feature")
     -m, --milestone NAME    Milestone name to assign
-        --login NAME        Act as this Gitea tea login (wins over MOSAIC_GIT_IDENTITY)
     -i, --interactive       Prompt for missing issue fields
     -h, --help              Show this help message
 
@@ -124,10 +97,6 @@ while [[ $# -gt 0 ]]; do
             MILESTONE="$2"
             shift 2
             ;;
-        --login)
-            LOGIN_OVERRIDE="$2"
-            shift 2
-            ;;
         -i|--interactive)
             INTERACTIVE=true
             shift
@@ -165,37 +134,13 @@ case "$PLATFORM" in
         "${CMD[@]}"
         ;;
     gitea)
-        # Resolve the acting principal identity-first (#1280). The tea login
-        # list is the LAST resort: it knows nothing about which seat is calling,
-        # and a login resolved from it first is what attributed issues to the
-        # wrong account even when MOSAIC_GIT_IDENTITY was set.
-        principal_host=$(get_remote_host 2>/dev/null || true)
-        if ! principal_resolved="$(resolve_gitea_principal "${LOGIN_OVERRIDE:-}" "$principal_host")"; then
-            # resolve_gitea_principal already printed the fail-loud diagnostic.
-            exit 1
-        fi
-        PRINCIPAL_MODE="$(printf '%s' "$principal_resolved" | cut -f1)"
-        PRINCIPAL_NAME="$(printf '%s' "$principal_resolved" | cut -f2)"
-
-        if [[ "$PRINCIPAL_MODE" == "identity" ]]; then
-            # HAPPY PATH for a requested identity: create through the REST API
-            # with the per-slot token and never invoke tea — the identity arm
-            # must be REACHED, not sit behind a tea failure (#1280).
-            gitea_issue_create_api
-            exit $?
-        fi
-
         if command -v tea >/dev/null 2>&1; then
             REPO_SLUG=$(get_repo_slug)
-            if [[ "$PRINCIPAL_MODE" == "login" ]]; then
-                GITEA_LOGIN_NAME="$PRINCIPAL_NAME"
-            else
-                GITEA_LOGIN_NAME=$(get_gitea_login) || {
-                    echo "Warning: could not resolve Gitea login for tea; trying Gitea API fallback..." >&2
-                    gitea_issue_create_api
-                    exit $?
-                }
-            fi
+            GITEA_LOGIN_NAME=$(get_gitea_login) || {
+                echo "Warning: could not resolve Gitea login for tea; trying Gitea API fallback..." >&2
+                gitea_issue_create_api
+                exit $?
+            }
             if ! get_gitea_authenticated_user "$GITEA_LOGIN_NAME" >/dev/null; then
                 echo "Warning: Tea authenticated-user validation failed (possible stale user/login); trying Gitea API fallback..." >&2
                 gitea_issue_create_api

packages/mosaic/framework/tools/git/pr-create.sh

@@ -1,15 +1,6 @@
 #!/bin/bash
 # pr-create.sh - Create pull requests on Gitea or GitHub
-# Usage: pr-create.sh -t "Title" [-b "Body"] [-B base] [-H head] [-l "labels"] [-m "milestone"] [--login <name>]
-#
-# Acting principal is resolved identity-first (#1280): an explicit --login
-# wins; otherwise MOSAIC_GIT_IDENTITY / per-worktree git config
-# mosaic.gitIdentity selects the principal when a per-slot token exists (and
-# the wrapper then creates the PR through the REST API with that identity's
-# token — tea is never invoked, so the tea login list cannot shadow the
-# requested principal); the tea login list is the LAST resort. A requested
-# identity with no per-slot token fails LOUD rather than writing under
-# whichever account tea happens to hold.
+# Usage: pr-create.sh -t "Title" [-b "Body"] [-B base] [-H head] [-l "labels"] [-m "milestone"]
 
 set -e
 
@@ -28,15 +19,6 @@ ISSUE=""
 
 # get_remote_host, get_gitea_token, get_repo_info, and get_gitea_repo_args are provided by detect-platform.sh
 
-# Acting-principal mode set in the Gitea branch below (from
-# resolve_gitea_principal): "login" when --login was given, "identity" when a
-# git identity bound, "default" otherwise. PRINCIPAL_MODE=login makes the API
-# arm resolve the --login principal's token too, so an explicit --login keeps
-# winning even on the tea-FAILURE fallback arm (otherwise the fallback would
-# silently re-resolve to the environment identity or shared credential).
-PRINCIPAL_MODE=""
-PRINCIPAL_NAME=""
-
 gitea_pr_create_api() {
     local host repo token url payload
     host=$(get_remote_host) || {
@@ -47,19 +29,10 @@ gitea_pr_create_api() {
         echo "Error: could not determine repo owner/name for API fallback" >&2
         return 1
     }
-    if [[ "$PRINCIPAL_MODE" == "login" ]]; then
-        token=$(get_gitea_token_for_login "$PRINCIPAL_NAME" "$host") || {
-            echo "Error: could not resolve a host-matched Gitea token for --login '$PRINCIPAL_NAME' on host '$host' (API path)" >&2
-            return 1
-        }
-    else
-        # Identity-first when MOSAIC_GIT_IDENTITY / git config mosaic.gitIdentity
-        # is set (per-slot token, fail-loud on absence); shared default otherwise.
-        token=$(get_gitea_token "$host") || {
-            echo "Error: Gitea token not found for API fallback (set GITEA_TOKEN or configure ~/.git-credentials)" >&2
-            return 1
-        }
-    fi
+    token=$(get_gitea_token "$host") || {
+        echo "Error: Gitea token not found for API fallback (set GITEA_TOKEN or configure ~/.git-credentials)" >&2
+        return 1
+    }
 
     if [[ -n "$LABELS" || -n "$MILESTONE" || "$DRAFT" == true ]]; then
         echo "Warning: API fallback applies title/body/head/base only; labels/milestone/draft require authenticated tea setup." >&2
@@ -103,7 +76,6 @@ Options:
     -H, --head BRANCH       Head branch with changes (default: current branch)
     -l, --labels LABELS     Comma-separated labels
     -m, --milestone NAME    Milestone name
-        --login NAME        Act as this Gitea tea login (wins over MOSAIC_GIT_IDENTITY)
     -i, --issue NUMBER      Link to issue (auto-generates title if not provided)
     -d, --draft             Create as draft PR
     -h, --help              Show this help message
@@ -144,10 +116,6 @@ while [[ $# -gt 0 ]]; do
             MILESTONE="$2"
             shift 2
             ;;
-        --login)
-            LOGIN_OVERRIDE="$2"
-            shift 2
-            ;;
         -i|--issue)
             ISSUE="$2"
             shift 2
@@ -206,41 +174,15 @@ case "$PLATFORM" in
         "${CMD[@]}"
         ;;
     gitea)
-        # Resolve the acting principal identity-first (#1280). The tea login
-        # list is the LAST resort: it knows nothing about which seat is calling,
-        # and a login resolved from it first is what attributed PRs to the wrong
-        # account even when MOSAIC_GIT_IDENTITY was set.
-        principal_host=$(get_remote_host 2>/dev/null || true)
-        if ! principal_resolved="$(resolve_gitea_principal "${LOGIN_OVERRIDE:-}" "$principal_host")"; then
-            # resolve_gitea_principal already printed the fail-loud diagnostic.
-            exit 1
-        fi
-        PRINCIPAL_MODE="$(printf '%s' "$principal_resolved" | cut -f1)"
-        PRINCIPAL_NAME="$(printf '%s' "$principal_resolved" | cut -f2)"
-
-        if [[ "$PRINCIPAL_MODE" == "identity" ]]; then
-            # HAPPY PATH for a requested identity: the per-slot token IS the
-            # credential, so create through the REST API directly and never
-            # invoke tea — the identity arm must be REACHED, not sit behind a
-            # tea failure (#1280). Fail-loud on a missing slot already happened
-            # in resolve_gitea_principal.
-            gitea_pr_create_api
-            exit $?
-        fi
-
         # tea pull create syntax. Always pass --repo because tea repo inference
         # is unreliable in Mosaic worktrees/profile shells. Use arrays instead
         # of eval so markdown backticks/body content are not shell-executed.
         REPO_SLUG=$(get_repo_slug)
-        if [[ "$PRINCIPAL_MODE" == "login" ]]; then
-            GITEA_LOGIN_NAME="$PRINCIPAL_NAME"
-        else
-            GITEA_LOGIN_NAME=$(get_gitea_login) || {
-                echo "Warning: could not resolve Gitea login for tea; trying Gitea API fallback..." >&2
-                gitea_pr_create_api
-                exit $?
-            }
-        fi
+        GITEA_LOGIN_NAME=$(get_gitea_login) || {
+            echo "Warning: could not resolve Gitea login for tea; trying Gitea API fallback..." >&2
+            gitea_pr_create_api
+            exit $?
+        }
         if ! get_gitea_authenticated_user "$GITEA_LOGIN_NAME" >/dev/null; then
             echo "Warning: Tea authenticated-user validation failed (possible stale user/login); trying Gitea API fallback..." >&2
             gitea_pr_create_api

packages/mosaic/framework/tools/git/pr-merge.sh

@@ -1,13 +1,6 @@
 #!/bin/bash
 # pr-merge.sh - Merge pull requests on Gitea or GitHub
-# Usage: pr-merge.sh -n PR_NUMBER [-m squash] [-d] [--expect-head SHA] [--co-author-trailers --escalate-to PRINCIPAL] [--login <name>]
-#
-# Acting principal is resolved identity-first (#1280): an explicit --login
-# wins; otherwise MOSAIC_GIT_IDENTITY / per-worktree git config
-# mosaic.gitIdentity selects the credential (per-slot token, fail-loud when
-# absent); the shared host credential is the last resort. The merge is
-# performed with the resolved credential only — never a cross-principal
-# fallback (an HTTP 401 from the identity-bound token is a hard stop).
+# Usage: pr-merge.sh -n PR_NUMBER [-m squash] [-d] [--expect-head SHA] [--co-author-trailers --escalate-to PRINCIPAL]
 
 set -euo pipefail
 
@@ -23,7 +16,6 @@ DRY_RUN=false
 EXPECT_HEAD=""
 CO_AUTHOR_TRAILERS=false
 ESCALATE_TO=""
-LOGIN_OVERRIDE=""
 
 usage() {
     cat <<EOF
@@ -39,7 +31,6 @@ Options:
         --expect-head SHA   Refuse unless the PR head matches this full commit SHA
         --co-author-trailers Build verified trailers from linked PR commit authors
         --escalate-to NAME  Named principal for an unresolved-author BLOCK
-        --login NAME        Act as this Gitea tea login (wins over MOSAIC_GIT_IDENTITY)
     -h, --help              Show this help message
 
 Examples:
@@ -48,7 +39,6 @@ Examples:
     $(basename "$0") -n 42 -d                 # Squash merge and delete branch
     $(basename "$0") -n 42 --expect-head 0123456789abcdef0123456789abcdef01234567
     $(basename "$0") -n 42 --co-author-trailers --escalate-to tl-mosaic
-    $(basename "$0") -n 42 --login fred-ms    # Merge under the fred-ms tea login
 EOF
     exit "${1:-1}"
 }
@@ -92,14 +82,6 @@ while [[ $# -gt 0 ]]; do
             ESCALATE_TO="$2"
             shift 2
             ;;
-        --login|-l)
-            if [[ $# -lt 2 ]]; then
-                echo "Error: --login requires one tea login name." >&2
-                exit 1
-            fi
-            LOGIN_OVERRIDE="$2"
-            shift 2
-            ;;
         -h|--help)
             usage 0
             ;;
@@ -590,22 +572,9 @@ PY
 merge_gitea_with_api() {
     local host="$1" token attempt_rc
 
-    # Identity-first principal resolution (#1280): an explicit --login wins
-    # over MOSAIC_GIT_IDENTITY (operator intent beats environment); otherwise
-    # get_gitea_token resolves the identity's per-slot token when an identity
-    # is requested (fail-loud when absent) and the shared host credential only
-    # when no identity is set. No cross-principal fallback: whatever resolves
-    # here is the ONLY credential the merge is attempted with.
-    if [[ -n "$LOGIN_OVERRIDE" ]]; then
-        if ! token=$(get_gitea_token_for_login "$LOGIN_OVERRIDE" "$host"); then
-            echo "Error: --login '$LOGIN_OVERRIDE' has no host-matched token on host '$host'; refusing to merge under any other principal (#1280 identity-first resolution)." >&2
-            return 1
-        fi
-    else
-        if ! token=$(get_gitea_token "$host"); then
-            echo "Error: Could not resolve the required Gitea token; refusing merge without changing principals." >&2
-            return 1
-        fi
+    if ! token=$(get_gitea_token "$host"); then
+        echo "Error: Could not resolve the required Gitea token; refusing merge without changing principals." >&2
+        return 1
     fi
     if [[ -z "$token" ]]; then
         echo "Error: Required Gitea token resolved empty; refusing merge without changing principals." >&2
@@ -633,25 +602,10 @@ if [[ "$DRY_RUN" == true ]]; then
             echo "Error: Cannot determine host from origin remote URL" >&2
             exit 1
         }
-        # Report the acting principal the merge WOULD use, resolved the same
-        # way the real merge resolves it (#1280) — a dry run that names a
-        # different principal than the merge would act as is a lie.
-        if ! principal_resolved="$(resolve_gitea_principal "$LOGIN_OVERRIDE" "$HOST")"; then
-            # Fail-loud diagnostic already printed (unresolvable --login or a
-            # requested identity with no per-slot token).
-            exit 1
-        fi
-        DRY_PRINCIPAL_MODE="$(printf '%s' "$principal_resolved" | cut -f1)"
-        DRY_PRINCIPAL_NAME="$(printf '%s' "$principal_resolved" | cut -f2)"
-        case "$DRY_PRINCIPAL_MODE" in
-            login)    DRY_PRINCIPAL_DESC="tea login '$DRY_PRINCIPAL_NAME'" ;;
-            identity) DRY_PRINCIPAL_DESC="git identity '$DRY_PRINCIPAL_NAME' (per-slot credential)" ;;
-            *)        DRY_PRINCIPAL_DESC="default host credential" ;;
-        esac
         if [[ "$CO_AUTHOR_TRAILERS" == true ]]; then
-            echo "Dry run: would verify PR commit authors and merge PR #$PR_NUMBER on $HOST as $DRY_PRINCIPAL_DESC with authenticated Gitea API message fields (base=$BASE_BRANCH, method=squash)."
+            echo "Dry run: would verify PR commit authors and merge PR #$PR_NUMBER on $HOST with authenticated Gitea API message fields (base=$BASE_BRANCH, method=squash)."
         else
-            echo "Dry run: would merge PR #$PR_NUMBER on $HOST as $DRY_PRINCIPAL_DESC with the authenticated exact-head Gitea API path (base=$BASE_BRANCH, method=squash)."
+            echo "Dry run: would merge PR #$PR_NUMBER on $HOST with the authenticated exact-head Gitea API path (base=$BASE_BRANCH, method=squash)."
         fi
     else
         echo "Dry run: would merge PR #$PR_NUMBER on $PLATFORM (base=$BASE_BRANCH, method=squash)."

packages/mosaic/framework/tools/git/pr-review.sh

@@ -76,7 +76,7 @@ while [[ $# -gt 0 ]]; do
             echo "  -n, --number   PR number (required)"
             echo "  -a, --action   Review action: approve, request-changes, comment (required)"
             echo "  -c, --comment  Review comment (required for request-changes)"
-            echo "  -l, --login    Override the detected Gitea tea login (all actions; wins over MOSAIC_GIT_IDENTITY)"
+            echo "  -l, --login    Override the detected Gitea tea login (approve/request-changes only)"
             echo "  -r, --repo     Explicit owner/repo slug (skips git-remote slug inference)"
             echo "  -H, --host     Explicit Gitea host (skips remote-host inference)"
             echo "  -h, --help     Show this help"
@@ -346,14 +346,7 @@ gitea_resolve_api_for_login() {
     else
         host=$(get_remote_host)
     fi
-    if [[ "$override_explicit" == "identity" ]]; then
-        # Requested git identity (#1280): the per-slot token MUST resolve via
-        # get_gitea_token's identity arm; never borrow the tea default login.
-        GITEA_API_TOKEN=$(get_gitea_token "$host") || {
-            echo "Error: could not resolve the per-slot token for requested git identity '$effective_login' on host '$host'; refusing to fall back to the tea login list or shared credentials (review write/read-back, #1280)." >&2
-            return 1
-        }
-    elif [[ -n "$override_explicit" ]]; then
+    if [[ -n "$override_explicit" ]]; then
         GITEA_API_TOKEN=$(get_gitea_token_for_login "$effective_login" "$host") || {
             echo "Error: could not resolve a host-matched Gitea token for --login '$effective_login' on host '$host'; refusing to fall back to the host default identity or a cross-host credential (review write/read-back)" >&2
             return 1
@@ -683,32 +676,29 @@ if [[ "$PLATFORM" == "github" ]]; then
             ;;
     esac
 elif [[ "$PLATFORM" == "gitea" ]]; then
-    # Resolve the acting principal ONCE for every action, identity-first
-    # (#1280): an explicit --login wins; otherwise MOSAIC_GIT_IDENTITY /
-    # per-worktree git config mosaic.gitIdentity selects the principal when a
-    # per-slot token exists (fail-loud when it does not); the tea login list is
-    # the LAST resort — it enumerates whatever logins this host happens to hold
-    # and knows nothing about which seat is calling, so resolving from it first
-    # wrote under whichever account tea had configured (the #1280 family).
-    principal_host="${HOST_OVERRIDE:-$(get_remote_host 2>/dev/null || true)}"
-    if ! principal_resolved="$(resolve_gitea_principal "$LOGIN_OVERRIDE" "$principal_host")"; then
-        # resolve_gitea_principal already printed the fail-loud diagnostic.
-        exit 1
-    fi
-    PRINCIPAL_MODE="$(printf '%s' "$principal_resolved" | cut -f1)"
-    PRINCIPAL_NAME="$(printf '%s' "$principal_resolved" | cut -f2)"
     case $ACTION in
         approve)
-            # Identity-first principal resolution (#1280): PRINCIPAL_MODE /
-            # PRINCIPAL_NAME were resolved once above from --login >
-            # MOSAIC_GIT_IDENTITY / git config > tea login list (last resort).
-            if [[ "$PRINCIPAL_MODE" == "identity" ]]; then
-                gitea_resolve_api_for_login "$PRINCIPAL_NAME" identity || exit 1
-            elif [[ "$PRINCIPAL_MODE" == "login" ]]; then
-                gitea_resolve_api_for_login "$PRINCIPAL_NAME" explicit || exit 1
-            else
-                gitea_resolve_api_for_login "$PRINCIPAL_NAME" "" || exit 1
-            fi
+            # Best-effort host for the tea-login GUESS only (gitea_resolve_api_for_login
+            # below re-derives the real host from HOST_OVERRIDE/remote independently and
+            # is authoritative). Prefer an explicit -H/--host; otherwise best-effort
+            # git-remote inference, tolerating its ABSENCE (a bare `get_remote_host` here
+            # under `set -e`, with no origin and no -H, previously killed the script
+            # SILENTLY — exit 1, zero output — even though -r/-H are exactly the flags
+            # that support running with no usable origin at all).
+            host="${HOST_OVERRIDE:-$(get_remote_host 2>/dev/null || true)}"
+            # A --login override always wins. Otherwise name this host's login
+            # only as a best effort: the login name merely selects a per-login
+            # token, and gitea_resolve_api_for_login falls back to the host
+            # credential (get_gitea_token) when no tea login is named — so a host
+            # tea's login list need not enumerate exotic (e.g. ported) hosts for
+            # the default credential to resolve. The single resolved token is
+            # then used for the write, the /user identity, and the read-back.
+            EFFECTIVE_LOGIN="$LOGIN_OVERRIDE"
+            [[ -n "$EFFECTIVE_LOGIN" ]] || EFFECTIVE_LOGIN=$(get_gitea_login_for_host "$host" 2>/dev/null || true)
+            # Bind the REST endpoint + token to the effective login, then derive
+            # the acting identity from that SAME credential so the review submit
+            # and its read-back verify against the identity that performed them.
+            gitea_resolve_api_for_login "$EFFECTIVE_LOGIN" "${LOGIN_OVERRIDE:+explicit}" || exit 1
             ACTING_LOGIN=$(gitea_authenticated_login) || exit 1
             head_sha=$(gitea_pr_head_sha "$PR_NUMBER") || exit 1
             # The review body (if any) travels with the review itself in the REST
@@ -725,16 +715,24 @@ elif [[ "$PLATFORM" == "gitea" ]]; then
                 echo "Error: Comment required for request-changes"
                 exit 1
             fi
-            # Identity-first principal resolution (#1280): PRINCIPAL_MODE /
-            # PRINCIPAL_NAME were resolved once above from --login >
-            # MOSAIC_GIT_IDENTITY / git config > tea login list (last resort).
-            if [[ "$PRINCIPAL_MODE" == "identity" ]]; then
-                gitea_resolve_api_for_login "$PRINCIPAL_NAME" identity || exit 1
-            elif [[ "$PRINCIPAL_MODE" == "login" ]]; then
-                gitea_resolve_api_for_login "$PRINCIPAL_NAME" explicit || exit 1
-            else
-                gitea_resolve_api_for_login "$PRINCIPAL_NAME" "" || exit 1
-            fi
+            # Best-effort host for the tea-login GUESS only (gitea_resolve_api_for_login
+            # below re-derives the real host from HOST_OVERRIDE/remote independently and
+            # is authoritative). Prefer an explicit -H/--host; otherwise best-effort
+            # git-remote inference, tolerating its ABSENCE (a bare `get_remote_host` here
+            # under `set -e`, with no origin and no -H, previously killed the script
+            # SILENTLY — exit 1, zero output — even though -r/-H are exactly the flags
+            # that support running with no usable origin at all).
+            host="${HOST_OVERRIDE:-$(get_remote_host 2>/dev/null || true)}"
+            # A --login override always wins. Otherwise name this host's login
+            # only as a best effort: the login name merely selects a per-login
+            # token, and gitea_resolve_api_for_login falls back to the host
+            # credential (get_gitea_token) when no tea login is named — so a host
+            # tea's login list need not enumerate exotic (e.g. ported) hosts for
+            # the default credential to resolve. The single resolved token is
+            # then used for the write, the /user identity, and the read-back.
+            EFFECTIVE_LOGIN="$LOGIN_OVERRIDE"
+            [[ -n "$EFFECTIVE_LOGIN" ]] || EFFECTIVE_LOGIN=$(get_gitea_login_for_host "$host" 2>/dev/null || true)
+            gitea_resolve_api_for_login "$EFFECTIVE_LOGIN" "${LOGIN_OVERRIDE:+explicit}" || exit 1
             ACTING_LOGIN=$(gitea_authenticated_login) || exit 1
             head_sha=$(gitea_pr_head_sha "$PR_NUMBER") || exit 1
             review_id=$(gitea_submit_review_verified "$PR_NUMBER" "REQUEST_CHANGES" "$COMMENT" "$ACTING_LOGIN" "$head_sha") || {
@@ -748,16 +746,24 @@ elif [[ "$PLATFORM" == "gitea" ]]; then
                 echo "Error: Comment required"
                 exit 1
             fi
-            # Identity-first principal resolution (#1280): PRINCIPAL_MODE /
-            # PRINCIPAL_NAME were resolved once above from --login >
-            # MOSAIC_GIT_IDENTITY / git config > tea login list (last resort).
-            if [[ "$PRINCIPAL_MODE" == "identity" ]]; then
-                gitea_resolve_api_for_login "$PRINCIPAL_NAME" identity || exit 1
-            elif [[ "$PRINCIPAL_MODE" == "login" ]]; then
-                gitea_resolve_api_for_login "$PRINCIPAL_NAME" explicit || exit 1
-            else
-                gitea_resolve_api_for_login "$PRINCIPAL_NAME" "" || exit 1
-            fi
+            # Best-effort host for the tea-login GUESS only (gitea_resolve_api_for_login
+            # below re-derives the real host from HOST_OVERRIDE/remote independently and
+            # is authoritative). Prefer an explicit -H/--host; otherwise best-effort
+            # git-remote inference, tolerating its ABSENCE (a bare `get_remote_host` here
+            # under `set -e`, with no origin and no -H, previously killed the script
+            # SILENTLY — exit 1, zero output — even though -r/-H are exactly the flags
+            # that support running with no usable origin at all).
+            host="${HOST_OVERRIDE:-$(get_remote_host 2>/dev/null || true)}"
+            # A --login override always wins. Otherwise name this host's login
+            # only as a best effort: the login name merely selects a per-login
+            # token, and gitea_resolve_api_for_login falls back to the host
+            # credential (get_gitea_token) when no tea login is named — so a host
+            # tea's login list need not enumerate exotic (e.g. ported) hosts for
+            # the default credential to resolve. The single resolved token is
+            # then used for the write, the /user identity, and the read-back.
+            EFFECTIVE_LOGIN="$LOGIN_OVERRIDE"
+            [[ -n "$EFFECTIVE_LOGIN" ]] || EFFECTIVE_LOGIN=$(get_gitea_login_for_host "$host" 2>/dev/null || true)
+            gitea_resolve_api_for_login "$EFFECTIVE_LOGIN" "${LOGIN_OVERRIDE:+explicit}" || exit 1
             ACTING_LOGIN=$(gitea_authenticated_login) || exit 1
             comment_id=$(gitea_create_comment_verified "$PR_NUMBER" "$COMMENT" "$ACTING_LOGIN") || {
                 echo "Error: could not create and verify a comment on Gitea PR #$PR_NUMBER via a provider-returned created id (#865)." >&2

packages/mosaic/framework/tools/git/test-gitea-principal-resolution.sh

@@ -1,255 +0,0 @@
-#!/usr/bin/env bash
-# Regression harness for detect-platform.sh's resolve_gitea_principal() — the
-# identity-first acting-principal resolution shared by the git wrappers
-# (mosaicstack/stack #1280).
-#
-# The contract under test (precedence: --login > MOSAIC_GIT_IDENTITY /
-# git config mosaic.gitIdentity > tea login list, which is the LAST resort):
-#   1. identity env + per-slot token present -> mode=identity, principal=
-#      identity name, source names the identity's slot PATH (never a token
-#      value).
-#   2. identity env + per-slot token ABSENT -> FAIL LOUD: nonzero, empty
-#      stdout, stderr naming the identity and the expected slot path.
-#   3. identity env + --login -> --login wins (login mode resolves even when
-#      the identity has no slot — operator intent beats environment).
-#   4. identity unset + no --login -> default mode: the tea login list
-#      resolves the principal exactly as before (preserved behavior).
-#   5. no identity + no host-matching tea login -> default/host-credential
-#      (preserved behavior; absence is not an error on the default path).
-#   6. identity on an UNRECOGNIZED host (no per-slot scheme) -> does not bind;
-#      default mode (containment, mirroring get_gitea_token).
-#   7. --login with no host-bound token for that login -> FAIL LOUD, stderr
-#      naming the login and the host.
-#   8. git config mosaic.gitIdentity is honored when the env var is unset.
-#   9. The resolver NEVER emits a token value — stdout/stderr of every
-#      successful resolution must not contain the slot file's contents.
-#
-# Uses a stubbed tea binary, stubbed tea config.yml, stubbed credentials.json
-# and stubbed per-slot token files under a fake HOME. NEVER reads real secrets.
-
-set -euo pipefail
-
-SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
-WORK_DIR="${MOSAIC_TEST_WORK_DIR:-$PWD/.mosaic-test-work/gitea-principal-resolution}"
-FAKE_HOME="$WORK_DIR/home"
-REPO_DIR="$WORK_DIR/repo"
-BIN_DIR="$WORK_DIR/bin"
-CREDENTIALS_FILE="$FAKE_HOME/.config/mosaic/credentials.json"
-
-rm -rf "$WORK_DIR"
-mkdir -p "$FAKE_HOME/.config/mosaic/secrets/gitea-tokens" "$FAKE_HOME/.config/tea" "$REPO_DIR" "$BIN_DIR"
-
-git -C "$REPO_DIR" init -q
-git -C "$REPO_DIR" remote add origin https://git.mosaicstack.dev/mosaicstack/stack.git
-
-cat > "$CREDENTIALS_FILE" <<'JSON'
-{
-  "gitea": {
-    "mosaicstack": {
-      "url": "https://git.mosaicstack.dev",
-      "token": "shared-mosaicstack-token"
-    },
-    "usc": {
-      "url": "https://git.uscllc.com",
-      "token": "shared-usc-token"
-    }
-  }
-}
-JSON
-
-# tea's own config store: the source get_gitea_token_for_login reads. Logins
-# "alice" (mosaicstack) and "bob-usc" (usc) carry sentinel token values that
-# the assertions prove are NEVER emitted by the resolver.
-cat > "$FAKE_HOME/.config/tea/config.yml" <<'YAML'
-logins:
-- name: alice
-  url: https://git.mosaicstack.dev
-  token: SECRET-alice-tea-token
-- name: bob-usc
-  url: https://git.uscllc.com
-  token: SECRET-bob-usc-tea-token
-YAML
-
-# Stubbed tea: only what login resolution needs (`login list --output json`).
-cat > "$BIN_DIR/tea" <<'SH'
-#!/usr/bin/env bash
-set -euo pipefail
-if [[ "$*" == "login list --output json" ]]; then
-    cat <<'JSON'
-[
-  {"name":"alice","url":"https://git.mosaicstack.dev","default":true},
-  {"name":"bob-usc","url":"https://git.uscllc.com"}
-]
-JSON
-    exit 0
-fi
-exit 0
-SH
-chmod +x "$BIN_DIR/tea"
-
-# Per-slot identity token with a sentinel value the assertions prove is never
-# emitted (proving "token came from the identity's slot BY PATH, not by value").
-echo -n "SECRET-agentX-slot-token" > "$FAKE_HOME/.config/mosaic/secrets/gitea-tokens/gitea-mosaicstack-agentX.token"
-
-fail=0
-assert_eq() {
-  local desc="$1" expected="$2" actual="$3"
-  if [[ "$expected" != "$actual" ]]; then
-    echo "FAIL: $desc — expected '$expected', got '$actual'" >&2
-    fail=1
-  fi
-}
-assert_contains() {
-  local desc="$1" haystack="$2" needle="$3"
-  if [[ "$haystack" != *"$needle"* ]]; then
-    echo "FAIL: $desc — missing '$needle' in: $haystack" >&2
-    fail=1
-  fi
-}
-assert_not_contains() {
-  local desc="$1" haystack="$2" needle="$3"
-  if [[ "$haystack" == *"$needle"* ]]; then
-    echo "FAIL: $desc — must not contain '$needle', got: $haystack" >&2
-    fail=1
-  fi
-}
-
-# Runs resolve_gitea_principal for $1=login_override $2=host inside REPO_DIR
-# (per-worktree git config resolves there) under a fake HOME, stubbed tea, and
-# stubbed credentials. Extra env (e.g. MOSAIC_GIT_IDENTITY) via $@.
-call_resolver() {
-  local login="$1" host="$2"; shift 2
-  (
-    cd "$REPO_DIR"
-    env -i HOME="$FAKE_HOME" PATH="$BIN_DIR:$PATH" \
-      GIT_CONFIG_GLOBAL=/dev/null GIT_CONFIG_SYSTEM=/dev/null \
-      MOSAIC_CREDENTIALS_FILE="$CREDENTIALS_FILE" \
-      DETECT_PLATFORM_SH="$SCRIPT_DIR/detect-platform.sh" "$@" \
-      bash -c 'source "$DETECT_PLATFORM_SH"; resolve_gitea_principal "$1" "$2"' _ "$login" "$host"
-  )
-}
-
-field() { printf '%s' "$1" | cut -f"$2"; }
-
-# ---------------------------------------------------------------------------
-# 1. Identity env + slot present -> identity mode, slot named BY PATH, and no
-#    token value ever emitted.
-# ---------------------------------------------------------------------------
-git -C "$REPO_DIR" config --unset mosaic.gitIdentity 2>/dev/null || true
-out=$(call_resolver "" "git.mosaicstack.dev" MOSAIC_GIT_IDENTITY=agentX)
-assert_eq "identity mode"            "identity" "$(field "$out" 1)"
-assert_eq "identity principal"       "agentX"   "$(field "$out" 2)"
-assert_eq "identity slot source" \
-  "identity-slot:$FAKE_HOME/.config/mosaic/secrets/gitea-tokens/gitea-mosaicstack-agentX.token" \
-  "$(field "$out" 3)"
-assert_not_contains "identity stdout leaks token" "$out" "SECRET"
-
-# ---------------------------------------------------------------------------
-# 2. Identity env + slot ABSENT -> fail loud: nonzero, empty stdout, stderr
-#    naming the identity and the expected slot path.
-# ---------------------------------------------------------------------------
-stderr_file="$WORK_DIR/stderr.tmp"
-set +e
-out=$(call_resolver "" "git.mosaicstack.dev" MOSAIC_GIT_IDENTITY=agentNoSlot 2>"$stderr_file")
-rc=$?
-set -e
-if [[ "$rc" -eq 0 ]]; then
-  echo "FAIL: missing slot — expected nonzero return, got 0 (stdout='$out')" >&2
-  fail=1
-fi
-if [[ -n "$out" ]]; then
-  echo "FAIL: missing slot — expected empty stdout, got '$out'" >&2
-  fail=1
-fi
-err=$(cat "$stderr_file")
-assert_contains "missing slot names identity"  "$err" "agentNoSlot"
-assert_contains "missing slot names slot path" "$err" \
-  "$FAKE_HOME/.config/mosaic/secrets/gitea-tokens/gitea-mosaicstack-agentNoSlot.token"
-assert_not_contains "missing-slot stderr leaks token" "$err" "SECRET"
-
-# ---------------------------------------------------------------------------
-# 3. Identity + --login -> --login wins. Also wins when the identity has NO
-#    slot (no identity check may veto an explicit login).
-# ---------------------------------------------------------------------------
-out=$(call_resolver "alice" "git.mosaicstack.dev" MOSAIC_GIT_IDENTITY=agentX)
-assert_eq "login beats identity (mode)"      "login" "$(field "$out" 1)"
-assert_eq "login beats identity (principal)" "alice" "$(field "$out" 2)"
-assert_eq "login source" "tea-login:alice" "$(field "$out" 3)"
-out=$(call_resolver "alice" "git.mosaicstack.dev" MOSAIC_GIT_IDENTITY=agentNoSlot)
-assert_eq "login beats slot-less identity" "login" "$(field "$out" 1)"
-
-# ---------------------------------------------------------------------------
-# 4. No identity, no --login -> default mode via the tea login list
-#    (preserved behavior).
-# ---------------------------------------------------------------------------
-out=$(call_resolver "" "git.mosaicstack.dev")
-assert_eq "default mode"          "default" "$(field "$out" 1)"
-assert_eq "default principal"     "alice"   "$(field "$out" 2)"
-assert_eq "default source"        "tea-default" "$(field "$out" 3)"
-
-# ---------------------------------------------------------------------------
-# 5. No identity, no --login, no host-matching tea login -> default with the
-#    host credential (absence is not an error on the default path).
-# ---------------------------------------------------------------------------
-out=$(call_resolver "" "git.unknown.test")
-assert_eq "no-match default mode"  "default" "$(field "$out" 1)"
-assert_eq "no-match default principal" "" "$(field "$out" 2)"
-assert_eq "no-match default source" "host-credential" "$(field "$out" 3)"
-
-# ---------------------------------------------------------------------------
-# 6. Identity on an UNRECOGNIZED host -> does not bind; default mode
-#    (containment, mirroring get_gitea_token's scope).
-# ---------------------------------------------------------------------------
-out=$(call_resolver "" "github.com" MOSAIC_GIT_IDENTITY=agentX)
-assert_eq "unrecognized host falls to default" "default" "$(field "$out" 1)"
-
-# ---------------------------------------------------------------------------
-# 7. --login with no host-bound token for that login -> fail loud, stderr
-#    naming the login and the host.
-# ---------------------------------------------------------------------------
-: > "$stderr_file"
-set +e
-out=$(call_resolver "ghost-login" "git.mosaicstack.dev" 2>"$stderr_file")
-rc=$?
-set -e
-if [[ "$rc" -eq 0 ]]; then
-  echo "FAIL: unknown --login — expected nonzero return, got 0 (stdout='$out')" >&2
-  fail=1
-fi
-err=$(cat "$stderr_file")
-assert_contains "unknown login names login" "$err" "ghost-login"
-assert_contains "unknown login names host"  "$err" "git.mosaicstack.dev"
-# A cross-host login (exists, but for usc) must ALSO fail loud for mosaicstack.
-set +e
-out=$(call_resolver "bob-usc" "git.mosaicstack.dev" 2>"$stderr_file")
-rc=$?
-set -e
-if [[ "$rc" -eq 0 ]]; then
-  echo "FAIL: cross-host --login — expected nonzero return, got 0" >&2
-  fail=1
-fi
-
-# ---------------------------------------------------------------------------
-# 8. git config mosaic.gitIdentity honored when env is unset.
-# ---------------------------------------------------------------------------
-git -C "$REPO_DIR" config mosaic.gitIdentity agentX
-out=$(call_resolver "" "git.mosaicstack.dev")
-assert_eq "git-config identity mode" "identity" "$(field "$out" 1)"
-assert_eq "git-config identity principal" "agentX" "$(field "$out" 2)"
-git -C "$REPO_DIR" config --unset mosaic.gitIdentity
-
-# ---------------------------------------------------------------------------
-# 9. Cross-host slot layout: the usc slot path is chosen for the usc host.
-# ---------------------------------------------------------------------------
-echo -n "SECRET-agentX-usc-token" > "$FAKE_HOME/.config/mosaic/secrets/gitea-tokens/gitea-usc-agentX.token"
-out=$(call_resolver "" "git.uscllc.com" MOSAIC_GIT_IDENTITY=agentX)
-assert_eq "usc identity mode" "identity" "$(field "$out" 1)"
-assert_eq "usc slot source" \
-  "identity-slot:$FAKE_HOME/.config/mosaic/secrets/gitea-tokens/gitea-usc-agentX.token" \
-  "$(field "$out" 3)"
-
-if [[ "$fail" -eq 0 ]]; then
-  echo "resolve_gitea_principal identity-first resolution regression passed"
-fi
-
-exit "$fail"

packages/mosaic/framework/tools/git/test-issue-create-body-safety.sh

@@ -77,30 +77,12 @@ exit 0
 SH
 chmod +x "$BIN_DIR/tea"
 
-# TRIPWIRE provider stub: this harness tests argv construction, so ANY curl
-# call is a failure of that contract (and, before this stub existed, a LIVE
-# write — the #1282–#1287 incident: the seat's real HOME leaked a global
-# mosaic.gitIdentity, flipping the wrapper into identity mode whose real
-# per-slot token created real issues on the forge). Fail loudly instead.
-cat > "$BIN_DIR/curl" <<'SH'
-#!/usr/bin/env bash
-echo "FAIL: body-safety harness reached a provider request — this test must never curl" >&2
-exit 99
-SH
-chmod +x "$BIN_DIR/curl"
-
-# Hermetic invocation: fake HOME (no credentials, no tea config, no token
-# slots) and GIT_CONFIG_GLOBAL severed — `git config --get mosaic.gitIdentity`
-# otherwise resolves the WORKSTATION's global identity (mos-dt-0 on the seat
-# that wrote this) and reroutes the wrapper into identity mode (#1280 family).
 (
     cd "$REPO_DIR"
-    env -i HOME="$WORK_DIR/home" PATH="$BIN_DIR:$PATH" \
-      GIT_CONFIG_GLOBAL=/dev/null GIT_CONFIG_SYSTEM=/dev/null \
-      MOSAIC_TEST_RECEIVED="$RECEIVED_FILE" \
-      "$SCRIPT_DIR/issue-create.sh" -t "Body safety test" -b "$BODY"
+    PATH="$BIN_DIR:$PATH" \
+    MOSAIC_TEST_RECEIVED="$RECEIVED_FILE" \
+    "$SCRIPT_DIR/issue-create.sh" -t "Body safety test" -b "$BODY"
 ) >/dev/null
-mkdir -p "$WORK_DIR/home"
 
 # 1. No command substitution executed anywhere in the pipeline.
 if [[ -e "$SENTINEL" ]]; then

packages/mosaic/framework/tools/git/test-issue-create-interactive-auth.sh

@@ -47,31 +47,14 @@ SH
 chmod +x "$BIN_DIR/tea" "$BIN_DIR/curl"
 
 run_wrapper() {
-  # Hermetic: fake HOME (fixture credentials only, no token slots, no tea
-  # config) and GIT_CONFIG_GLOBAL severed — `git config --get
-  # mosaic.gitIdentity` otherwise resolves the WORKSTATION's global identity
-  # and reroutes the wrapper into identity mode before the tea paths this
-  # harness exercises (#1280 family; see test-issue-create-body-safety.sh).
-  # An `env …` prefix (used for MOSAIC_TEA_STALE_USER) is re-wrapped, not
-  # doubled: arguments beginning with "env" are shifted past.
-  local env_pairs=()
-  if [[ "${1:-}" == "env" ]]; then
-    shift
-    while [[ "$#" -gt 0 && "$1" == *=* ]]; do
-      env_pairs+=("$1")
-      shift
-    done
-  fi
   (
     cd "$REPO_DIR"
-    env -i HOME="$WORK_DIR/home" PATH="$BIN_DIR:$PATH" \
-      GIT_CONFIG_GLOBAL=/dev/null GIT_CONFIG_SYSTEM=/dev/null \
-      MOSAIC_CREDENTIALS_FILE="$CREDENTIALS_FILE" \
-      MOSAIC_TEST_LOG="$LOG_FILE" "${env_pairs[@]}" \
-      "$@"
+    PATH="$BIN_DIR:$PATH" \
+    MOSAIC_CREDENTIALS_FILE="$CREDENTIALS_FILE" \
+    MOSAIC_TEST_LOG="$LOG_FILE" \
+    "$@"
   )
 }
-mkdir -p "$WORK_DIR/home"
 
 : > "$LOG_FILE"
 printf 'Interactive title\nInteractive body\nlabel-a,label-b\nM1\n' | run_wrapper "$SCRIPT_DIR/issue-create.sh" -i >/dev/null

packages/mosaic/framework/tools/git/test-pr-create-identity-first.sh

@@ -1,244 +0,0 @@
-#!/usr/bin/env bash
-# Load-bearing regression harness for pr-create.sh identity-first principal
-# resolution (mosaicstack/stack #1280).
-#
-# The failure this harness is written down to catch: `MOSAIC_GIT_IDENTITY=fargo
-# pr-create.sh …` produces a PR attributed to `mos-dt-0` (whichever account the
-# tea login list happens to hold). Before #1280 the identity-aware code existed
-# but sat on the API arm that only ran when the tea path FAILED — tea succeeded,
-# so the identity arm never executed, and every test that did not check ORDERING
-# passed. This harness checks ordering directly:
-#
-#   1. identity set + slot present -> the PR is created via the REST API with
-#      the identity's per-slot token (asserted by sentinel value AT the fake
-#      provider), and tea's `pr create` is NEVER invoked.
-#   2. identity set + slot ABSENT -> nonzero, stderr naming the identity and
-#      the expected slot path; neither tea `pr create` nor any API request
-#      fires. No silent fallback to the tea login list.
-#   3. identity set + --login -> --login wins: tea runs WITH the explicit
-#      --login, no API request.
-#   4. nothing set -> preserved behavior: tea path with the tea-list login.
-#
-# Uses a stubbed tea, a stubbed curl provider, stubbed credentials.json and
-# per-slot token under a fake HOME. NEVER reads real secrets or hits a live
-# forge — all assertions are against the stubs' logs.
-
-set -euo pipefail
-
-SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
-WORK_DIR="${MOSAIC_TEST_WORK_DIR:-$PWD/.mosaic-test-work/pr-create-identity-first}"
-FAKE_HOME="$WORK_DIR/home"
-REPO_DIR="$WORK_DIR/repo"
-TOOLS_DIR="$WORK_DIR/tools"
-BIN_DIR="$WORK_DIR/bin"
-LOG_FILE="$WORK_DIR/calls.log"
-CREDENTIALS_FILE="$FAKE_HOME/.config/mosaic/credentials.json"
-
-rm -rf "$WORK_DIR"
-mkdir -p "$FAKE_HOME/.config/mosaic/secrets/gitea-tokens" "$FAKE_HOME/.config/tea" \
-  "$REPO_DIR" "$TOOLS_DIR/git" "$TOOLS_DIR/_lib" "$BIN_DIR"
-
-# Fixture: the real scripts under test, copied so sibling stubs (and the
-# ../_lib credential loader) resolve inside the fixture tree.
-cp "$SCRIPT_DIR/pr-create.sh" "$TOOLS_DIR/git/pr-create.sh"
-cp "$SCRIPT_DIR/detect-platform.sh" "$TOOLS_DIR/git/detect-platform.sh"
-cp "$SCRIPT_DIR/../_lib/credentials.sh" "$TOOLS_DIR/_lib/credentials.sh"
-chmod +x "$TOOLS_DIR/git/pr-create.sh"
-
-git -C "$REPO_DIR" init -q
-git -C "$REPO_DIR" remote add origin https://git.mosaicstack.dev/mosaicstack/stack.git
-
-cat > "$CREDENTIALS_FILE" <<'JSON'
-{
-  "gitea": {
-    "mosaicstack": {
-      "url": "https://git.mosaicstack.dev",
-      "token": "shared-mosaicstack-token"
-    }
-  }
-}
-JSON
-
-cat > "$FAKE_HOME/.config/tea/config.yml" <<'YAML'
-logins:
-- name: alice
-  url: https://git.mosaicstack.dev
-  token: SECRET-alice-tea-token
-YAML
-
-echo -n "SECRET-agentX-slot-token" > "$FAKE_HOME/.config/mosaic/secrets/gitea-tokens/gitea-mosaicstack-agentX.token"
-
-: > "$LOG_FILE"
-
-# Stubbed tea: records every invocation; `login list` feeds login resolution;
-# `api --login <n> /user` feeds get_gitea_authenticated_user; `pr create` marks
-# the marker file (its presence fails the identity-mode assertions).
-cat > "$BIN_DIR/tea" <<SH
-#!/usr/bin/env bash
-set -euo pipefail
-printf 'TEA: %s\n' "\$*" >> "$LOG_FILE"
-if [[ "\$*" == "login list --output json" ]]; then
-    cat <<'JSON'
-[
-  {"name":"alice","url":"https://git.mosaicstack.dev","default":true}
-]
-JSON
-    exit 0
-fi
-if [[ "\${1:-}" == "api" ]]; then
-    printf '%s\n' '{"login":"alice"}'
-    exit 0
-fi
-if [[ "\$*" == pr\ create* ]]; then
-    echo "TEA-PR-CREATE-INVOKED" >> "$LOG_FILE"
-    exit 0
-fi
-exit 0
-SH
-chmod +x "$BIN_DIR/tea"
-
-# Stubbed provider: records the URL and the Authorization header VALUE it
-# received, answers 201 with a created-PR object. The sentinel token values are
-# synthetic fixtures — asserting them at the provider proves WHICH slot's
-# credential carried the write.
-cat > "$BIN_DIR/curl" <<SH
-#!/usr/bin/env bash
-set -euo pipefail
-url=""
-auth=""
-while [[ \$# -gt 0 ]]; do
-    case "\$1" in
-        -H)
-            case "\$2" in
-                Authorization*) auth="\$2" ;;
-            esac
-            shift 2
-            ;;
-        *)  [[ -n "\$1" && "\$1" != -* ]] && url="\$1"
-            shift
-            ;;
-    esac
-done
-printf 'CURL-URL: %s\nCURL-AUTH: %s\n' "\$url" "\$auth" >> "$LOG_FILE"
-cat <<'JSON'
-{"number": 1299, "html_url": "https://git.mosaicstack.dev/mosaicstack/stack/pulls/1299"}
-JSON
-exit 0
-SH
-chmod +x "$BIN_DIR/curl"
-
-fail=0
-assert_contains() {
-  local desc="$1" needle="$2"
-  if ! grep -qF -- "$needle" "$LOG_FILE"; then
-    echo "FAIL: $desc — log does not contain '$needle':" >&2
-    cat "$LOG_FILE" >&2
-    fail=1
-  fi
-}
-assert_not_contains() {
-  local desc="$1" needle="$2"
-  if grep -qF -- "$needle" "$LOG_FILE"; then
-    echo "FAIL: $desc — log must not contain '$needle':" >&2
-    cat "$LOG_FILE" >&2
-    fail=1
-  fi
-}
-
-EXTRA_ARGS=""
-run_pr_create() {
-  # "$@" carries ONLY environment assignments (VAR=value); EXTRA_ARGS (if
-  # set) carries wrapper arguments, so `env` never mistakes a wrapper flag
-  # like --login for one of its own.
-  (
-    cd "$REPO_DIR"
-    # shellcheck disable=SC2086 # EXTRA_ARGS is deliberately word-split wrapper args
-    env -i HOME="$FAKE_HOME" PATH="$BIN_DIR:$PATH" \
-      GIT_CONFIG_GLOBAL=/dev/null GIT_CONFIG_SYSTEM=/dev/null \
-      MOSAIC_CREDENTIALS_FILE="$CREDENTIALS_FILE" "$@" \
-      bash "$TOOLS_DIR/git/pr-create.sh" -t "Test PR" -B next -H fix/test $EXTRA_ARGS
-  )
-}
-
-# ---------------------------------------------------------------------------
-# 1. HAPPY PATH (the load-bearing ordering test): identity set + slot present
-#    -> REST API with the per-slot token; tea `pr create` NEVER invoked.
-# ---------------------------------------------------------------------------
-set +e
-out=$(run_pr_create MOSAIC_GIT_IDENTITY=agentX 2>"$WORK_DIR/stderr-1.tmp")
-rc=$?
-set -e
-if [[ "$rc" -ne 0 ]]; then
-  echo "FAIL: identity happy path — expected rc=0, got $rc" >&2
-  cat "$WORK_DIR/stderr-1.tmp" >&2
-  fail=1
-fi
-assert_contains "identity happy path reaches the API" "CURL-URL: https://git.mosaicstack.dev/api/v1/repos/mosaicstack/stack/pulls"
-assert_contains "identity happy path carries the slot token" "CURL-AUTH: Authorization: token SECRET-agentX-slot-token"
-assert_not_contains "identity happy path must NOT invoke tea pr create" "TEA-PR-CREATE-INVOKED"
-
-# ---------------------------------------------------------------------------
-# 2. Identity set + slot ABSENT -> fail loud BEFORE any write: nonzero, stderr
-#    naming identity + slot path, no tea pr create, no API request.
-# ---------------------------------------------------------------------------
-: > "$LOG_FILE"
-set +e
-out=$(run_pr_create MOSAIC_GIT_IDENTITY=agentNoSlot 2>"$WORK_DIR/stderr-2.tmp")
-rc=$?
-set -e
-if [[ "$rc" -eq 0 ]]; then
-  echo "FAIL: missing slot — expected nonzero return, got 0 (stdout='$out')" >&2
-  fail=1
-fi
-err=$(cat "$WORK_DIR/stderr-2.tmp")
-if [[ "$err" != *"agentNoSlot"* ]]; then
-  echo "FAIL: missing slot — stderr does not name the identity:" >&2
-  echo "$err" >&2
-  fail=1
-fi
-if [[ "$err" != *"$FAKE_HOME/.config/mosaic/secrets/gitea-tokens/gitea-mosaicstack-agentNoSlot.token"* ]]; then
-  echo "FAIL: missing slot — stderr does not name the expected slot path:" >&2
-  echo "$err" >&2
-  fail=1
-fi
-assert_not_contains "missing slot must not reach tea pr create" "TEA-PR-CREATE-INVOKED"
-assert_not_contains "missing slot must not reach the API" "CURL-URL"
-
-# ---------------------------------------------------------------------------
-# 3. Identity set + --login -> --login wins: tea runs WITH the explicit login.
-# ---------------------------------------------------------------------------
-: > "$LOG_FILE"
-EXTRA_ARGS="--login alice"
-set +e
-out=$(run_pr_create MOSAIC_GIT_IDENTITY=agentX 2>"$WORK_DIR/stderr-3.tmp")
-rc=$?
-set -e
-EXTRA_ARGS=""
-if [[ "$rc" -ne 0 ]]; then
-  echo "FAIL: login override — expected rc=0, got $rc" >&2
-  cat "$WORK_DIR/stderr-3.tmp" >&2
-  fail=1
-fi
-assert_contains "login override drives tea with the explicit login" "TEA: pr create --repo mosaicstack/stack --login alice"
-assert_not_contains "login override must not hit the API" "CURL-URL"
-
-# ---------------------------------------------------------------------------
-# 4. Nothing set -> preserved behavior: tea path with the tea-list login.
-# ---------------------------------------------------------------------------
-: > "$LOG_FILE"
-set +e
-out=$(run_pr_create 2>"$WORK_DIR/stderr-4.tmp")
-rc=$?
-set -e
-if [[ "$rc" -ne 0 ]]; then
-  echo "FAIL: default path — expected rc=0, got $rc" >&2
-  cat "$WORK_DIR/stderr-4.tmp" >&2
-  fail=1
-fi
-assert_contains "default path still uses the tea-list login" "TEA: pr create --repo mosaicstack/stack --login alice"
-
-if [[ "$fail" -eq 0 ]]; then
-  echo "pr-create identity-first happy-path regression passed"
-fi
-
-exit "$fail"

packages/mosaic/framework/tools/git/test-pr-merge-principal-resolution.sh

@@ -1,247 +0,0 @@
-#!/usr/bin/env bash
-# Regression harness for pr-merge.sh identity-first principal resolution
-# (mosaicstack/stack #1280).
-#
-# Covers:
-#   1. --dry-run reports the acting principal the merge WOULD use, resolved the
-#      same way the real merge resolves it: --login > MOSAIC_GIT_IDENTITY /
-#      git config mosaic.gitIdentity > shared host credential. (The pre-#1280
-#      deployed copy reported a tea login that the merge would not act as.)
-#   2. --dry-run fails closed when the requested principal has no credential:
-#      unknown --login, or an identity with no per-slot token (stderr names
-#      the login / the identity and its slot path).
-#   3. The real merge POST carries the resolved principal's credential and no
-#      other: --login merges with that login's tea-config token; an identity
-#      merges with the per-slot token; an unresolvable --login never reaches
-#      the provider.
-#
-# Fixture pattern from test-pr-merge-head-pin.sh: the scripts under test are
-# copied into a fixture tree with stubbed pr-metadata.sh / ci-queue-wait.sh
-# siblings; the provider is a stubbed curl that records the credential it
-# received. NEVER reads real secrets or hits a live forge.
-
-set -euo pipefail
-
-SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
-WORK_DIR="${MOSAIC_TEST_WORK_DIR:-$PWD/.mosaic-test-work/pr-merge-principal-resolution}"
-FAKE_HOME="$WORK_DIR/home"
-REPO_DIR="$WORK_DIR/repo"
-TOOLS_DIR="$WORK_DIR/tools"
-BIN_DIR="$WORK_DIR/bin"
-LOG_FILE="$WORK_DIR/calls.log"
-CREDENTIALS_FILE="$FAKE_HOME/.config/mosaic/credentials.json"
-SHA=0123456789abcdef0123456789abcdef01234567
-
-rm -rf "$WORK_DIR"
-mkdir -p "$FAKE_HOME/.config/mosaic/secrets/gitea-tokens" "$FAKE_HOME/.config/tea" \
-  "$REPO_DIR" "$TOOLS_DIR/git" "$TOOLS_DIR/_lib" "$BIN_DIR"
-
-cp "$SCRIPT_DIR/pr-merge.sh" "$TOOLS_DIR/git/pr-merge.sh"
-cp "$SCRIPT_DIR/detect-platform.sh" "$TOOLS_DIR/git/detect-platform.sh"
-cp "$SCRIPT_DIR/../_lib/credentials.sh" "$TOOLS_DIR/_lib/credentials.sh"
-chmod +x "$TOOLS_DIR/git/pr-merge.sh"
-
-git -C "$REPO_DIR" init -q
-git -C "$REPO_DIR" remote add origin https://git.mosaicstack.dev/mosaicstack/stack.git
-
-# Stubbed siblings pr-merge.sh resolves relative to its own SCRIPT_DIR.
-cat > "$TOOLS_DIR/git/pr-metadata.sh" <<SH
-#!/usr/bin/env bash
-printf '%s\n' '{"baseRefName":"next","headRefName":"fix/pinned","headRefOid":"$SHA","headRepository":"mosaicstack/stack","title":"Test PR","author":{"login":"contributor"}}'
-SH
-cat > "$TOOLS_DIR/git/ci-queue-wait.sh" <<'SH'
-#!/usr/bin/env bash
-exit 0
-SH
-chmod +x "$TOOLS_DIR/git/pr-metadata.sh" "$TOOLS_DIR/git/ci-queue-wait.sh"
-
-cat > "$CREDENTIALS_FILE" <<'JSON'
-{
-  "gitea": {
-    "mosaicstack": {
-      "url": "https://git.mosaicstack.dev",
-      "token": "shared-mosaicstack-token"
-    }
-  }
-}
-JSON
-
-cat > "$FAKE_HOME/.config/tea/config.yml" <<'YAML'
-logins:
-- name: fred-ms
-  url: https://git.mosaicstack.dev
-  token: SECRET-fred-ms-tea-token
-YAML
-
-echo -n "SECRET-agentX-slot-token" > "$FAKE_HOME/.config/mosaic/secrets/gitea-tokens/gitea-mosaicstack-agentX.token"
-
-: > "$LOG_FILE"
-
-# Stubbed tea for login-list resolution only.
-cat > "$BIN_DIR/tea" <<'SH'
-#!/usr/bin/env bash
-set -euo pipefail
-if [[ "$*" == "login list --output json" ]]; then
-    cat <<'JSON'
-[
-  {"name":"fred-ms","url":"https://git.mosaicstack.dev","default":true}
-]
-JSON
-    exit 0
-fi
-exit 0
-SH
-chmod +x "$BIN_DIR/tea"
-
-# Stubbed provider. pr-merge passes curl config on STDIN with -K -; the stub
-# reads stdin, records the Authorization header it received, answers 200.
-cat > "$BIN_DIR/curl" <<SH
-#!/usr/bin/env bash
-set -euo pipefail
-url=""
-out_file=""
-stdin_config=""
-if [[ ! -t 0 ]]; then
-    stdin_config="\$(cat || true)"
-fi
-while [[ \$# -gt 0 ]]; do
-    case "\$1" in
-        -o) out_file="\$2"; shift 2 ;;
-        -K|-w|--max-filesize|--max-time|--connect-timeout|-sS) shift 2 ;;
-        *)  [[ -n "\$1" && "\$1" != -* && -z "\$url" ]] && url="\$1"
-            shift
-            ;;
-    esac
-done
-auth="\$(printf '%s' "\$stdin_config" | grep -o 'Authorization: token [^"]*' || true)"
-printf 'CURL-URL: %s\nCURL-AUTH: %s\n' "\$url" "\$auth" >> "$LOG_FILE"
-[[ -n "\$out_file" ]] && printf '{}' > "\$out_file"
-printf '200\n'
-exit 0
-SH
-chmod +x "$BIN_DIR/curl"
-
-fail=0
-assert_contains_log() {
-  local desc="$1" needle="$2"
-  if ! grep -qF -- "$needle" "$LOG_FILE"; then
-    echo "FAIL: $desc — log does not contain '$needle':" >&2
-    cat "$LOG_FILE" >&2
-    fail=1
-  fi
-}
-assert_not_contains_log() {
-  local desc="$1" needle="$2"
-  if grep -qF -- "$needle" "$LOG_FILE"; then
-    echo "FAIL: $desc — log must not contain '$needle':" >&2
-    cat "$LOG_FILE" >&2
-    fail=1
-  fi
-}
-
-run_pr_merge() {
-  local extra_args="$1"; shift
-  (
-    cd "$REPO_DIR"
-    # shellcheck disable=SC2086 # extra_args is deliberately word-split wrapper args
-    env -i HOME="$FAKE_HOME" PATH="$BIN_DIR:$PATH" \
-      GIT_CONFIG_GLOBAL=/dev/null GIT_CONFIG_SYSTEM=/dev/null \
-      MOSAIC_CREDENTIALS_FILE="$CREDENTIALS_FILE" "$@" \
-      bash "$TOOLS_DIR/git/pr-merge.sh" -n 42 $extra_args
-  )
-}
-
-# ---------------------------------------------------------------------------
-# 1. --dry-run reports the resolved acting principal truthfully.
-# ---------------------------------------------------------------------------
-out=$(run_pr_merge "--dry-run" MOSAIC_GIT_IDENTITY=agentX)
-if [[ "$out" != *"as git identity 'agentX' (per-slot credential)"* ]]; then
-  echo "FAIL: dry-run identity — principal not reported: $out" >&2
-  fail=1
-fi
-
-out=$(run_pr_merge "--dry-run --login fred-ms" MOSAIC_GIT_IDENTITY=agentX)
-if [[ "$out" != *"as tea login 'fred-ms'"* ]]; then
-  echo "FAIL: dry-run login override — login not reported (must beat env identity): $out" >&2
-  fail=1
-fi
-
-out=$(run_pr_merge "--dry-run")
-if [[ "$out" != *"as default host credential"* ]]; then
-  echo "FAIL: dry-run default — not reported: $out" >&2
-  fail=1
-fi
-
-# ---------------------------------------------------------------------------
-# 2. --dry-run fails closed when the requested principal has no credential.
-# ---------------------------------------------------------------------------
-stderr_file="$WORK_DIR/stderr.tmp"
-set +e
-out=$(run_pr_merge "--dry-run --login ghost" 2>"$stderr_file")
-rc=$?
-set -e
-if [[ "$rc" -eq 0 ]] || [[ "$(cat "$stderr_file")" != *"ghost"* ]]; then
-  echo "FAIL: dry-run unknown --login — expected fail-loud naming 'ghost', rc=$rc" >&2
-  cat "$stderr_file" >&2
-  fail=1
-fi
-
-: > "$stderr_file"
-set +e
-out=$(run_pr_merge "--dry-run" MOSAIC_GIT_IDENTITY=agentNoSlot 2>"$stderr_file")
-rc=$?
-set -e
-err=$(cat "$stderr_file")
-if [[ "$rc" -eq 0 ]] || [[ "$err" != *"agentNoSlot"* ]] \
-  || [[ "$err" != *"$FAKE_HOME/.config/mosaic/secrets/gitea-tokens/gitea-mosaicstack-agentNoSlot.token"* ]]; then
-  echo "FAIL: dry-run identity without slot — expected fail-loud naming identity + slot path, rc=$rc" >&2
-  echo "$err" >&2
-  fail=1
-fi
-
-# ---------------------------------------------------------------------------
-# 3. The real merge POST carries the resolved principal's credential ONLY.
-# ---------------------------------------------------------------------------
-: > "$LOG_FILE"
-set +e
-out=$(run_pr_merge "--login fred-ms" MOSAIC_GIT_IDENTITY=agentX 2>"$stderr_file")
-rc=$?
-set -e
-if [[ "$rc" -ne 0 ]]; then
-  echo "FAIL: merge with --login — expected rc=0, got $rc" >&2
-  cat "$stderr_file" >&2
-  fail=1
-fi
-assert_contains_log "merge --login uses the login token" "CURL-AUTH: Authorization: token SECRET-fred-ms-tea-token"
-assert_not_contains_log "merge --login must not use the identity slot token" "SECRET-agentX-slot-token"
-assert_not_contains_log "merge --login must not use the shared token" "shared-mosaicstack-token"
-
-: > "$LOG_FILE"
-set +e
-out=$(run_pr_merge "" MOSAIC_GIT_IDENTITY=agentX 2>"$stderr_file")
-rc=$?
-set -e
-if [[ "$rc" -ne 0 ]]; then
-  echo "FAIL: merge with identity — expected rc=0, got $rc" >&2
-  cat "$stderr_file" >&2
-  fail=1
-fi
-assert_contains_log "merge identity uses the per-slot token" "CURL-AUTH: Authorization: token SECRET-agentX-slot-token"
-assert_not_contains_log "merge identity must not use the shared token" "shared-mosaicstack-token"
-
-: > "$LOG_FILE"
-set +e
-out=$(run_pr_merge "--login ghost" 2>"$stderr_file")
-rc=$?
-set -e
-if [[ "$rc" -eq 0 ]]; then
-  echo "FAIL: merge with unknown --login — expected nonzero, got 0" >&2
-  fail=1
-fi
-assert_not_contains_log "merge with unknown --login must not reach the provider" "CURL-URL"
-
-if [[ "$fail" -eq 0 ]]; then
-  echo "pr-merge identity-first principal resolution regression passed"
-fi
-
-exit "$fail"

packages/mosaic/package.json

@@ -25,7 +25,7 @@
     "lint": "eslint src",
     "typecheck": "tsc --noEmit",
     "test": "vitest run --passWithNoTests && pnpm run test:framework-shell",
-    "test:framework-shell": "bash framework/tools/quality/scripts/check-test-enumeration.sh && bash framework/tools/quality/scripts/test-check-test-enumeration.sh && python3 src/lease-broker/daemon_deadline_unittest.py && python3 src/lease-broker/normative_fragments_unittest.py && python3 src/lease-broker/promotion_binding_unittest.py && python3 src/lease-broker/promotion_trigger_unittest.py && python3 src/lease-broker/receipt_challenge_unittest.py && python3 src/lease-broker/context_recovery_unittest.py && python3 src/lease-broker/recovery_runtime_unittest.py && python3 src/lease-broker/recovery_b1_adversarial_unittest.py && python3 src/lease-broker/receipt_observer_client_unittest.py && python3 src/lease-broker/invariant_r_unittest.py && python3 src/lease-broker/framework_skill_portability_unittest.py && python3 src/mutator-gate/runtime_tools_unittest.py && python3 src/mutator-gate/runtime_launch_guard_unittest.py && python3 src/mutator-gate/version_coupling_unittest.py && python3 framework/tools/lease-broker/check-runtime-launches.py --root ../.. && bash framework/tools/codex/test-pr-diff-context.sh && bash framework/tools/qa/test-deps-preflight.sh && bash framework/tools/git/test-pr-review-gitea-comment.sh && bash framework/tools/git/test-pr-review-repo-host-override.sh && bash framework/tools/git/test-ci-queue-wait-branch-absent.sh && bash framework/tools/git/test-ci-queue-wait-tristate.sh && bash framework/tools/git/test-ci-queue-wait-github-checks.sh && bash framework/tools/git/test-pr-merge-queue-branch.sh && bash framework/tools/git/test-pr-merge-head-pin.sh && bash framework/tools/git/test-pr-merge-message-field.sh && bash framework/tools/git/test-git-credential-mosaic.sh && bash framework/tools/git/test-gitea-token-identity.sh && bash framework/tools/git/test-gitea-principal-resolution.sh && bash framework/tools/git/test-pr-create-identity-first.sh && bash framework/tools/git/test-pr-merge-principal-resolution.sh && bash framework/tools/woodpecker/test-terminal-green-contract.sh && bash framework/tools/_scripts/test-install-ordering-guard.sh && bash framework/tools/_scripts/test-mosaic-init-rce.sh && bash framework/tools/tmux/agent-send.test.sh && bash framework/tools/wake/test-wake-store-ack.sh && bash framework/tools/wake/test-wake-store-enqueue-race.sh && bash framework/tools/wake/test-wake-digest-hmac.sh && bash framework/tools/wake/test-wake-digest-quarantine.sh && bash framework/tools/wake/test-wake-detector.sh && bash framework/tools/wake/test-wake-fn-oracle.sh && bash framework/tools/wake/test-wake-reconcile.sh && bash framework/tools/wake/test-wake-beacon.sh && bash framework/tools/wake/test-wake-preimage.sh && bash framework/tools/wake/test-wake-install.sh && bash framework/tools/glpi/test-list-http-status.sh && bash framework/tools/orchestrator/test-board-roll.sh && bash framework/tools/woodpecker/test-ci-wait-exit-matrix.sh && bash framework/tools/_scripts/test-fleet-transport-check.sh"
+    "test:framework-shell": "bash framework/tools/quality/scripts/check-test-enumeration.sh && bash framework/tools/quality/scripts/test-check-test-enumeration.sh && python3 src/lease-broker/daemon_deadline_unittest.py && python3 src/lease-broker/normative_fragments_unittest.py && python3 src/lease-broker/promotion_binding_unittest.py && python3 src/lease-broker/promotion_trigger_unittest.py && python3 src/lease-broker/receipt_challenge_unittest.py && python3 src/lease-broker/context_recovery_unittest.py && python3 src/lease-broker/recovery_runtime_unittest.py && python3 src/lease-broker/recovery_b1_adversarial_unittest.py && python3 src/lease-broker/receipt_observer_client_unittest.py && python3 src/lease-broker/invariant_r_unittest.py && python3 src/lease-broker/framework_skill_portability_unittest.py && python3 src/mutator-gate/runtime_tools_unittest.py && python3 src/mutator-gate/runtime_launch_guard_unittest.py && python3 src/mutator-gate/version_coupling_unittest.py && python3 framework/tools/lease-broker/check-runtime-launches.py --root ../.. && bash framework/tools/codex/test-pr-diff-context.sh && bash framework/tools/qa/test-deps-preflight.sh && bash framework/tools/git/test-pr-review-gitea-comment.sh && bash framework/tools/git/test-pr-review-repo-host-override.sh && bash framework/tools/git/test-ci-queue-wait-branch-absent.sh && bash framework/tools/git/test-ci-queue-wait-tristate.sh && bash framework/tools/git/test-ci-queue-wait-github-checks.sh && bash framework/tools/git/test-pr-merge-queue-branch.sh && bash framework/tools/git/test-pr-merge-head-pin.sh && bash framework/tools/git/test-pr-merge-message-field.sh && bash framework/tools/git/test-git-credential-mosaic.sh && bash framework/tools/git/test-gitea-token-identity.sh && bash framework/tools/woodpecker/test-terminal-green-contract.sh && bash framework/tools/_scripts/test-install-ordering-guard.sh && bash framework/tools/_scripts/test-mosaic-init-rce.sh && bash framework/tools/tmux/agent-send.test.sh && bash framework/tools/wake/test-wake-store-ack.sh && bash framework/tools/wake/test-wake-store-enqueue-race.sh && bash framework/tools/wake/test-wake-digest-hmac.sh && bash framework/tools/wake/test-wake-digest-quarantine.sh && bash framework/tools/wake/test-wake-detector.sh && bash framework/tools/wake/test-wake-fn-oracle.sh && bash framework/tools/wake/test-wake-reconcile.sh && bash framework/tools/wake/test-wake-beacon.sh && bash framework/tools/wake/test-wake-preimage.sh && bash framework/tools/wake/test-wake-install.sh && bash framework/tools/glpi/test-list-http-status.sh && bash framework/tools/orchestrator/test-board-roll.sh && bash framework/tools/woodpecker/test-ci-wait-exit-matrix.sh && bash framework/tools/_scripts/test-fleet-transport-check.sh && bash framework/tools/_scripts/test-brain-home-check.sh"
   },
   "dependencies": {
     "@mosaicstack/brain": "workspace:*",

packages/mosaic/src/commands/fleet-agent-crud-command.ts

@@ -1,5 +1,6 @@
 import { readFile } from 'node:fs/promises';
 import { join, resolve } from 'node:path';
+import { fleetAgentEnvDir, fleetRolesLocalDir } from '../fleet/brain-home.js';
 import type { Command } from 'commander';
 import {
   executeFleetAgentMutation,
@@ -149,9 +150,9 @@ async function executeCommand(
     request,
     mosaicHome,
     rosterPath,
-    agentEnvDir: join(mosaicHome, 'fleet', 'agents'),
+    agentEnvDir: fleetAgentEnvDir(mosaicHome),
     rolesDir: join(mosaicHome, 'fleet', 'roles'),
-    overrideDir: join(mosaicHome, 'fleet', 'roles.local'),
+    overrideDir: fleetRolesLocalDir(mosaicHome),
     dryRun: forceDryRun || opts.dryRun === true,
     ...(deps.projectionApplier === undefined ? {} : { projectionApplier: deps.projectionApplier }),
   });

packages/mosaic/src/commands/fleet-migration-command.ts

@@ -1,5 +1,6 @@
 import { readFile } from 'node:fs/promises';
 import { join } from 'node:path';
+import { fleetAgentEnvDir, fleetRolesLocalDir } from '../fleet/brain-home.js';
 import type { Command } from 'commander';
 import {
   parseV1MigrationObservations,
@@ -120,11 +121,11 @@ export function registerFleetMigrationCommand(
           observations,
           personaDirs: {
             rolesDir: deps.rolesDir ?? join(mosaicHome, 'fleet', 'roles'),
-            overrideDir: deps.overrideDir ?? join(mosaicHome, 'fleet', 'roles.local'),
+            overrideDir: deps.overrideDir ?? fleetRolesLocalDir(mosaicHome),
           },
           environment: {
             mosaicHome,
-            agentEnvDir: join(mosaicHome, 'fleet', 'agents'),
+            agentEnvDir: fleetAgentEnvDir(mosaicHome),
           },
         });
         printJson(preview);

packages/mosaic/src/commands/fleet-personas.ts

@@ -30,19 +30,21 @@ import { lstat, readFile, readdir, stat } from 'node:fs/promises';
 import { homedir } from 'node:os';
 import { basename, isAbsolute, join, sep } from 'node:path';
 import type { Command } from 'commander';
+import { fleetRolesLocalDir } from '../fleet/brain-home.js';
 
 function defaultMosaicHome(): string {
   return process.env['MOSAIC_HOME'] ?? join(homedir(), '.config', 'mosaic');
 }
 
-/** Baseline persona role contracts (reseeded on update). */
+/** Baseline persona role contracts (reseeded on update; config home — framework). */
 export function defaultRolesDir(mosaicHome = defaultMosaicHome()): string {
   return join(mosaicHome, 'fleet', 'roles');
 }
 
-/** PRESERVE-protected override layer (survives update; wins on merge). */
+/** PRESERVE-protected override layer (survives update; wins on merge).
+ * Brain home (`~/.mosaic/fleet/roles.local`) when a brain is active. */
 export function defaultOverrideDir(mosaicHome = defaultMosaicHome()): string {
-  return join(mosaicHome, 'fleet', 'roles.local');
+  return fleetRolesLocalDir(mosaicHome);
 }
 
 /**

packages/mosaic/src/commands/fleet-profiles.ts

@@ -25,6 +25,7 @@ import { homedir } from 'node:os';
 import { basename, join } from 'node:path';
 import type { Command } from 'commander';
 import YAML from 'yaml';
+import { fleetProfilesDir } from '../fleet/brain-home.js';
 import {
   defaultOverrideDir,
   extractClassesFromDir,
@@ -36,9 +37,10 @@ function defaultMosaicHome(): string {
   return process.env['MOSAIC_HOME'] ?? join(homedir(), '.config', 'mosaic');
 }
 
-/** Directory holding the seeded profile yaml files. */
+/** Directory holding the seeded profile yaml files — brain home when active
+ * (user working copies, committed), else the config home seed. */
 export function defaultProfilesDir(mosaicHome = defaultMosaicHome()): string {
-  return join(mosaicHome, 'fleet', 'profiles');
+  return fleetProfilesDir(mosaicHome);
 }
 
 /** Directory holding the persona role contracts. */

packages/mosaic/src/commands/fleet-regen-command.ts

@@ -3,6 +3,7 @@ import { homedir } from 'node:os';
 import { join, relative, resolve } from 'node:path';
 import type { Command } from 'commander';
 import type { CommandRunner } from './fleet.js';
+import { fleetAgentEnvDir } from '../fleet/brain-home.js';
 import {
   applyPreparedGeneratedAgentEnvironmentProjection,
   prepareGeneratedAgentEnvironmentProjection,
@@ -153,7 +154,7 @@ export async function executeFleetRegen(
   options: FleetRegenOptions,
 ): Promise<FleetRegenResult> {
   const mosaicHome = defaultMosaicHome(deps);
-  const agentEnvDir = join(mosaicHome, 'fleet', 'agents');
+  const agentEnvDir = fleetAgentEnvDir(mosaicHome);
   const rosterPath = join(mosaicHome, 'fleet', 'roster.yaml');
   const readRoster = deps.readRoster ?? defaultReadRoster(deps, mosaicHome);
   const prepare = deps.prepareProjection ?? prepareGeneratedAgentEnvironmentProjection;

packages/mosaic/src/commands/fleet.ts

@@ -13,6 +13,7 @@ import {
 import { randomUUID } from 'node:crypto';
 import { homedir, hostname, userInfo } from 'node:os';
 import { dirname, join, resolve } from 'node:path';
+import { fleetAgentEnvDir } from '../fleet/brain-home.js';
 import { fileURLToPath } from 'node:url';
 import { spawn } from 'node:child_process';
 import * as readline from 'node:readline';
@@ -158,7 +159,7 @@ export function resolveFleetPaths(mosaicHome = defaultMosaicHome()): FleetPaths
     fleetToolsDir: join(mosaicHome, 'tools', 'fleet'),
     tmuxToolsDir: join(mosaicHome, 'tools', 'tmux'),
     systemdUserDir: join(homedir(), '.config', 'systemd', 'user'),
-    agentEnvDir: join(mosaicHome, 'fleet', 'agents'),
+    agentEnvDir: fleetAgentEnvDir(mosaicHome),
   };
 }
 

packages/mosaic/src/commands/launch.spec.ts

@@ -349,3 +349,90 @@ describe('registerRuntimeLaunchers — claudex (EXPERIMENTAL overlay)', () => {
     expect(mockExit).not.toHaveBeenCalled();
   });
 });
+
+// ─── Seat harness homes (MOSAIC-D-002, brain-home split) ────────────────────
+
+import { activeSeatDir, seatPersonaOverlay } from './launch.js';
+
+describe('activeSeatDir — per-agent harness home resolution', () => {
+  let root: string;
+  const savedAgentName = process.env['MOSAIC_AGENT_NAME'];
+  const savedBrainHome = process.env['MOSAIC_BRAIN_HOME'];
+
+  beforeEach(() => {
+    root = mkdtempSync(join(tmpdir(), 'mosaic-seat-home-'));
+    delete process.env['MOSAIC_BRAIN_HOME'];
+  });
+
+  afterEach(() => {
+    rmSync(root, { recursive: true, force: true });
+    if (savedAgentName === undefined) {
+      delete process.env['MOSAIC_AGENT_NAME'];
+    } else {
+      process.env['MOSAIC_AGENT_NAME'] = savedAgentName;
+    }
+    if (savedBrainHome !== undefined) {
+      process.env['MOSAIC_BRAIN_HOME'] = savedBrainHome;
+    } else {
+      delete process.env['MOSAIC_BRAIN_HOME'];
+    }
+  });
+
+  it('resolves the seat dir when MOSAIC_BRAIN_HOME carries the seat', () => {
+    const seat = join(root, 'brain', 'fleet', 'agents', 'coder0');
+    mkdirSync(seat, { recursive: true });
+    process.env['MOSAIC_AGENT_NAME'] = 'coder0';
+    process.env['MOSAIC_BRAIN_HOME'] = join(root, 'brain');
+
+    expect(activeSeatDir(join(root, 'config', 'mosaic'))).toBe(seat);
+  });
+
+  it('returns undefined without an agent name (bare launches stay shared)', () => {
+    delete process.env['MOSAIC_AGENT_NAME'];
+    expect(activeSeatDir(join(root, 'config', 'mosaic'))).toBeUndefined();
+  });
+
+  it('returns undefined when the seat dir does not exist in the brain', () => {
+    process.env['MOSAIC_AGENT_NAME'] = 'ghost';
+    process.env['MOSAIC_BRAIN_HOME'] = join(root, 'brain');
+    mkdirSync(join(root, 'brain', 'fleet', 'agents'), { recursive: true });
+
+    expect(activeSeatDir(join(root, 'config', 'mosaic'))).toBeUndefined();
+  });
+
+  it.each(['../escape', 'a/b', '.hidden-start', '', 'spaced name'])(
+    'rejects unsafe agent name %j (path traversal cannot leave the seat store)',
+    (name: string) => {
+      process.env['MOSAIC_AGENT_NAME'] = name;
+      process.env['MOSAIC_BRAIN_HOME'] = join(root, 'brain');
+
+      expect(activeSeatDir(join(root, 'config', 'mosaic'))).toBeUndefined();
+    },
+  );
+
+  it('seatPersonaOverlay renders the seat SOUL.md as an overlay block', () => {
+    const seat = join(root, 'brain', 'fleet', 'agents', 'coder0');
+    mkdirSync(seat, { recursive: true });
+    writeFileSync(join(seat, 'SOUL.md'), '# coder0 — code seat persona\n\nShips tested code.\n');
+    process.env['MOSAIC_AGENT_NAME'] = 'coder0';
+    process.env['MOSAIC_BRAIN_HOME'] = join(root, 'brain');
+
+    const overlay = seatPersonaOverlay(join(root, 'config', 'mosaic'));
+    expect(overlay).toContain('## Seat Persona');
+    expect(overlay).toContain('coder0 — code seat persona');
+  });
+
+  it('seatPersonaOverlay is empty when the seat carries no SOUL.md', () => {
+    const seat = join(root, 'brain', 'fleet', 'agents', 'coder0');
+    mkdirSync(seat, { recursive: true });
+    process.env['MOSAIC_AGENT_NAME'] = 'coder0';
+    process.env['MOSAIC_BRAIN_HOME'] = join(root, 'brain');
+
+    expect(seatPersonaOverlay(join(root, 'config', 'mosaic'))).toBe('');
+  });
+
+  it('seatPersonaOverlay is empty when no agent name is set', () => {
+    delete process.env['MOSAIC_AGENT_NAME'];
+    expect(seatPersonaOverlay(join(root, 'config', 'mosaic'))).toBe('');
+  });
+});

packages/mosaic/src/commands/launch.ts

@@ -19,7 +19,7 @@ import {
 import { createHash, randomBytes } from 'node:crypto';
 import { createRequire } from 'node:module';
 import { homedir, hostname } from 'node:os';
-import { join, dirname } from 'node:path';
+import { join, dirname, resolve } from 'node:path';
 import type { Command } from 'commander';
 import {
   buildResolvedFleetCommsBlock,
@@ -29,6 +29,7 @@ import {
 import { readRegularFileSecure } from '../fleet/secure-file.js';
 import { readPersonaContractBlock } from '../fleet/persona-contract.js';
 import { canonicalizeRoleClass } from './fleet-personas.js';
+import { resolveBrainHome } from '../fleet/brain-home.js';
 import { launchClaudex, type ClaudexHarnessAdapter } from './claudex.js';
 import { runLeaseEnforcementDoctorCheck } from './lease-doctor-check.js';
 
@@ -64,9 +65,46 @@ const HARNESS_HOME_ENV: Record<RuntimeName, string> = {
   opencode: 'XDG_CONFIG_HOME',
 };
 
-/** Dedicated mosaic-owned home for a runtime: ~/.config/mosaic/.<runtime> */
-function harnessHome(runtime: RuntimeName): string {
-  return join(MOSAIC_HOME, `.${runtime}`);
+/** Dedicated mosaic-owned home for a runtime: ~/.config/mosaic/.<runtime>.
+ * With an active brain seat (MOSAIC_AGENT_NAME + seat dir in the brain home)
+ * the home is per-agent instead: <brainHome>/fleet/agents/<seat>/.<runtime> —
+ * per-agent sessions, settings, and auth inside the seat dir (canon §2,
+ * MOSAIC-D-002). Seat runtime dirs are dot-named so the brain's ignore policy
+ * (per-seat .pi/.claude/.codex dirs) keeps credential material untracked. */
+const SEAT_AGENT_NAME_RE = /^[a-zA-Z0-9][a-zA-Z0-9._-]*$/;
+
+export function activeSeatDir(mosaicHome: string = MOSAIC_HOME): string | undefined {
+  const agent = process.env['MOSAIC_AGENT_NAME']?.trim();
+  if (
+    agent === undefined ||
+    agent === '' ||
+    !SEAT_AGENT_NAME_RE.test(agent) ||
+    agent.includes('..')
+  ) {
+    return undefined;
+  }
+  const brain = resolveBrainHome(mosaicHome);
+  if (resolve(brain) === resolve(mosaicHome)) return undefined; // no brain
+  const seat = join(brain, 'fleet', 'agents', agent);
+  return existsSync(seat) ? seat : undefined;
+}
+
+function harnessHome(runtime: RuntimeName, mosaicHome: string = MOSAIC_HOME): string {
+  const seat = activeSeatDir(mosaicHome);
+  if (seat !== undefined) return join(seat, `.${runtime}`);
+  return join(mosaicHome, `.${runtime}`);
+}
+
+/** Seat persona block: with an active brain seat, <seat>/SOUL.md layers
+ * persona on the root generic base (canon invariant; MOSAIC-D-002). The base
+ * SOUL stays load-on-demand — only the seat delta is injected by value.
+ * Empty string when no seat is active or the seat carries no SOUL.md. */
+export function seatPersonaOverlay(mosaicHome: string = MOSAIC_HOME): string {
+  const seatDir = activeSeatDir(mosaicHome);
+  if (seatDir === undefined) return '';
+  const seatSoul = readOptional(join(seatDir, 'SOUL.md'));
+  if (!seatSoul.trim()) return '';
+  return '## Seat Persona\n\n' + seatSoul.trim();
 }
 
 /**
@@ -182,6 +220,8 @@ function recordLaunch(runtime: RuntimeName, cliArgs: string[], yolo: boolean): v
       cli_version: CLI_VERSION,
       config_home: harnessHome(runtime),
       config_home_isolated: true,
+      config_home_kind: activeSeatDir() !== undefined ? 'seat' : 'runtime-shared',
+      agent_name: process.env['MOSAIC_AGENT_NAME']?.trim() || null,
       config_home_env: HARNESS_HOME_ENV[runtime] ?? null,
       argv: redactArgv(cliArgs),
       normative_fragments: normativeFragmentDigests(runtime),
@@ -569,6 +609,11 @@ For required push/merge/issue-close/release actions, execute without routine con
   if (soulLocal.trim()) {
     overlayBlocks.push('## Persona Overlay (SOUL.local.md)\n\n' + soulLocal.trim());
   }
+  // Seat persona (MOSAIC-D-002): per-seat SOUL.md layers on the generic base.
+  const seatPersona = seatPersonaOverlay(mosaicHome);
+  if (seatPersona !== '') {
+    overlayBlocks.push(seatPersona);
+  }
   const standardsLocal = readOptional(join(mosaicHome, 'STANDARDS.local.md'));
   if (standardsLocal.trim()) {
     overlayBlocks.push('## Standards Overlay (STANDARDS.local.md)\n\n' + standardsLocal.trim());

packages/mosaic/src/fleet/brain-home.spec.ts

@@ -0,0 +1,114 @@
+import { mkdir, mkdtemp, rm } from 'node:fs/promises';
+import { homedir, tmpdir } from 'node:os';
+import { join } from 'node:path';
+import { afterEach, beforeEach, describe, expect, it } from 'vitest';
+import {
+  brainHomeIsActive,
+  fleetAgentEnvDir,
+  fleetProfilesDir,
+  fleetRolesLocalDir,
+  fleetStateDir,
+  resolveBrainHome,
+  type BrainHomeOptions,
+} from './brain-home.js';
+
+describe('fleet brain-home resolution', (): void => {
+  let cleanup: string | undefined;
+
+  const savedBrainEnv = process.env['MOSAIC_BRAIN_HOME'];
+
+  beforeEach((): void => {
+    delete process.env['MOSAIC_BRAIN_HOME'];
+  });
+
+  afterEach(async (): Promise<void> => {
+    if (savedBrainEnv === undefined) {
+      delete process.env['MOSAIC_BRAIN_HOME'];
+    } else {
+      process.env['MOSAIC_BRAIN_HOME'] = savedBrainEnv;
+    }
+    if (cleanup !== undefined) {
+      await rm(cleanup, { recursive: true, force: true });
+      cleanup = undefined;
+    }
+  });
+
+  async function makeTmp(): Promise<string> {
+    const root = await mkdtemp(join(tmpdir(), 'mosaic-brain-home-'));
+    cleanup = root;
+    return root;
+  }
+
+  it('MOSAIC_BRAIN_HOME env wins over every other signal', (): void => {
+    process.env['MOSAIC_BRAIN_HOME'] = '/explicit/brain';
+    expect(resolveBrainHome('/any/mosaic-home')).toBe('/explicit/brain');
+    expect(fleetAgentEnvDir('/any/mosaic-home')).toBe('/explicit/brain/fleet/agents');
+    expect(brainHomeIsActive('/any/mosaic-home')).toBe(true);
+  });
+
+  it('injected envBrainHome wins identically (test seam)', (): void => {
+    const opts: BrainHomeOptions = { envBrainHome: '/injected/brain' };
+    expect(resolveBrainHome('/any/mosaic-home', opts)).toBe('/injected/brain');
+    expect(fleetAgentEnvDir('/any/mosaic-home', opts)).toBe('/injected/brain/fleet/agents');
+  });
+
+  it('a non-default mosaicHome never adopts the canonical brain (hermetic legacy)', (): void => {
+    const mosaicHome = '/tmp/not-the-default-config-home';
+    expect(resolveBrainHome(mosaicHome)).toBe(mosaicHome);
+    expect(brainHomeIsActive(mosaicHome)).toBe(false);
+    expect(fleetAgentEnvDir(mosaicHome)).toBe(join(mosaicHome, 'fleet', 'agents'));
+  });
+
+  it('the default config home adopts the brain when it carries fleet/agents', async (): Promise<void> => {
+    const root = await makeTmp();
+    const brain = join(root, 'brain');
+    await mkdir(join(brain, 'fleet', 'agents'), { recursive: true });
+    const configHome = join(root, 'config', 'mosaic');
+    const opts: BrainHomeOptions = { homes: { brain, configDefault: configHome } };
+
+    expect(resolveBrainHome(configHome, opts)).toBe(brain);
+    expect(fleetAgentEnvDir(configHome, opts)).toBe(join(brain, 'fleet', 'agents'));
+    expect(fleetRolesLocalDir(configHome, opts)).toBe(join(brain, 'fleet', 'roles.local'));
+    expect(fleetProfilesDir(configHome, opts)).toBe(join(brain, 'fleet', 'profiles'));
+    expect(fleetStateDir(configHome, opts)).toBe(join(brain, 'fleet'));
+    expect(brainHomeIsActive(configHome, opts)).toBe(true);
+  });
+
+  it('the default config home stays legacy when no brain exists', async (): Promise<void> => {
+    const root = await makeTmp();
+    const configHome = join(root, 'config', 'mosaic');
+    const opts: BrainHomeOptions = {
+      homes: { brain: join(root, 'brain'), configDefault: configHome },
+    };
+
+    expect(resolveBrainHome(configHome, opts)).toBe(configHome);
+    expect(brainHomeIsActive(configHome, opts)).toBe(false);
+  });
+
+  it('an empty MOSAIC_BRAIN_HOME is ignored, not treated as set', (): void => {
+    process.env['MOSAIC_BRAIN_HOME'] = '   ';
+    expect(resolveBrainHome('/tmp/legacy-home')).toBe('/tmp/legacy-home');
+  });
+
+  it('adoption requires fleet/agents specifically, not any brain content', async (): Promise<void> => {
+    const root = await makeTmp();
+    const brain = join(root, 'brain');
+    await mkdir(join(brain, 'fleet'), { recursive: true }); // fleet without agents
+    const configHome = join(root, 'config', 'mosaic');
+    const opts: BrainHomeOptions = { homes: { brain, configDefault: configHome } };
+
+    expect(resolveBrainHome(configHome, opts)).toBe(configHome);
+  });
+
+  it('real-home control: a host brain is adopted only through the default home', (): void => {
+    // Control on the un-injected path: this host carries ~/.mosaic/fleet/agents,
+    // so the default config home resolves to the brain or legacy — both valid
+    // canonical endpoints — while a non-default home never adopts.
+    const defaultHome = join(homedir(), '.config', 'mosaic');
+    const resolved = resolveBrainHome(defaultHome);
+    expect([defaultHome, join(homedir(), '.mosaic')]).toContain(resolved);
+    expect(resolveBrainHome(join(homedir(), 'elsewhere', 'mosaic'))).toBe(
+      join(homedir(), 'elsewhere', 'mosaic'),
+    );
+  });
+});

packages/mosaic/src/fleet/brain-home.ts

@@ -0,0 +1,76 @@
+import { existsSync } from 'node:fs';
+import { homedir } from 'node:os';
+import { join, resolve } from 'node:path';
+
+/**
+ * Overridable resolution inputs (tests inject tmp homes; production reads
+ * the environment and the real home directory).
+ */
+export interface BrainHomeOptions {
+  /** Explicit brain home; defaults to `MOSAIC_BRAIN_HOME`. */
+  readonly envBrainHome?: string;
+  /**
+   * Canonical homes used for adoption. Defaults derive from the real
+   * `homedir()`: `{ brain: ~/.mosaic, configDefault: ~/.config/mosaic }`.
+   */
+  readonly homes?: { readonly brain: string; readonly configDefault: string };
+}
+
+/**
+ * Brain-home resolution — the three-tree fleet split (stack canon
+ * `docs/STRUCTURE-CANON.md` §2, first carried by the USC estate brain):
+ *
+ *   config home (~/.config/mosaic)  framework templates + dispatch state:
+ *                                   fleet/roles (baseline), fleet/roster.yaml,
+ *                                   fleet/run (heartbeats), fleet/services
+ *   brain home  (~/.mosaic)         user-owned fleet state, committed:
+ *                                   fleet/agents/<seat>.env.*, fleet/roles.local,
+ *                                   fleet/profiles working copies
+ *
+ * Resolution order:
+ *   1. `MOSAIC_BRAIN_HOME` env (explicit, always wins)
+ *   2. canonical `~/.mosaic` — adopted ONLY when mosaicHome is the real
+ *      default config home AND `~/.mosaic/fleet/agents` exists. Custom
+ *      `--mosaic-home` values (tests, sandboxes, canaries) never trigger
+ *      adoption, keeping them hermetic and deterministic.
+ *   3. mosaicHome itself (legacy single-tree behavior).
+ */
+export function resolveBrainHome(mosaicHome: string, options: BrainHomeOptions = {}): string {
+  const explicit = options.envBrainHome ?? process.env['MOSAIC_BRAIN_HOME'];
+  if (explicit !== undefined && explicit.trim() !== '') {
+    return explicit;
+  }
+  const homes = options.homes ?? {
+    brain: join(homedir(), '.mosaic'),
+    configDefault: join(homedir(), '.config', 'mosaic'),
+  };
+  if (resolve(mosaicHome) !== resolve(homes.configDefault)) {
+    return mosaicHome;
+  }
+  return existsSync(join(homes.brain, 'fleet', 'agents')) ? homes.brain : mosaicHome;
+}
+
+/** True when fleet state resolves somewhere other than the config home. */
+export function brainHomeIsActive(mosaicHome: string, options: BrainHomeOptions = {}): boolean {
+  return resolve(resolveBrainHome(mosaicHome, options)) !== resolve(mosaicHome);
+}
+
+/** Fleet state root (brain home when active, else the config home). */
+export function fleetStateDir(mosaicHome: string, options: BrainHomeOptions = {}): string {
+  return join(resolveBrainHome(mosaicHome, options), 'fleet');
+}
+
+/** Seat launch envs — `<brainHome>/fleet/agents` when a brain is active. */
+export function fleetAgentEnvDir(mosaicHome: string, options: BrainHomeOptions = {}): string {
+  return join(fleetStateDir(mosaicHome, options), 'agents');
+}
+
+/** PRESERVE-protected persona override layer — `<brainHome>/fleet/roles.local`. */
+export function fleetRolesLocalDir(mosaicHome: string, options: BrainHomeOptions = {}): string {
+  return join(fleetStateDir(mosaicHome, options), 'roles.local');
+}
+
+/** System-type profiles (user working copies) — `<brainHome>/fleet/profiles`. */
+export function fleetProfilesDir(mosaicHome: string, options: BrainHomeOptions = {}): string {
+  return join(fleetStateDir(mosaicHome, options), 'profiles');
+}

packages/mosaic/src/fleet/fleet-reconciler.ts

@@ -3,6 +3,7 @@ import { lstat, open, readFile, unlink, type FileHandle } from 'node:fs/promises
 import { randomUUID } from 'node:crypto';
 import { homedir } from 'node:os';
 import { join } from 'node:path';
+import { fleetAgentEnvDir } from './brain-home.js';
 import {
   applyPreparedAgentEnvironmentProjection,
   prepareAgentEnvironmentProjection,
@@ -617,7 +618,7 @@ function defaultPrepareProjections(
         (agent: FleetRosterV2Agent): Promise<PreparedAgentEnvironmentProjection> =>
           prepareAgentEnvironmentProjection({
             mosaicHome,
-            agentEnvDir: join(mosaicHome, 'fleet', 'agents'),
+            agentEnvDir: fleetAgentEnvDir(mosaicHome),
             agentName: agent.name,
             generated: projectRosterV2AgentGeneratedEnv(roster, agent),
           }),

packages/mosaic/src/fleet/generated-env-boundary.spec.ts

@@ -176,6 +176,52 @@ describe('generated fleet agent environment boundary', (): void => {
     expect((await stat(result.generatedPath)).mode & 0o777).toBe(0o600);
   });
 
+  it('brain home: accepts and writes projections under MOSAIC_BRAIN_HOME/fleet/agents', async (): Promise<void> => {
+    const savedBrainHome = process.env['MOSAIC_BRAIN_HOME'];
+    try {
+      cleanup = await mkdtemp(join(tmpdir(), 'mosaic-generated-env-'));
+      const mosaicHome = join(cleanup, 'config-home');
+      const brainHome = join(cleanup, 'brain');
+      const agentEnvDir = join(brainHome, 'fleet', 'agents');
+      process.env['MOSAIC_BRAIN_HOME'] = brainHome;
+
+      const result = await writeAgentEnvironmentProjection({
+        mosaicHome,
+        agentEnvDir,
+        agentName: 'coder0',
+        generated: generatedValues,
+      });
+
+      // Projection landed in the brain tree, not under the config home.
+      expect(result.generatedPath).toBe(join(agentEnvDir, 'coder0.env.generated'));
+      expect((await stat(join(brainHome, 'fleet'))).mode & 0o777).toBe(0o700);
+      expect((await stat(agentEnvDir)).mode & 0o777).toBe(0o700);
+      expect((await stat(result.generatedPath)).mode & 0o777).toBe(0o600);
+      await expect(stat(join(mosaicHome, 'fleet'))).rejects.toThrow();
+
+      // A config-home agentEnvDir is now REJECTED while the brain is active —
+      // the boundary must not silently split state across two trees.
+      let rejected: unknown;
+      try {
+        await writeAgentEnvironmentProjection({
+          mosaicHome,
+          agentEnvDir: join(mosaicHome, 'fleet', 'agents'),
+          agentName: 'coder1',
+          generated: { ...generatedValues, MOSAIC_AGENT_NAME: 'coder1' },
+        });
+      } catch (caught: unknown) {
+        rejected = caught;
+      }
+      expect(rejected).toBeInstanceOf(AgentEnvBoundaryError);
+    } finally {
+      if (savedBrainHome === undefined) {
+        delete process.env['MOSAIC_BRAIN_HOME'];
+      } else {
+        process.env['MOSAIC_BRAIN_HOME'] = savedBrainHome;
+      }
+    }
+  });
+
   it('regenerates desired keys, relocates safe legacy local data, and quarantines forbidden legacy input', async (): Promise<void> => {
     cleanup = await mkdtemp(join(tmpdir(), 'mosaic-generated-env-'));
     const mosaicHome = join(cleanup, 'mosaic');

packages/mosaic/src/fleet/generated-env-boundary.ts

@@ -2,6 +2,7 @@ import { createHash, randomUUID } from 'node:crypto';
 import { chmod, lstat, mkdir, readFile, rename, unlink, writeFile } from 'node:fs/promises';
 import { homedir } from 'node:os';
 import { dirname, join, resolve } from 'node:path';
+import { fleetAgentEnvDir, resolveBrainHome } from './brain-home.js';
 import { compareCodePoints } from './deterministic-order.js';
 
 export type AgentEnvironmentKind = 'generated' | 'local';
@@ -528,12 +529,15 @@ async function validatePrivateProjectionDirectory(
   mosaicHome: string,
   agentEnvDir: string,
 ): Promise<void> {
-  const fleetDir = join(mosaicHome, 'fleet');
-  const expectedAgentEnvDir = join(fleetDir, 'agents');
+  // Brain-home split (canon §2): seat envs live under the brain home's
+  // fleet/agents when a brain is active; roster + templates stay config-home.
+  const expectedAgentEnvDir = fleetAgentEnvDir(mosaicHome);
   if (resolve(agentEnvDir) !== resolve(expectedAgentEnvDir)) {
     throw new AgentEnvBoundaryError('unsafe-directory', '(directory)', agentEnvDir);
   }
-  await assertManagedDirectoryIfPresent(mosaicHome, false);
+  const stateHome = resolveBrainHome(mosaicHome);
+  const fleetDir = join(stateHome, 'fleet');
+  await assertManagedDirectoryIfPresent(stateHome, false);
   await assertManagedDirectoryIfPresent(fleetDir, false);
   await assertManagedDirectoryIfPresent(agentEnvDir, true);
 }
@@ -543,8 +547,9 @@ async function ensurePrivateProjectionDirectory(
   agentEnvDir: string,
 ): Promise<void> {
   await validatePrivateProjectionDirectory(mosaicHome, agentEnvDir);
-  const fleetDir = join(mosaicHome, 'fleet');
-  await ensureManagedDirectory(mosaicHome, false);
+  const stateHome = resolveBrainHome(mosaicHome);
+  const fleetDir = join(stateHome, 'fleet');
+  await ensureManagedDirectory(stateHome, false);
   await ensureManagedDirectory(fleetDir, false);
   await ensureManagedDirectory(agentEnvDir, true);
 }

scripts/verify-release.mjs

@@ -0,0 +1,166 @@
+#!/usr/bin/env node
+// verify-release.mjs — the ONE canonical terminal verification command
+// (SDLC-D-034, `pnpm verify:release`).
+//
+// Publication (.woodpecker/publish.yml `verify` step) is bound to terminal
+// verification of the exact commit through this command, which is composed
+// from the SAME commands the PR CI pipeline (.woodpecker/ci.yml) runs — CI and
+// publish share one semantic checklist:
+//
+//   stage         | mirrors ci.yml step | commands
+//   --------------|---------------------|------------------------------------------
+//   sanitization  | sanitization        | verify-sanitized.sh, check-resident-
+//                |                     | budget.sh (--self-test + run),
+//                |                     | check-test-enumeration.sh
+//   upgrade-guard | upgrade-guard       | test-upgrade-manifest-guard.sh,
+//                |                     | test-upgrade-rollback.sh,
+//                |                     | test-upgrade-durable-snapshot.sh,
+//                |                     | test-install-migration.sh
+//   typecheck     | typecheck           | pnpm typecheck (runs the checkout
+//                |                     | preflight, then turbo typecheck)
+//   lint          | lint                | pnpm lint
+//   format        | format              | pnpm format:check
+//   test          | test                | pnpm test
+//   build         | publish.yml build   | pnpm build
+//
+// Caller-provided prerequisites (kept at the pipeline level — see the comments
+// in .woodpecker/ci.yml): `bash` + `rsync` for the guard stages, `openssl` and
+// the pinned @earendil-works/pi-coding-agent for the test stage, and — on the
+// postgres path only — the ci-postgres service plus
+// `pnpm --filter @mosaicstack/db run db:migrate` before the test stage.
+//
+// This command works with DATABASE_URL set (CI postgres path) or unset (local
+// PGlite path); it never sets, exports, or requires a database itself.
+//
+// scripts/verify-release.test.mjs enforces that this stage table keeps
+// matching .woodpecker/ci.yml step-for-step, so the two surfaces cannot drift
+// apart silently.
+
+import { spawnSync } from 'node:child_process';
+import path from 'node:path';
+import { fileURLToPath } from 'node:url';
+
+export const STAGES = [
+  {
+    name: 'sanitization',
+    // Mirror of the .woodpecker/ci.yml `sanitization` step (minus its
+    // `apk add` environment prep). Kept as direct command strings here: the
+    // #1017 test-enumeration guard audits these paths through the ci.yml
+    // surface, so indirection from ci.yml into this file is not possible.
+    commands: [
+      'bash packages/mosaic/framework/tools/quality/scripts/verify-sanitized.sh',
+      'bash packages/mosaic/framework/tools/quality/scripts/check-resident-budget.sh --self-test',
+      'bash packages/mosaic/framework/tools/quality/scripts/check-resident-budget.sh',
+      'bash packages/mosaic/framework/tools/quality/scripts/check-test-enumeration.sh',
+    ],
+  },
+  {
+    name: 'upgrade-guard',
+    // Mirror of the .woodpecker/ci.yml `upgrade-guard` step (minus its
+    // `apk add` environment prep).
+    commands: [
+      'bash packages/mosaic/framework/tools/quality/scripts/test-upgrade-manifest-guard.sh',
+      'bash packages/mosaic/framework/tools/quality/scripts/test-upgrade-rollback.sh',
+      'bash packages/mosaic/framework/tools/quality/scripts/test-upgrade-durable-snapshot.sh',
+      'bash packages/mosaic/framework/tools/quality/scripts/test-install-migration.sh',
+    ],
+  },
+  {
+    // `pnpm typecheck` is `pnpm preflight && turbo run typecheck`, so the
+    // checkout preflight (scripts/preflight.mjs) is part of this stage exactly
+    // as it is part of the ci.yml `typecheck` step.
+    name: 'typecheck',
+    commands: ['pnpm typecheck'],
+  },
+  {
+    name: 'lint',
+    commands: ['pnpm lint'],
+  },
+  {
+    name: 'format',
+    commands: ['pnpm format:check'],
+  },
+  {
+    // Requires `openssl` and the pinned `pi` binary on the pipeline path; see
+    // the caller-provided prerequisites above.
+    name: 'test',
+    commands: ['pnpm test'],
+  },
+  {
+    name: 'build',
+    commands: ['pnpm build'],
+  },
+];
+
+export function stageByName(name) {
+  return STAGES.find((stage) => stage.name === name);
+}
+
+function missingBinaries(bins) {
+  return bins.filter(
+    (bin) => spawnSync('sh', ['-c', `command -v ${bin} >/dev/null 2>&1`]).status !== 0,
+  );
+}
+
+function runCommand(command) {
+  const result = spawnSync(command, { shell: true, stdio: 'inherit' });
+  if (result.error) {
+    console.error(`[verify:release] failed to launch '${command}': ${result.error.message}`);
+    return false;
+  }
+  if (result.status !== 0) {
+    const reason = result.signal ? `terminated by ${result.signal}` : `exited ${result.status}`;
+    console.error(`[verify:release] command '${command}' ${reason}`);
+    return false;
+  }
+  return true;
+}
+
+// Runs the complete mandatory verification set (or, with --stage <name>, the
+// single named stage — used for wiring/smoke-testing, not for gating: only a
+// run of every stage is a terminal verification). Fails fast: the first
+// failing command aborts with a non-zero exit code. Returns the exit code.
+export function verifyRelease({ stages = STAGES } = {}) {
+  const missing = missingBinaries(['bash', 'rsync']);
+  if (missing.length > 0) {
+    console.error(
+      `[verify:release] FATAL: required binaries missing from PATH: ${missing.join(', ')}. ` +
+        'The caller provides them (ci-base bakes bash; pipelines apk add rsync).',
+    );
+    return 1;
+  }
+  for (const stage of stages) {
+    console.log(`\n[verify:release] === stage: ${stage.name} ===`);
+    for (const command of stage.commands) {
+      console.log(`[verify:release] $ ${command}`);
+      if (!runCommand(command)) {
+        console.error(
+          `[verify:release] FATAL: stage '${stage.name}' failed — verification inconclusive`,
+        );
+        return 1;
+      }
+    }
+  }
+  console.log(`\n[verify:release] all ${stages.length} stage(s) passed`);
+  return 0;
+}
+
+function main(argv) {
+  const stageFlagIndex = argv.indexOf('--stage');
+  if (stageFlagIndex !== -1) {
+    const name = argv[stageFlagIndex + 1];
+    const stage = stageByName(name);
+    if (!stage) {
+      console.error(
+        `[verify:release] unknown stage '${name ?? ''}' — expected one of: ${STAGES.map((entry) => entry.name).join(', ')}`,
+      );
+      process.exit(2);
+    }
+    process.exit(verifyRelease({ stages: [stage] }));
+  }
+  process.exit(verifyRelease());
+}
+
+if (process.argv[1] && path.resolve(process.argv[1]) === fileURLToPath(import.meta.url)) {
+  main(process.argv.slice(2));
+}

scripts/verify-release.test.mjs

@@ -0,0 +1,277 @@
+import assert from 'node:assert/strict';
+import { readFile } from 'node:fs/promises';
+import { createRequire } from 'node:module';
+import path from 'node:path';
+import test from 'node:test';
+
+import { STAGES } from './verify-release.mjs';
+
+// SDLC-D-034 checkout invariant: publication in .woodpecker/publish.yml is
+// bound to exact-commit terminal verification. This suite parses the real
+// pipeline files and fails red when the gate is bypassed, weakened, or drifts
+// out of sync with the canonical `pnpm verify:release` command.
+
+// Reuse the monorepo's existing YAML parser (@mosaicstack/mosaic's direct
+// dependency) instead of adding a root dependency or vendoring a parser.
+const mosaicRequire = createRequire(
+  path.resolve(process.cwd(), 'packages', 'mosaic', 'package.json'),
+);
+const { parse: parseYaml } = mosaicRequire('yaml');
+
+const publishYmlPath = path.join(process.cwd(), '.woodpecker', 'publish.yml');
+const ciYmlPath = path.join(process.cwd(), '.woodpecker', 'ci.yml');
+
+async function readPublishPipeline() {
+  return parseYaml(await readFile(publishYmlPath, 'utf8'));
+}
+
+// A step has an external publication effect when its name starts with
+// `publish` or when any command pushes an image to a registry.
+function pushesImage(step) {
+  return (step.commands ?? []).some((command) =>
+    /(^|\s)(\/kaniko\/executor|docker push)\b|--destination/.test(command),
+  );
+}
+
+function publishEffectSteps(pipeline) {
+  return Object.entries(pipeline.steps ?? {})
+    .filter(([name, step]) => name.startsWith('publish') || pushesImage(step))
+    .map(([name]) => name);
+}
+
+// Transitive closure of a step's depends_on graph.
+function dependencyClosure(pipeline, stepName, seen = new Set()) {
+  const dependencies = pipeline.steps?.[stepName]?.depends_on ?? [];
+  for (const dependency of dependencies) {
+    if (seen.has(dependency)) continue;
+    seen.add(dependency);
+    dependencyClosure(pipeline, dependency, seen);
+  }
+  return seen;
+}
+
+function verifyCommands(pipeline) {
+  const verify = pipeline.steps?.verify;
+  assert.ok(verify, 'publish pipeline must define a `verify` step');
+  assert.ok(Array.isArray(verify.commands), '`verify` step must have commands');
+  return verify.commands;
+}
+
+function assertCommitIdentityAssertion(commands) {
+  const text = commands.join('\n');
+  assert.match(
+    text,
+    /CI_COMMIT_SHA/,
+    '`verify` must compare the provider commit identity (CI_COMMIT_SHA)',
+  );
+  assert.match(text, /git rev-parse HEAD/, '`verify` must compare against git rev-parse HEAD');
+  assert.match(
+    text,
+    /exit 1/,
+    '`verify` must fail closed (exit 1) on identity mismatch or emptiness',
+  );
+}
+
+function assertCanonicalCommand(commands) {
+  assert.ok(
+    commands.some((command) => /^pnpm verify:release\b/.test(command.trim())),
+    '`verify` must run the canonical terminal verification command `pnpm verify:release`',
+  );
+}
+
+function assertPublishGate(pipeline) {
+  assert.ok(pipeline.steps, 'publish pipeline must define steps');
+
+  const commands = verifyCommands(pipeline);
+  assertCommitIdentityAssertion(commands);
+  assertCanonicalCommand(commands);
+
+  const effects = publishEffectSteps(pipeline);
+  assert.ok(effects.length > 0, 'publish pipeline must contain publish effect steps to guard');
+
+  for (const stepName of effects) {
+    const step = pipeline.steps[stepName];
+    assert.ok(
+      Array.isArray(step.depends_on) && step.depends_on.includes('verify'),
+      `publish effect '${stepName}' must depend DIRECTLY on the verify step (SDLC-D-034: transitively through build is not enough)`,
+    );
+    assert.ok(
+      dependencyClosure(pipeline, stepName).has('verify'),
+      `publish effect '${stepName}' must depend on a chain that includes verify`,
+    );
+  }
+  return effects;
+}
+
+test('the publish pipeline gates every publish effect behind exact-commit verification', async () => {
+  const pipeline = await readPublishPipeline();
+  const effects = assertPublishGate(pipeline);
+  assert.deepEqual(effects.sort(), [
+    'build-appservice',
+    'build-gateway',
+    'build-web',
+    'publish-next-npm',
+    'publish-npm',
+  ]);
+});
+
+test('the verify step carries no path/event short-circuit of its own', async () => {
+  const pipeline = await readPublishPipeline();
+  // A `when` filter on `verify` would let a publish effect fire on an event
+  // class that skipped verification — the gate must be unconditional.
+  assert.equal(pipeline.steps.verify.when, undefined);
+});
+
+test('a publish step that bypasses verify fails the gate checker', () => {
+  // Negative fixture: a plausible publish pipeline where `publish-npm` hangs
+  // off `build` only and `build` never chains to `verify` — the exact bypass
+  // class SDLC-D-034 closes. The checker must go red on it.
+  const bypassingPipeline = `
+steps:
+  install:
+    image: node:24-alpine
+    commands:
+      - pnpm install --frozen-lockfile
+  verify:
+    image: node:24-alpine
+    commands:
+      - |
+        if [ -z "$CI_COMMIT_SHA" ] || [ "$CI_COMMIT_SHA" != "$(git rev-parse HEAD)" ]; then
+          echo "identity mismatch" >&2
+          exit 1
+        fi
+      - pnpm verify:release
+    depends_on:
+      - install
+  build:
+    image: node:24-alpine
+    commands:
+      - pnpm build
+    depends_on:
+      - install
+  publish-npm:
+    image: node:24-alpine
+    commands:
+      - pnpm publish
+    depends_on:
+      - build
+`;
+  assert.throws(
+    () => assertPublishGate(parseYaml(bypassingPipeline)),
+    /publish-npm.*DIRECTLY.*verify/s,
+  );
+});
+
+test('a publish step chained to verify only transitively fails the gate checker', () => {
+  // Negative fixture: `build` depends on verify but `publish-npm` does not
+  // carry the direct edge — weaker than SDLC-D-034 requires of the real DAG.
+  const transitiveOnlyPipeline = `
+steps:
+  install:
+    image: node:24-alpine
+    commands:
+      - pnpm install --frozen-lockfile
+  verify:
+    image: node:24-alpine
+    commands:
+      - |
+        if [ -z "$CI_COMMIT_SHA" ] || [ "$CI_COMMIT_SHA" != "$(git rev-parse HEAD)" ]; then
+          echo "identity mismatch" >&2
+          exit 1
+        fi
+      - pnpm verify:release
+    depends_on:
+      - install
+  build:
+    image: node:24-alpine
+    commands:
+      - pnpm build
+    depends_on:
+      - install
+      - verify
+  publish-npm:
+    image: node:24-alpine
+    commands:
+      - pnpm publish
+    depends_on:
+      - build
+`;
+  assert.throws(
+    () => assertPublishGate(parseYaml(transitiveOnlyPipeline)),
+    /publish-npm.*DIRECTLY.*verify/s,
+  );
+});
+
+test('a verify step without the commit-identity assertion fails the gate checker', () => {
+  const noIdentityPipeline = `
+steps:
+  verify:
+    image: node:24-alpine
+    commands:
+      - pnpm verify:release
+  publish-npm:
+    image: node:24-alpine
+    commands:
+      - pnpm publish
+    depends_on:
+      - verify
+`;
+  assert.throws(() => assertPublishGate(parseYaml(noIdentityPipeline)), /CI_COMMIT_SHA/);
+});
+
+test('the canonical verify:release stages mirror the PR CI pipeline one-for-one', async () => {
+  const ci = parseYaml(await readFile(ciYmlPath, 'utf8'));
+  const canonical = Object.fromEntries(STAGES.map((stage) => [stage.name, stage.commands]));
+
+  // The complete mandatory set, in gate order.
+  assert.deepEqual(
+    STAGES.map((stage) => stage.name),
+    ['sanitization', 'upgrade-guard', 'typecheck', 'lint', 'format', 'test', 'build'],
+  );
+
+  // Guard stages: ci.yml commands minus its `apk add` environment prep must be
+  // exactly the canonical stage commands (order included).
+  for (const stageName of ['sanitization', 'upgrade-guard']) {
+    assert.deepEqual(
+      ci.steps[stageName].commands.filter((command) => !command.startsWith('apk add')),
+      canonical[stageName],
+      `canonical '${stageName}' stage must match the ci.yml step`,
+    );
+  }
+
+  // pnpm stages: ci.yml commands minus `corepack enable` must be exactly the
+  // canonical stage commands.
+  for (const stepName of ['typecheck', 'lint', 'format']) {
+    assert.deepEqual(
+      ci.steps[stepName].commands.filter((command) => command !== 'corepack enable'),
+      canonical[stepName],
+      `canonical '${stepName}' stage must match the ci.yml step`,
+    );
+  }
+
+  // The test stage is shared, but ci.yml wraps it in pipeline-level
+  // prerequisites the canonical command expects its caller to provide
+  // (SDLC-D-034): the postgres service + readiness wait + db:migrate, openssl,
+  // and the pinned pi runtime. None of those may be dropped silently.
+  for (const command of canonical.test) {
+    assert.ok(
+      ci.steps.test.commands.includes(command),
+      `ci.yml test step must run the canonical test stage command '${command}'`,
+    );
+  }
+  for (const fragment of [
+    'pg_isready -h ci-postgres',
+    'pnpm --filter @mosaicstack/db run db:migrate',
+    'npm install -g @earendil-works/[email protected]',
+  ]) {
+    assert.ok(
+      ci.steps.test.commands.some((command) => command.includes(fragment)),
+      `ci.yml test step must keep its pipeline-level prerequisite '${fragment}'`,
+    );
+  }
+});
+
+test('the root package.json exposes verify:release as the canonical command', async () => {
+  const packageJson = JSON.parse(await readFile(path.join(process.cwd(), 'package.json'), 'utf8'));
+  assert.match(packageJson.scripts['verify:release'], /scripts\/verify-release\.mjs/);
+});

skills/mosaic-forge/SKILL.md

@@ -133,10 +133,13 @@ When the full `@mosaicstack/forge` package is available, Forge uses MACP task ex
 
 ```bash
 # Run from CLI
-mosaic forge run path/to/brief.md
+# Fails closed with a typed FORGE_NO_EXECUTOR capability error when no real
+# executor is wired — pass --simulate to opt into explicit typed simulation
+# (every result carries status `simulated`, which satisfies nothing).
+mosaic forge run path/to/brief.md [--simulate]
 
-# Resume interrupted run
-mosaic forge resume .forge/runs/20260401-143022/
+# Resume interrupted run (same fail-closed rule as forge run)
+mosaic forge resume .forge/runs/20260401-143022/ [--simulate]
 
 # Check status
 mosaic forge status .forge/runs/20260401-143022/