chore(ci): repin the own-CI reviewer to claude-sonnet-5

[tigers1997]

What & why

One line: the own-CI reviewer moves from claude-sonnet-4-6 to claude-sonnet-5. Same tier, same list price — the v2.7.0 maintainer cost decision stands unchanged. This is a currency bump, not a re-opening of that call.

Why it is its own PR, and not part of the currency stack. It started life inside chore/upstream-currency (#92), which turned out to break the review gate on the entire stack:

Skipping action due to workflow validation: Workflow validation failed. The workflow file must exist and have identical content to the version on the repository's default branch.

anthropics/claude-code-action@v1 refuses to run when the PR's review.yml differs from the default branch's copy — a sensible guard against a PR rewriting the workflow that reviews it. The failure mode is the nasty kind: the action skips rather than fails, so there is no red check. #92–#95 all sat there looking like they were waiting on a reviewer that was never going to start. Confirmed it wasn't a credentials problem first — #88–#91 posted real VERDICT comments through 2026-06-27 on the same secret.

So the repin has been lifted out of that stack (all four branches rebased; review.yml is now byte-identical to main on each) and lands here instead, where verdict-gate's self-bootstrap escape hatch is the designed handling for exactly this case.

Worth writing down: any future review.yml change has to be its own PR for the same reason. The CHANGELOG entry says so, since the symptom gives no hint of the cause.

Type of change

• feat — new module / skill / feature (minor bump)
• fix — bug fix (patch bump)
• docs — documentation only
• chore — tooling, CI, release plumbing
• refactor — no behavioral change
• BREAKING

Scope

• One logical change. One commit, one token changed in review.yml plus its CHANGELOG entry.
• Modules affected: none — own CI only, nothing under templates/ moves.
• Personas affected: none.

Tests

• python3 configure.py --check passes locally.
• Verified the committed review.yml blob is byte-identical to main's apart from the model token — no line-ending or whitespace churn.
• No fixture applies; this is a one-token CI change.

CHANGELOG

• Added an entry under ## Unreleased.
• I will SHA-anchor the entry after merge.

License & NOTICE

• My contribution is my own work.
• No AGPL-incompatible code.
• No third-party code added.
• templates/discipline-skills/ untouched by this PR.
• LICENSE / NOTICE untouched.

Signing

• All commits are signed. — Not signed (no key on the authoring machine). A squash-merge signs the result.
• Conventional Commits prefix.

I understand

• An automated AI review will run on this PR. — It will not, and cannot: this PR changes review.yml, which is precisely the condition that makes the action self-skip. That is the self-bootstrap case CONTRIBUTING.md describes, and it needs a maintainer's eye on the one-line diff instead.
• No merge is possible while any required check is red.
• My contribution rights are described in CONTRIBUTING.md.

---

Merge this LAST, after #92–#95. The same validation rule cuts both ways: the moment this lands, main's review.yml says claude-sonnet-5 while #92–#95 still carry claude-sonnet-4-6, and the action self-skips on them all over again. They have already been reviewed against the currently-pinned model and passed, so there is nothing to gain by going first and a re-broken gate to lose. Order: #92 → #93 → #94 → #95 → this.