Twinned from https://github.com/cfsh/ninefold/issues/791 by tools/beads/import_github.py's reconcile pass.
Suggestions from ninefold-reviewer's review of #788. Below the gate's severity floor, recorded rather than dropped. Unclaimed; take or close freely.
tools/pr/perimeter_check.py:161 — _needs_app()'s guidance still says *"let the dispatcher do it — it publishes this check after every per-PR sweep, including on-demand review_now.py pulls."* Functionally still true (a review_now.py run does publish), but "sweep" and "the dispatcher" are the vocabulary #788 retired everywhere else. Same file, :61 and :285, carry WAKE: STALE-DISPATCH in comments — those are provenance for why the check exists and are fine to leave.tools/tests/test_dispatch.py now ends without a trailing newline (git flags it as "No newline at end of file" in the patch). The python lane is unittest only, so nothing fails on it — it just makes the next diff on that file noisier than it needs to be.DEVEL.md's gotcha now reads "~/.ninefold-reviewer/ is created for you by review_now.py". It is dispatch.py that mkdirs locks/, logs/ and worktrees/; review_now.py creates it only transitively. True enough for an operator, imprecise for someone debugging a permissions failure.notes/history.md has no entry for the deletion, and three of its sections now describe rules that no longer exist: *"Why the dev asks for review instead of the cron re-reviewing (#702)"* (:500), *"STALE-DISPATCH, and the false alarm that followed it"* (:665), *"The cron dies silently without self-healing"* (:702). Deliberately not raised in the gate — history.md is case law, not a rule source, and a rule always wins over it. Listed here only because DEV_PRINCIPLES C6 names it as the home for the tombstone prose that F5 in the review flagged, so the two items pair naturally if anyone takes either.None.
No comments.