# Cross-review — root Markdown cleanup, follow-up

Job: `fd874949`  
Reviewer: `planner-reviewer-codex-01` (`planner,reviewer`)  
Baseline: `07cae71`, current working tree including the staged roadmap move.

No blocking findings remain. Both findings from review `2601270d` are resolved by documentation changes that match the implementation.

## Previous findings verified as resolved

1. **Optional dispatch-token comparison — resolved (`MESSAGING.md:350`).** The paragraph now distinguishes the token stored for an attempt from the optional `expected_dispatch_token` supplied by callers. It explicitly states that ordinary publisher status updates do not supply the precondition. This matches `lib_py/delegation/engine.py:181` and `:270`, the publisher update path in `lib_py/delegation/mqtt_common.py:435`, and inferred completion in `lib_py/loop/autocomplete.py:224`. Token generation and transition ordering also match `engine.py:296-300`.

2. **Subscriber readiness fallback — resolved (`MESSAGING.md:354`).** The handshake bullet now describes a bounded readiness attempt, warning and continuation after the polling budget expires, and continuation on early exit status 0 or 3. These statements match `multi-agent-mux-delegate-job/multi-agent-mux-delegate-job:172-194`; the previous unconditional readiness guarantee has been removed.

## Verification and review scope

- Compared the complete embedded diff payloads of jobs `2601270d` and `fd874949`: only the two passages above and the corresponding `MESSAGING.md` blob hash changed. The previously reviewed roadmap move, backlog consolidation, bilingual trees, bootstrap updates, and archived reports are unchanged.
- Verified every file included in the new brief against its recorded new Git blob hash; all match the current files.
- `git diff HEAD --check`: clean.
- No changes under `.agents/skills/`, `tests/`, or `deploy/`.
- Root Markdown inventory remains the required eight files. Backlog section counts remain **1 / 3 / 1 / 0 / 31**, with no completed blocks in sections 2 or 3.
- Rechecked **34 local link targets**, including six references tracked inside the `nats-docker` submodule: all exist and are tracked. The plan's parent-repository-only tracking check produces submodule false positives; validating those targets within their owning repository resolves them.
- Prior independent preservation checks remain applicable because the relevant files are byte-identical to the preceding review: O-1 content, moved completed blocks, legacy diagnosis details, README tree parity, and roadmap links showed no content-loss defect.
- Prior independent focused regression result from job `2601270d`: **49 passed in 54.88s**, using `tests/test_deploy_layout.py`, `tests/test_o1_rebuttal.py`, `tests/test_deploy_freshness.py`, and `tests/test_version_consistency.py`. This run was not repeated for the two prose corrections. The attached peer report's full-suite result of 750 passing tests and minimal/full installation checks remain attributed peer evidence, not an independent rerun in this job.

The two requested wording corrections close the review. No runtime changes or planner escalation are needed.

[VERDICT: PASS]