chore(porch): 271 state commits stranded after #274 merged

[pseudoseed]

Refs #233. Refs #271.

Porch writes its state commits after the PR merges, so the four below could not have been
in #274:

e383ffc05 chore(porch): 271 protocol complete
869474aee chore(porch): 271 pr gate-approved
3b847c4ab chore(porch): 271 pr gate-requested
53c2c7961 chore(porch): 271 PR #274 merged

This is #233 — porch state commits stranded on a completed builder branch — happening again,
with the same shape.

The diff is one file

codev/projects/271-safety-adjacent-an-architect-c/status.yaml. Main's copy still reads
phase: pr with the gate pending; this branch reads phase: verified with
approved_at: 2026-08-31T17:19:19 and merged: true on the PR history entry. This branch is
the correct state and main is stale.

What is not in the diff, and why the branch had to be updated first

origin/main was eleven commits ahead of this branch — #274's own merge commit plus all of
bugfix-264 (#270), which merged in between. Opening this PR from the branch as it stood would
have shown those eleven as deletions: a state-commit PR that silently reverted someone else's
merged work, and it would have looked small in the title.

So origin/main was merged into the branch first. That is not a cherry-pick and it does not
touch main; it is the branch catching up to what it was already merged into. After it, the diff
against main is the single file above and nothing else.

Test plan

• No source file changes — the diff is one YAML state file
• Branch is up to date with origin/main, so the PR reverts nothing
• The state it carries was written by porch, not by hand

🤖 Generated with Claude Code