[Jammy2211]

Summary

The PyAutoConf → PyAutoNerves rename (package autoconf → autonerves) was
completed in the reader-facing docs but left eight sites naming the old
repo/package. Four were live defects, all silent — nothing failed, the surfaces
just stopped seeing the Nerves repo.

The root cause sat under the symptom the prompt reported. repo_aliases split
one repo across two dead keys: @PyAutoConf normalised to autoconf and
@PyAutoNerves to pyautonerves, and neither is in the body map, so neither
resolved. autonerves is now the single canonical key with the other three
spellings aliased onto it — the shape the existing pyautobuild: pyautohands
entry already set. That also collapses the three target_default_wiki rows that
existed only to paper over the split.

The other three live defects:

• test_witness pointed at PyAutoConf/test_autoconf, reachable only
  through the stale alias. pyauto-brain refactor therefore reported
  [unwitnessed: pyautonerves] and advised "strengthen tests first" for a repo
  that has a real suite.
• release.relevant_repos listed PyAutoConf, and nightly.sh fetches
  each name verbatim. Whether a rename redirect carried that call is unverified
  (see below); naming PyAutoNerves removes the dependency rather than trusting
  it.
• hygiene.sh timed import autoconf, whose non-zero rc is recorded as
  autoconf:n/a and skipped — so Nerves import timing went unmeasured.

Four comment/prose sites follow. Two were stale beyond the name:
ensure_workspace_labels.sh and skills/repo_cleanup/reference.md both mapped
repos to rhayes777/Jammy2211 owners the migration retired — the labels
script's own live list already disagreed with its header. Both now defer to the
body map's github: field instead of hardcoding an owner.

Fixes the headline finding of
PyAutoMind/draft/maintenance/pyautobrain/refactor_witness_map_missing_autonerves.md;
that prompt's broader "audit the other organ/library repos" ask is deliberately
not absorbed here and stays a separate task.

API Changes

None. PyAutoBrain is an organ, not a released library, and nothing imports it —
these are internal policy maps and agent prose. config/policy.yaml is a
declared config surface, so the key changes are noted for any adopting fork:
test_witness and target_default_wiki are keyed autonerves (was autoconf),
and release.relevant_repos names PyAutoNerves. Old spellings keep working
through repo_aliases.

Test Plan

• python3 -m pytest tests/ -q → 467 passed (462 baseline + 5 new)
• python3 PyAutoMind/scripts/repos_sync.py --check → all 11 checks OK
  (its tenant-firewall leg caught my own test comments hardcoding the org
  name; fixed before commit)
• bash -n on all four touched shell scripts
• Live conductor runs against the changed policy — refactor, bug,
  sizing all behave

Live proof of the headline defect being fixed — bin/pyauto-brain refactor on
this branch:

Repos affected:       autofit, autonerves
Witnesses:            {'autofit': 'PyAutoFit/test_autofit',
                       'autonerves': 'PyAutoNerves/test_autonerves'}

No [unwitnessed: pyautonerves].

New test coverage

• tests/test_policy_seams.py — all four spellings normalise to one key that is
  in KNOWN_REPOS; the witness map resolves under each; and every name in
  release.relevant_repos must exist in the body map. That last one is the
  drift guard whose absence let this survive the rename.
• tests/test_activity_gate.py — fixture moved off the old name, plus a case
  asserting a PyAutoNerves commit qualifies a night (the prompt's acceptance
  criterion, offline and deterministic).
• tests/test_nerves_rename_complete.py (new) — enforces the prompt's grep
  acceptance repo-wide, allowlisting only the sites deliberately about the
  rename, and failing if an allowlist entry goes dead so it can't rot into a
  blanket waiver. It found the eighth site (skills/OPERATIONS.md) on its
  first run.

Two things a reviewer should check rather than take on trust

1. Heart was never consulted. pyauto-heart is not on PATH in this
   web-github session and PyAutoHeart is not in its repo scope, so
   bin/pyauto-brain vitals returns not-found. The per-repo suite stood in per
   the WORKFLOW.md fallback and is green, but no YELLOW was acknowledged and
   no RED was seen, because nothing was asked.

2. The commit message says "157-test suite… verified live" — the count is
   weaker than that reads. The path PyAutoNerves/test_autonerves is
   verified properly: cloned PyAutoLabs/PyAutoNerves@main and confirmed the
   directory exists beside the autonerves package with 17 test_*.py files.
   The count is cited from PyAutoMind/complete/2026/08/cli-noise-autonerves-batch.md
   and was not re-run here. Nothing in the change depends on it — only the path
   is load-bearing — but the commit states the number without that qualifier and
   pushed history isn't rewritten in this org, so the correction lives here.

Gate at sign-off

Effective autonomy supervised (header supervised, bug cap supervised).
Plan was pre-approved before any edit and is on #267, unmodified since.
Tests PASS (467) · smoke n/a (organism repo, no downstream script surface) ·
review CLEAN with claim dispositions on
#267 ·
Heart NOT EVALUATED (above). Human signed off on PR-open; merge is yours.

No pending-release label — PyAutoBrain is outside the release chain.

Closes #267

🤖 Generated with Claude Code

https://claude.ai/code/session_01MPhEXMoPYa7H1sNHWR2GNW

---

Generated by Claude Code