@event4u/agent-config 9.32.0 → 9.34.0

package/.claude-plugin/marketplace.json

@@ -6,7 +6,7 @@
   },
   "metadata": {
     "description": "Bootstrap shim for the event4u agent-config suite. This plugin ships ONLY the dispatcher hooks plus one install-pointer skill — ALL content (skills, rules, commands) is installed via `npx -y @event4u/agent-config init`, which writes the file projection AND registers the same hooks in ~/.claude/settings.json. DEPRECATED as a content channel; kept so the marketplace listing stays discoverable and existing installs keep working hooks.",
-    "version": "9.32.0",
+    "version": "9.34.0",
     "keywords": [
       "agent-config",
       "skills",

package/CHANGELOG.md

@@ -314,107 +314,120 @@ Entry-shape contract: [`docs/contracts/CHANGELOG-conventions.md`](docs/contracts
 > keeps the active era under the 250-line drift cap enforced by
 > `tests/test_changelog_eras.py`.
 
-# Era: 9.31.x — current
+# Era: pre-9.34.0 — archived
 
-> Started at `9.31.0`. Full entries live inline below.
+> All entries before `9.34.0` live in
+> [`docs/archive/CHANGELOG-pre-9.34.0.md`](docs/archive/CHANGELOG-pre-9.34.0.md).
+> The archive is read-only; git tags remain the canonical
+> source for what shipped. Splitting them out of the main file
+> keeps the active era under the 250-line drift cap enforced by
+> `tests/test_changelog_eras.py`.
+
+# Era: 9.34.x — current
+
+> Started at `9.34.0`. Full entries live inline below.
 > The drift test caps this era at 250 lines of entry body; growth past
-> that forces a new era split (`# Era: 9.32.x`, etc.) — see
+> that forces a new era split (`# Era: 9.35.x`, etc.) — see
 > [`docs/contracts/CHANGELOG-conventions.md § Era splits`](docs/contracts/CHANGELOG-conventions.md).
 
-## [9.32.0](https://github.com/event4u-app/agent-config/compare/9.31.0...9.32.0) (2026-08-10)
+## [9.34.0](https://github.com/event4u-app/agent-config/compare/9.33.0...9.34.0) (2026-08-11)
 
 ### Release highlights
 
 <!-- Curated head: fill before merge, keep it under 10 lines, and leave `_none_` where it is genuinely the answer. The generated log below is unchanged. -->
-- **Behaviour changes:** _auto-derived, rewrite before merge:_ rule/schema diffs, breaking commits or removed public surface in 1f01490, e05de77.
-- **Default changes + migration:** _none_
-- **Security and correctness:** _none_
+- **Behaviour changes:** _none_
+- **Default changes + migration:** _auto-derived, rewrite before merge:_ commits naming a default, migration or migrate in 4420340.
+- **Security and correctness:** _auto-derived, rewrite before merge:_ security-scoped commits in 4420340, b2fc429.
 - **Honest nulls:** _none_
 - **Known limitations:** _none_
 
 ### Features
 
-* **work-engine:** route both red-check lanes through the self-fix loop ([10c8f7e](https://github.com/event4u-app/agent-config/commit/10c8f7eafbf21fc03a5f301163c941a20df4a885))
-* **work-engine:** bounded self-fix loop primitive + executable dod schema ([1f01490](https://github.com/event4u-app/agent-config/commit/1f0149035c1a27729b729dab1dedd861ccdf9812))
-* **envelope:** drift anchor, scripted grounding, verbatim error strings ([6c9187e](https://github.com/event4u-app/agent-config/commit/6c9187e30f41c87553764759096a3f7f4d19a126))
-* **envelope:** successor tailoring, mandatory failed_approaches, data-never-instruction boundary ([b03beac](https://github.com/event4u-app/agent-config/commit/b03beac1ad1325253c89a63eac9abc9f051269d3))
-* **orchestration:** report per-field provenance for host capabilities ([f4958de](https://github.com/event4u-app/agent-config/commit/f4958decf4a70b445a38747bff768ab449877d52))
+* **handoff-envelope:** a checkable off-limits path list, and a validated reversibility tag ([33fd14c](https://github.com/event4u-app/agent-config/commit/33fd14ca8c8aa3ade0081e5f5bf33b6baef6f3f7))
+* **r2-dispatch:** tell the three stale-artefact states apart, and route each to its contract path ([ca417a1](https://github.com/event4u-app/agent-config/commit/ca417a13ebd7890bbefedec89399add96a3f8eca))
+* **lint-handoffs:** make a blank Open-questions section a finding, not a pass ([a22534b](https://github.com/event4u-app/agent-config/commit/a22534bbf8e80eb5e32d7e547651b8efa55a47dc))
+* **self-repair:** bound record creation per source, and count what the cap refuses ([76aa7b5](https://github.com/event4u-app/agent-config/commit/76aa7b56dd53212f0b4f5d556cae27054bb2a358))
+* **discovery:** add a dormancy report that names an unavailable signal ([99cfa2f](https://github.com/event4u-app/agent-config/commit/99cfa2f632320b7e1300e0d539292e1c038437a6))
+* **discovery-graph:** add a zero-inbound query, per-pass stats and error containment ([ed3d532](https://github.com/event4u-app/agent-config/commit/ed3d5325285476e86fae62c679b6d20cbbff923f))
+* **lint-hedge-words:** diff-scoped advisory hedge lint with a declared stage ([364ef70](https://github.com/event4u-app/agent-config/commit/364ef707dce9e29917d4bc9f5d2b2aeedc4572da))
+* **claims:** bind six external research cites, gate the successor pointer ([37cc8fe](https://github.com/event4u-app/agent-config/commit/37cc8fee84343b8cb92edf7778b97ce235086dc8))
+* **skill-linter:** gate Security constraints on script-bearing skills ([b2fc429](https://github.com/event4u-app/agent-config/commit/b2fc4292540fe140ca9ead90b32bebaac133d2d0))
+* **council:** check a synthesis verdict against its own stance tally ([37103c3](https://github.com/event4u-app/agent-config/commit/37103c3e68efd749dd1d386f3402b8bc0679255d))
 
 ### Bug Fixes
 
-* **gates:** stop the R2 review-input copy from laundering fixture secrets ([c63d350](https://github.com/event4u-app/agent-config/commit/c63d3502013e7bc261096d02e2c102a52e2f1ec7))
-* **golden:** teach the replay the delegated bad-verdict surface, re-lock baselines ([1ff480c](https://github.com/event4u-app/agent-config/commit/1ff480c6adeb88af32198a19d41838e2f9f5dcc0))
-* **envelope:** close the eight R2 review findings ([b0d153c](https://github.com/event4u-app/agent-config/commit/b0d153cb3084dac2a20093b73d28a3ce3c12324c))
-* **tests:** type the self-fix state helpers against the wire types ([60d2724](https://github.com/event4u-app/agent-config/commit/60d272419b07b5abea1ce6ffa430faf127fcde4a))
-* **handoff:** stop offering the caller its own empty session ([b72f772](https://github.com/event4u-app/agent-config/commit/b72f772a03207aa4feeded3acd5b68823e3fd79c))
-* **tests:** narrow regex captures in the capability parity block ([6f27615](https://github.com/event4u-app/agent-config/commit/6f276154b83781404d6551113c6d43e5cf040423))
-* **carrier:** a paths: disagreement is a delivery defect, not inert metadata ([318fb81](https://github.com/event4u-app/agent-config/commit/318fb818eb543f27fc4c3b138ce762dc8c9e9640))
-
-### Documentation
-
-* **review:** re-bind the findings artefact to the post-fix scope ([7a9bbbb](https://github.com/event4u-app/agent-config/commit/7a9bbbb2d7f4f04786771914bf7b3b68aba23e10))
-* **review:** record the R2 completion-review findings before fixing them ([27322e8](https://github.com/event4u-app/agent-config/commit/27322e876c76f96d6a67131f9e96a1759290a8f5))
-* **roadmap:** P2.2 build half shipped, its threshold unevaluated with a blocker ([1e478fd](https://github.com/event4u-app/agent-config/commit/1e478fdc317fdf12df001c1b98ef264df54a3742))
-* **contracts:** the flow contract documents the self-fix loop and the dod slot ([0ebaaba](https://github.com/event4u-app/agent-config/commit/0ebaabae9e12c75446efefc5be6f6b33b9d0435f))
-* **roadmap:** capability-answerability Phase 1 closed, blocker narrowed to one question ([c37af04](https://github.com/event4u-app/agent-config/commit/c37af045ecbf62b31e6878befdc88d763498d7a9))
-* **rules:** delegation-policy names the capability provenance check ([e05de77](https://github.com/event4u-app/agent-config/commit/e05de771d74e8284e2e64e91e33be66b1bd15e23))
-* **contracts:** the host-capability contract documents six fields, not five ([566b984](https://github.com/event4u-app/agent-config/commit/566b984681d1a1a5ef76643fa6f90f297c284e81))
-
-### Tests
-
-* **work-engine:** pin the loop contract and update the changed halt shapes ([eb88aa6](https://github.com/event4u-app/agent-config/commit/eb88aa695121a01c0c1e10de9804f1df3424c334))
+* **tests:** satisfy exactOptionalPropertyTypes in the two new spec files ([4f9cd9b](https://github.com/event4u-app/agent-config/commit/4f9cd9b46370920daa0a39aa5cb10a65b70e70a4))
+* **security-lint:** report the corpus a gate actually walked, not the shared default ([4420340](https://github.com/event4u-app/agent-config/commit/4420340cba6f5f091d2b836f21e747878fd71ed3))
+* **roadmap-archival:** never rewrite path strings inside frozen records ([8805814](https://github.com/event4u-app/agent-config/commit/88058140884f14c275737e0e612daee005db2bd7))
+* **review:** re-align the round-3 dispositions and escape the table pipes ([c569ba5](https://github.com/event4u-app/agent-config/commit/c569ba54c95e0d16ab315cb7a5223a33d34f3268))
+* **review:** close the round-3 findings, including four npm ci sites the verify regex could not see ([57867df](https://github.com/event4u-app/agent-config/commit/57867dfd5dbb2d00bbc02deace9c0312e999b423))
+* **review:** repair the round-2 findings, including a self-falsifying cost table ([90cb3ca](https://github.com/event4u-app/agent-config/commit/90cb3caff641d4bb4ef515813b7ec44af92d6636))
+* **lint-hedge-words:** harden the gate scope so the coverage ratchet holds ([d0c9d9e](https://github.com/event4u-app/agent-config/commit/d0c9d9e9b6812d59a7a89e8287bc4b89800aa28d))
+* **council:** let the reserved split label yield to a real option ([30328d8](https://github.com/event4u-app/agent-config/commit/30328d8811bd65f35f2bf42bc6b4fbe8cc76fa4c))
+* **council:** address the R2 completion-review findings ([02c786c](https://github.com/event4u-app/agent-config/commit/02c786c28646da19e8c9cee94ab49b826f91a611))
+* **review:** repair the seven completion-review findings ([efa7cc1](https://github.com/event4u-app/agent-config/commit/efa7cc1d150e0679678db22252dedbbcb838051f))
+* address the nine R2 completion-review findings ([02cdfbb](https://github.com/event4u-app/agent-config/commit/02cdfbb28f04e432cccff7c14ba9f12e177feb0e))
+* **ci-time-ratio:** correct the output-path docstring and register a task target ([1e28e18](https://github.com/event4u-app/agent-config/commit/1e28e1884ea8be49f6b8412e3b1c2d8c7240e9c9))
 
-### Other
-
-* **carrier:** re-bind to the post-merge scope — the fixed point ([884104c](https://github.com/event4u-app/agent-config/commit/884104c15d8257579a320932c671b489d96e7b06))
-* **carrier:** re-bind the R2 round to the shipping content, all 6 terminal ([9bbfbc3](https://github.com/event4u-app/agent-config/commit/9bbfbc35388f54c77fdae27315025843a6ea79f5))
-* **carrier:** R2 completion review round 1 — 6 findings, all open ([75be940](https://github.com/event4u-app/agent-config/commit/75be940f4f9bcbf71e1f1248467637fc51c6b63b))
+### Performance
 
-Tests: 12818 (+89 since 9.31.0)
+* **tests:** render the proof page once in build_proof.test.ts ([c4d699b](https://github.com/event4u-app/agent-config/commit/c4d699ba61c4ad5a1a19ca292b4d3218748069b5))
 
-## [9.31.0](https://github.com/event4u-app/agent-config/compare/9.30.0...9.31.0) (2026-08-10)
-
-### Release highlights
-
-<!-- Curated head: fill before merge, keep it under 10 lines, and leave `_none_` where it is genuinely the answer. The generated log below is unchanged. -->
-- **Behaviour changes:** _none_
-- **Default changes + migration:** _none_
-- **Security and correctness:** _none_
-- **Honest nulls:** _none_
-- **Known limitations:** _none_
-
-### Features
+### Documentation
 
-* **settings:** give the scripts read path the template-defaults layer ([a67a640](https://github.com/event4u-app/agent-config/commit/a67a640cbf64262d07026853b7b5ead21b509e5d))
+* **dispatch-safety:** close 3.1, 3.2 and 3.3, and record why 3.4 waits for producers ([9efe07e](https://github.com/event4u-app/agent-config/commit/9efe07e11b40eb5231d267f89968045af87163d9))
+* **dispatch-safety:** ADR-109 amendment 4, and close 1.3, 1.4 and 4.1 ([5f7e83a](https://github.com/event4u-app/agent-config/commit/5f7e83ad61488c1098ce5696731664b24a13ac2f))
+* **review:** declare the no-code-surface skip for the org-pack decision ([4a573c0](https://github.com/event4u-app/agent-config/commit/4a573c0565854f0f561de8258eb776b1f51eb1fd))
+* **council:** record the gate-scoped solo-attendance floor as ADR-224 ([686c23d](https://github.com/event4u-app/agent-config/commit/686c23df83b84c33f5bfac5257d1b187a0174f73))
+* **roadmap:** close estate-lifecycle Phases 2-4 and record the two refusals ([f64b7fc](https://github.com/event4u-app/agent-config/commit/f64b7fc87fd6c0701ff5601676f067af2fe024fe))
+* **governance:** separate the two lifecycle vocabularies and answer the deferred field ([212a38c](https://github.com/event4u-app/agent-config/commit/212a38c45622b9ee2329fe2929db2949a2c612de))
+* **decision:** decline the external pack source root ([87f8943](https://github.com/event4u-app/agent-config/commit/87f894310ec1b07d64d86cef7d9b8a3cb5595a0a))
+* **review:** disposition the round-3 findings ([c78afd5](https://github.com/event4u-app/agent-config/commit/c78afd510497d08b3684fc5fe42230b7a82b2af3))
+* **review:** disposition the round-2 findings and preserve both rounds ([59b4561](https://github.com/event4u-app/agent-config/commit/59b45613fdd7eb3ffef275b5062f36c366428015))
+* **review:** re-bind the findings after the gate-hardening fix ([0928422](https://github.com/event4u-app/agent-config/commit/092842250cb6b861e9cc217e0708bc63d411513b))
+* **review:** disposition the seven findings against the repair commit ([910f7c4](https://github.com/event4u-app/agent-config/commit/910f7c4e85886f194f214ceeafe06c2b479b154e))
+* **review:** record the completion-review findings for ci-economy ([d5e1fbd](https://github.com/event4u-app/agent-config/commit/d5e1fbdd3184adc4cd0c4f2a88776ae475e99422))
+* **roadmap:** resolve the council-integrity deferral into a READY follow-up ([94f25c2](https://github.com/event4u-app/agent-config/commit/94f25c25d7911ef9fc177db21dba0e294e7e143a))
+* **review:** re-bind the completion-review findings to the post-fix scope ([0208175](https://github.com/event4u-app/agent-config/commit/0208175e677af8db7632be58e7384209dceae45e))
+* **roadmap:** close 15 of 16 ci-economy steps, cancel the build fan-out ([3a3523d](https://github.com/event4u-app/agent-config/commit/3a3523de29b269b9b580fd9bd071d94cafa0a2cd))
+* **adr:** ADR-223 records no required-check demotion on cost grounds ([6e0d45a](https://github.com/event4u-app/agent-config/commit/6e0d45ae673f2cd8fa8c0e36f09110b4a573de7e))
+* **development:** correct the testing section and document the in-process runner ([53ce1f6](https://github.com/event4u-app/agent-config/commit/53ce1f6fa2ef56ff3bc1737c987dd6e918c81928))
+* **ci-cost-budget:** re-anchor the baseline to CI-recorded figures ([2972545](https://github.com/event4u-app/agent-config/commit/2972545e966319ebf4c632e1931ebc8c084b8816))
+* **review:** R2 completion-review findings for the authoring-contract branch ([d97c428](https://github.com/event4u-app/agent-config/commit/d97c42860ea1ce4b6b63a347c824f00a536d853f))
+* **failure-signatures:** stable ids, per-row drills, capability-claim row ([c18295e](https://github.com/event4u-app/agent-config/commit/c18295e0108955418eabf064502b591432cdeb1a))
+* **skill-writing:** reclassify two unenforced sections, add three patterns ([07d8ce1](https://github.com/event4u-app/agent-config/commit/07d8ce1198699833f0154112831ae7fcf4703ba0))
+* **roadmap:** close council-integrity 2.1 and 3.1, record two premise corrections ([088f7ab](https://github.com/event4u-app/agent-config/commit/088f7ab87de41ecbe21c46d7efefef1ed5f04a60))
+
+### Refactoring
+
+* one definition of the env kill-switch predicate ([689cc55](https://github.com/event4u-app/agent-config/commit/689cc5558a55ccab4728175955459973a5b03e4e))
 
-### Bug Fixes
+### Build
 
-* **carrier:** a frontmatter-only difference is not body divergence ([c48b6c8](https://github.com/event4u-app/agent-config/commit/c48b6c88cf274e3e24b0795d8fb4d19be7637557))
-* **hooks:** replace the fence regex with a scanner, and stop shipping a moving figure ([2fe01e7](https://github.com/event4u-app/agent-config/commit/2fe01e79804c37380e8c968b6efdf71b7d9d33fa))
-* **hooks:** per-session pin marker, atomic write, enforced read cap ([67abbb6](https://github.com/event4u-app/agent-config/commit/67abbb63028ab79fa6d04083523fbb6ffb9033c3))
-* **scripts:** measure the gate per TURN, and withdraw the figures that were not ([313e665](https://github.com/event4u-app/agent-config/commit/313e665354c1dd0512aceb21475dbc532f6b9ba5))
-* **hooks:** turn-end-gate — key the guard on the turn, not the prompt text ([b3392e6](https://github.com/event4u-app/agent-config/commit/b3392e60ceeae500b639040e8849b801fc661008))
+* **toolchain:** enable tsc incremental buildinfo and the eslint cache ([344bf97](https://github.com/event4u-app/agent-config/commit/344bf979e7c355c1a97d7407706f198fd5bf78ee))
 
-### Documentation
+### CI
 
-* **carrier:** the header says three classes and there are now four ([1955234](https://github.com/event4u-app/agent-config/commit/195523465db8fabb84be3f4ad27f8293cf5e4ac5))
-* **evidence:** classify the 109 carrier divergences, and correct two records ([c2fc109](https://github.com/event4u-app/agent-config/commit/c2fc109f7ca30f621455a9a754803055a0b1a0d9))
-* **evidence:** map the scripts settings read path against the server one ([858a256](https://github.com/event4u-app/agent-config/commit/858a25646352fd3603800e5657527591a00e16ae))
-* **review:** record the R2 findings on the fix pass, before touching it ([65aee68](https://github.com/event4u-app/agent-config/commit/65aee68691d85bdba928b4de65f4e28025c80852))
-* **review:** record the R2 findings for turn-end-gate before fixing anything ([9f204c5](https://github.com/event4u-app/agent-config/commit/9f204c55a73a45bada0baa7925bc000df7bee912))
+* **workflows:** add PR-scoped concurrency and flag the bare npm ci calls ([522c052](https://github.com/event4u-app/agent-config/commit/522c05283ea2f5d4b7ca65c5862d21520291f0db))
 
-### Build
+### Chores
 
-* **install:** rebuild the bundle without the worktree symlink poison ([741806c](https://github.com/event4u-app/agent-config/commit/741806c91a9774b1b521ae34487d271d0c5dee93))
-* **install:** refresh the committed install bundle ([7bbcfdc](https://github.com/event4u-app/agent-config/commit/7bbcfdce9355708ef6785ec640af64d23a5ac4a0))
+* **roadmap:** regenerate the dashboard after the second main merge ([92bf6fb](https://github.com/event4u-app/agent-config/commit/92bf6fb58e50170176a76d5892098b0d19d21e8c))
+* **dist:** project the reversibility-tag paragraphs into the command tree ([2d33951](https://github.com/event4u-app/agent-config/commit/2d339519eee0dd84b4b876de8222e5503c08d04c))
+* **tests:** drop two dead symbols left by the py2ts parity teardown ([cfecb9d](https://github.com/event4u-app/agent-config/commit/cfecb9dd1466af91791ad60fcdb768a85ab5cf4f))
+* **roadmap:** close council-integrity-followup, open the implementation plan ([6efce85](https://github.com/event4u-app/agent-config/commit/6efce8541f6e3ec48789d2e4dc6333adcfb8148d))
+* **roadmap:** archive install-lifecycle and re-depth its moved links ([6790d85](https://github.com/event4u-app/agent-config/commit/6790d857c0869cf2d08c8d74c149d94dcf0a346b))
+* **tests:** drop the dead tail in the discovery-manifest test ([fe258f9](https://github.com/event4u-app/agent-config/commit/fe258f9da91b594f06837694619a5f6934e5a0dc))
+* **roadmap:** archive the authoring-contract roadmap ([4aea8f7](https://github.com/event4u-app/agent-config/commit/4aea8f7a241a063ed371edd1f53fad31e56340a4))
 
-### Chores
+### Other
 
-* **roadmap:** close carrier-layer-convergence Phases 1-2 on the measurement ([bc42809](https://github.com/event4u-app/agent-config/commit/bc42809689634578e52501470f299822624e15ae))
-* **roadmap:** archive road-to-scripts-settings-defaults, complete ([e3af4c5](https://github.com/event4u-app/agent-config/commit/e3af4c56c1149a96254f1e7e3c1807a8f53744b5))
-* **budget:** re-baseline the packed-tarball cap, 6.9 to 7.8, maintainer-decided ([04ca7d8](https://github.com/event4u-app/agent-config/commit/04ca7d84f93c33999b02361c9baa4e61a181a07c))
+* **council-integrity:** re-bind after merging origin/main a second time ([dd58e0c](https://github.com/event4u-app/agent-config/commit/dd58e0c7c46c27ec0784621096c7f091fd29c44e))
+* **council-integrity:** re-bind after merging origin/main ([eb84ecb](https://github.com/event4u-app/agent-config/commit/eb84ecb6c0bee33c4d9ec43b85b50493badf4b26))
+* **council-integrity:** re-bind the findings artefact to the fixed scope ([267ff3f](https://github.com/event4u-app/agent-config/commit/267ff3f3b4dae46a82bc38a4d1399d93b0304e57))
+* **council-integrity:** record R2 completion-review findings before any fix ([ccf6650](https://github.com/event4u-app/agent-config/commit/ccf6650bbe7287a92e58177d6fe9ef8fc1cbb6a3))
 
-Tests: 12729 (+38 since 9.30.0)
+Tests: 13068 (+133 since 9.33.0)
 
 # Era: pre-4.0.0 — archived
 

package/dist/agent-src/commands/agent-handoff.md

@@ -167,9 +167,11 @@ Markdown, no host API). Required fields, in order:
 ## Contract owed
 {what the current phase must produce before yielding}
 ## Decisions
-- {decisions taken, with one-line rationale}
+- {decision taken, with one-line rationale} {optionally close the line with
+  `[reversible]` or `[irreversible]` — those two spellings exactly}
 ## Open questions
-- {unresolved items the next session must not silently drop}
+- {unresolved items the next session must not silently drop — each as a
+  question ending in `?`; write `none` when there genuinely are none}
 ## Next command
 {the single command or step to run first on resume}
 ```
@@ -177,7 +179,11 @@ Markdown, no host API). Required fields, in order:
 **Resume rule:** a workflow skill's step 0 checks for this file and resumes
 from its contract (mode-inference table) instead of re-deriving state; a
 long phase refreshes the file before yielding. Validated by
-`lint_handoffs.ts` when present — a missing required field is red.
+`lint_handoffs.ts` when present — a missing required field is red, and so is an
+`## Open questions` section that answers neither way (blank, or a bare `TBD` /
+`TODO` / `...`). A `?`-terminated question passes; so does an explicit `none` —
+the check exists to stop a blank section reading as an all-clear, not to force a
+question where there is none.
 
 **Critical-planning-file safety protocol** (applies to HANDOFF.md and agent
 roadmap edits): read the current file FIRST; take a timestamped backup copy
@@ -225,6 +231,12 @@ do not auto-execute.
 - **Branch name is critical** — always include it.
 - **Open tasks are critical** — the new chat needs to know what's left.
 - **Decisions are important** — prevents the new chat from re-asking settled questions.
+  A decision line may close with `[reversible]` or `[irreversible]`, which tells the
+  successor "we picked A over B, easy to revisit" apart from "we already migrated the
+  data". The tag is optional; those two spellings are the only accepted ones, and a
+  near-miss (`[reversble]`, `[Irreversible]`) is a validation error rather than a
+  silently untagged line. No claim is made that tagging improves resumption — that
+  stays the registered, unmeasured `envelope_resume_success` metric.
 - **File list is optional** — only include if the new chat will need to edit specific files.
 - **NEVER render a printed handoff as live markdown** — see Iron Law above.
 

package/dist/agent-src/commands/review/routing.md

@@ -59,9 +59,8 @@ changed-file list. The skill:
 4. Computes overall severity (`high` wins).
 5. Returns the structured block.
 
-If neither data file exists, the skill returns the generic fallback
-from [`reviewer-awareness`](../../rules/reviewer-awareness.md). That is
-valid output — do not invent owners.
+If neither data file exists, the skill returns the generic fallback from
+its own common role vocabulary. That is valid output — do not invent owners.
 
 ### 3. Present the result
 
@@ -121,7 +120,7 @@ After the block, ask:
 ## See also
 
 - [`review-routing`](../../skills/review-routing/SKILL.md) — the resolver
-- [`reviewer-awareness`](../../rules/reviewer-awareness.md) — role vocabulary + data-source rules
+- [`reviewer-awareness`](../../rules/reviewer-awareness.md) — the rule that routes here: paths and risk over seniority, primary + secondary on medium/high
 - [`review-routing-data-format`](../../../docs/guidelines/agent-infra/review-routing-data-format.md)
   — YAML schemas
 - [`create-pr-description`](../../skills/create-pr:description-only/SKILL.md) —

package/dist/agent-src/contexts/communication/rules-auto/reply-close-mechanics.md

@@ -23,6 +23,53 @@ hunts for it. One PR → one URL line.
 - Only when a PR was actually created/updated **this turn**. No PR → no link line.
 - The raw URL (not a masked markdown link) so it is unmistakable and clickable.
 
+## Blocker handover — when the reply leaves work stuck
+
+A work reply that stops because something is waiting on the **user** closes
+with the blocker, in the same one summary, above the PR-URL line. Not a second
+summary — this is what "what remains" looks like when what remains is not yours
+to do.
+
+`agent-config gates --reply` renders it. Call it, paste what it returns, and
+mirror the labels into the user's language. It prints **nothing** when nothing
+is owned by the user, so it is safe to call unconditionally: the "no blocker →
+no block" rule is a property of the command, not a judgement you have to make.
+
+Three obligations the command cannot carry for you:
+
+- **The options ARE the blocker decision — never a choice of substitute work.**
+  This is the measured failure, not a hypothetical one: a session that ended
+  every reply with well-formed numbered options still left its real blockers
+  unresolved for nine rounds, because each option block asked *"which work
+  should I do while your decision sits there"* (option 3 was reliably "nothing
+  more autonomously"). The user answered every one of them with a digit and
+  decided nothing. Offering the agent's next task is not a handover.
+- **The action travels, not a pointer to it.** "The three readings are in the
+  roadmap" is not a decidable option — the user cannot answer without reading
+  first. Where a real choice exists, write the alternatives out, one line each
+  for what speaks for them, then the single recommendation line
+  ([`user-interaction`](../../../rules/user-interaction.md)).
+- **Owner ≠ agent.** If you could clear it with a tool or a command, it is not
+  a blocker, it is unfinished work — do it now. A blocker whose owner is the
+  agent, announced instead of executed, is the promissory close that
+  [`verify-before-complete`](../../../rules/verify-before-complete.md) forbids.
+
+## Why the blocker handover is not a trailing checklist
+
+[`mandated-lines`](../../execution/mandated-lines.md) bans the end-of-reply
+checklist, and this shape has to answer that ban rather than sit beside it.
+
+- **It fires only when owed.** Nothing waiting on the user ⇒ nothing emitted,
+  enforced by the renderer. It never appears as evidence that blockers were
+  checked for — which is exactly the ceremony the ban targets.
+- **The end of the reply IS the decision point here.** The turn stops because
+  the decision is the user's. A line emitted where the decision lands is the
+  rule's own criterion, not an exception to it.
+- **One blocker in full, the rest as a count.** The blocking one is written
+  out; everything else is `N other decisions also wait on you` plus the
+  command. A roster is the failure mode — "twelve decisions across two
+  documents, ~850 lines" produced no decision at all.
+
 ## Why it does not collide with sibling rules
 
 - **`language-and-tone`** last-line rule: a raw URL is language-neutral, so it
@@ -34,6 +81,15 @@ hunts for it. One PR → one URL line.
   last among interactive elements; the PR URL sits below it as the final line and
   is not an "open question," so it does not trip the no-trailing-open-question
   rule.
+- **`ask-when-uncertain`** (one question per turn): reporting N blockers is
+  **zero** questions — it is a report. What the turn may carry is at most ONE
+  numbered-options block, for the one blocking decision, with one recommendation
+  line. Per-blocker questions are the violation that rule names.
+- **`no-cheap-questions`** IL4: the handover ends on `Done when:` — a decidable
+  signal, not "shall I continue once you've sorted this out?".
+- **`scope-control`** post-artifact hard stop: the steps address the **user**.
+  A handover that routes a slash-command back to the agent ("Blocker X — shall I
+  start Phase 1?") is the execution menu that rule forbids, wearing a new label.
 
 ## Failure modes
 
@@ -43,3 +99,12 @@ hunts for it. One PR → one URL line.
   scroll/search.
 - Masking the URL behind link text so it is not obviously the PR.
 - Adding a PR-link line when no PR was created this turn.
+- Closing on a blocker while offering the agent's next task instead of the
+  decision the blocker actually needs.
+- Handing over options as a count ("four options, see the template") or as a
+  file reference — the user cannot answer either without reading first.
+- Listing every open blocker flat, leaving the user to work out which one is
+  the one holding the chain.
+- Reporting the same blocker again, turn after turn, with no resolution path —
+  a repeat with nothing moved is itself the finding, and belongs in the reply
+  as such.

package/dist/agent-src/contexts/execution/orchestration-telemetry.md

@@ -68,6 +68,9 @@ v1 forward-compat rule on unknown fields).
 | `rules_used` | int \| null | **Lean-init extension (L6 demand signal).** Rules the worker actually applied/cited. Validation: never exceeds `rules_carried`. `null` = not measured. Counts only. |
 | `work_tokens` | int \| null | **Dispatch-economy extension.** Tokens the worker consumed AFTER init — the delta from first worker turn to envelope close. With `init_tokens`, yields the `dispatch_floor` ratio (`src/config/dispatch-economy-metrics.json`). Sourced from the transcript ledger (`_lib/cc_transcript.ts` `billable_input` semantics) by the report/backfill, never model-estimated without the tag below. `null` = not measured. |
 | `floor_provenance` | enum \| null | **Dispatch-economy extension.** Provenance of the `init_tokens`/`work_tokens` pair: `measured` (transcript ledger) · `estimated`. Auto-defaults to `estimated` when either field is present untagged — mirrors `token_delta_provenance`. `null` = neither field present. |
+| `model_requested` | string \| null | **Served-model-truth extension (`inbox-harvest-2026-08-b-ledger-truth` Phase 1).** The model id that was ASKED for — the id `tier_chosen`, `tier_source`, `session_tier` and the downshift cost-% are all derived from. `null` = not recorded. Id only, never free-form. |
+| `model_served` | string \| null | **Served-model-truth extension.** The model id the provider reported ANSWERING with. `''`/`null` when the transport reports none (every CLI client) — absent is the honest value, not a defect. Attribution only: never route or tier on it. |
+| `model_divergent` | bool \| null | **Served-model-truth extension.** DERIVED, never supplied: `true` when both ids are present and differ (an alias or provider substitution — every tier-derived figure on this line attributes to a model that never ran), `false` when both are present and match. `null` when either is absent, because a comparison that never happened must not read as "checked, and they matched". |
 | `return_channel_chars` | int \| null | **Dispatch-economy extension (Phase 6.3).** Serialized chars of the tool result returned into the orchestrator context (sync completions, hook-derived count — content never enters the record). Detector for "isolation win refunded through the return channel"; committed cap: `MAX_ENVELOPE_CHARS` (12,000) via `return_channel` in `dispatch-economy-metrics.json`. `null` = async ack. |
 
 These routing + cost fields are additive and optional — a line without them is

package/dist/agent-src/rules/delegation-policy.md

@@ -66,6 +66,12 @@ primitive → run in-session (clean no-op).
 2. **Tier-size** each slice (`lite|medium|high`) per the classification context —
    never the orchestrator's own session tier (the cost win is the per-call
    downsize; see [`subagent-routing`](../contexts/execution/subagent-routing.md)).
+   Where `subagents.model_ceiling` caps the tier and the slice does not fit under
+   it, the worker **escalates and never silently delivers the degraded result**;
+   the orchestrator re-slices, runs in-session, or surfaces the ceiling to the
+   human, and never raises the ceiling itself (class C — `settings:set` refuses
+   it by construction). Absent is *uncapped*, never a low cap. Full contract:
+   [`subagent-boundary § The model ceiling`](../docs/contracts/subagent-boundary.md).
 3. **Dispatch** via the form gate + matching `subagent-orchestration` mode
    (independent → do-in-parallel; ordered → do-in-steps; risk/correctness →
    + judge; UI-heavy → live-app judge, efficacy-gated).

package/dist/agent-src/scripts/archive_completed_roadmaps.ts

@@ -128,12 +128,42 @@ function _branch_touched_paths(root: string, base: string): Set<string> | null {
     return out;
 }
 
+/**
+ * Paths where a repo-relative path is DATA, not a live reference — so
+ * rewriting it falsifies a record instead of keeping a link resolvable.
+ *
+ * Two classes, and the second is why this predicate exists at all:
+ *
+ * - `agents/evidence/**` — a committed record of what was true, or of what a
+ *   reviewer was shown, at one point in time. Retroactively editing it makes it
+ *   claim something that did not happen, which is the failure
+ *   `evaluator-independence` exists to prevent ("the prompt ships with the
+ *   verdict" is worthless if a later sweep can edit the prompt).
+ * - `*.patch` anywhere — a patch's `diff --git a/… b/…` and `+++ b/…` lines are
+ *   the payload. Rewriting them detaches the patch from the commit it records
+ *   AND makes `git apply` target a path the diff never created.
+ *
+ * Measured 2026-08-11: archiving one roadmap rewrote four frozen artefacts
+ * under `agents/evidence/reviews/`, including the `diff.patch` headers of a
+ * recorded review input. Nothing caught it — the sweep reported it as
+ * "4 ref(s) migrated", which reads like success.
+ */
+const _FROZEN_RECORD_PREFIXES: readonly string[] = ['agents/evidence/'];
+
+/** Is `rel` a frozen record whose path strings must not be rewritten? */
+export function _is_frozen_record(rel: string): boolean {
+    const norm = rel.split(path.sep).join('/');
+    return _FROZEN_RECORD_PREFIXES.some((p) => norm.startsWith(p)) || norm.endsWith('.patch');
+}
+
 /**
  * Rewrite full-path references `old_rel` → `new_rel` in tracked files.
  *
  * Only the exact repo-relative path is rewritten (bare-filename mentions are
  * left alone). The archived file's own path never matches because the search
- * string is the un-archived path.
+ * string is the un-archived path. Frozen records are skipped — see
+ * `_is_frozen_record`; a link left dangling in a historical record is correct,
+ * because the record describes a tree that no longer exists.
  */
 function _inbound_ref_rewrite(
     root: string,
@@ -165,6 +195,10 @@ function _inbound_ref_rewrite(
             // skip the roadmap file itself
             continue;
         }
+        if (_is_frozen_record(rel)) {
+            // A record of a past state — rewriting it falsifies the record.
+            continue;
+        }
         const fp = path.join(root, rel);
         let text: string;
         try {

package/dist/agent-src/scripts/roadmap_gates.ts

@@ -18,10 +18,34 @@
  * Blocking a pipeline on "the user has an open decision" would turn a
  * visibility aid into a second thing to fight.
  *
+ * `--reply` is the same projection shaped for the END OF A CHAT REPLY rather
+ * than a terminal pane, and it exists because the pull channel above is only
+ * read by someone who already knows to ask. Transcript forensics over three
+ * consumer sessions found five "what do I do now?" follow-ups, every one of
+ * them next to a blocker the agent HAD reported — as a file reference, as a
+ * bare option count, or as a choice of substitute work. What the same corpus
+ * also shows is that a tool's output gets carried into the reply verbatim
+ * (the dashboard's step count appears word-for-word inside eight option
+ * blocks). So the fix is not to make the agent format better from memory; it
+ * is to hand it the finished text. See ADR-222.
+ *
+ * Two properties carry the contract mechanically:
+ *   - Nothing owned by the user open ⇒ EMPTY output. "No blocker → no block"
+ *     stops being a judgement call, so the command is safe to call
+ *     unconditionally at reply-close.
+ *   - Exactly ONE blocker is rendered in full — the one that unblocks most.
+ *     Every other one is a single trailing line. The user is never handed a
+ *     flat list to weigh themselves.
+ *
+ * Labels are English like the rest of the CLI surface; the agent mirrors them
+ * into the user's language when it lifts them into a reply, exactly as it does
+ * for `Recommendation:` / `Empfehlung:`.
+ *
  * Invocation (from project root):
  *   ./agent-config gates            # decisions owned by the user
  *   ./agent-config gates --all      # every open blocker, grouped by owner
  *   ./agent-config gates --json     # machine-readable
+ *   ./agent-config gates --reply    # reply-close form; empty when none
  */
 
 import * as path from 'node:path';
@@ -209,6 +233,76 @@ function render(entries: readonly Entry[], all: boolean): string {
     return lines.join('\n') + '\n';
 }
 
+/**
+ * Reply-close form: the one blocking decision in full, the rest as one line.
+ *
+ * Deliberately narrower than `render()`. `Status` and `Owner` are roadmap-file
+ * metadata that a reader of a reply cannot act on, so they are dropped; what
+ * survives is what the user has to DO and how they will know it is done. The
+ * five-field shape stays where it is enforced — in the roadmap file — and does
+ * not follow the blocker into prose it would only pad.
+ *
+ * Returns '' when nothing is owned by the user, so the caller can append the
+ * result unconditionally and get silence when silence is correct.
+ */
+function renderReply(entries: readonly Entry[]): string {
+    const mine = entries.filter((e) => needsUser(e.blocker.owner));
+    if (mine.length === 0) {
+        return '';
+    }
+
+    // `collectEntries` already sorted by unblocking weight, so the first entry
+    // IS the one that matters most. Picking it here rather than asking the
+    // caller to choose is the point: a flat list is the failure mode.
+    const lead = mine[0] as Entry;
+    const rest = mine.slice(1);
+    const lines: string[] = [];
+
+    lines.push(`Blocked: ${lead.blocker.blocks} — ${lead.roadmapRel}`);
+
+    const steps = regroupTodo(lead.blocker.todo);
+    const doLabel = isLegacy(lead.blocker) ? 'Blocked until:' : 'Do:';
+    if (steps.length === 0) {
+        lines.push(
+            `${doLabel} (no steps recorded — the blocker entry needs a **What to do:** list)`,
+        );
+    } else if (steps.length === 1) {
+        lines.push(...wrap(`${doLabel} ${steps[0] as string}`, WIDTH, '    '));
+    } else {
+        // Emitted as authored, never re-numbered: real blocker entries already
+        // carry their own `1.` / `2.` markers (the CI-enforced `What to do:`
+        // shape), so adding our own would produce "1. 1. …". An entry that
+        // arrives unnumbered stays unnumbered — that is a fact about the
+        // roadmap entry, and papering over it here would hide it from the
+        // person who could fix it.
+        lines.push(`${doLabel}`);
+        for (const s of steps) {
+            lines.push(...wrap(`  ${s}`, WIDTH, '     '));
+        }
+    }
+    lines.push(...wrap(`Done when: ${lead.blocker.resolvedWhen}`, WIDTH, '    '));
+
+    if (rest.length > 0) {
+        // A count, not a roster. Naming all of them reproduces the exact
+        // failure this form exists to end — "zwölf Entscheidungen bei Dir …
+        // rund 850 Zeilen", which the transcript shows produced no decision at
+        // all. The one that blocks is written out above; the rest are one
+        // command away, and saying so is shorter than listing them and more
+        // honest than hiding them.
+        lines.push('');
+        lines.push(
+            ...wrap(
+                `${rest.length} other decision${rest.length !== 1 ? 's' : ''} also wait${rest.length === 1 ? 's' : ''} on you, ` +
+                    'none blocking this — `agent-config gates`.',
+                WIDTH,
+                '    ',
+            ),
+        );
+    }
+
+    return lines.join('\n') + '\n';
+}
+
 function renderJson(entries: readonly Entry[], all: boolean): string {
     const pick = all ? entries : entries.filter((e) => needsUser(e.blocker.owner));
     return (
@@ -270,18 +364,28 @@ function main(argv?: readonly string[]): number {
     const args = argv ?? process.argv.slice(2);
     const all = args.includes('--all');
     const json = args.includes('--json');
+    const reply = args.includes('--reply');
 
     const repoRoot = _resolveRepoRoot(process.cwd());
     const roadmapRoot = path.join(repoRoot, 'agents', 'roadmaps');
     if (!_isDir(roadmapRoot)) {
+        // `--reply` stays silent here too: a project without roadmaps has no
+        // blocker to hand over, and a reply-close line saying so would be the
+        // ceremony the form exists to avoid.
         process.stdout.write(
-            json ? '{"needsYou":0,"other":0,"blockers":[]}\n' : 'No roadmaps directory — nothing to report.\n',
+            reply
+                ? ''
+                : json
+                  ? '{"needsYou":0,"other":0,"blockers":[]}\n'
+                  : 'No roadmaps directory — nothing to report.\n',
         );
         return 0;
     }
 
     const entries = collectEntries(roadmapRoot);
-    process.stdout.write(json ? renderJson(entries, all) : render(entries, all));
+    process.stdout.write(
+        reply ? renderReply(entries) : json ? renderJson(entries, all) : render(entries, all),
+    );
     return 0;
 }
 
@@ -308,5 +412,5 @@ if (_isCliEntry()) {
     process.exitCode = main();
 }
 
-export { main, needsUser, regroupTodo, wrap, collectEntries, render, renderJson };
+export { main, needsUser, regroupTodo, wrap, collectEntries, render, renderJson, renderReply };
 export type { Entry };