Conciseness

The 19-line body contains no concept explanations and every sentence carries repo-specific knowledge Claude cannot infer ("A PR into `next` gets its body from `node .agents/pstack/land.mjs`. An agent that writes the body itself runs on a branch older than `next`'s agent rules.") or direct instruction, matching anchor 5's lean-and-efficient standard.

Actionability

Every step ships a copy-paste-ready command (`git fetch <remote> next && git rev-list --count HEAD..<remote>/next -- AGENTS.md .agents .claude .codex`, the gh pr list/view/edit invocations, `node .agents/rules/contributor-setup/scripts/doctor.mjs`), and the doctor script was verified to print `fix:` lines and `ok:` as the body claims; the only delegation ("as the `plan-page` skill's PR section says, with `--into <file>`") is an explicitly signaled conditional branch, matching anchor 5.

Workflow Clarity

The four steps are clearly sequenced with explicit gates and feedback loops even though the operation is batch (multiple PRs): "When it prints more than 0 ... and stop here", "apply what it prints until it prints `ok:`", and "This step is done when step 3's command prints nothing", so the batch-operation cap does not apply and anchor 5 fits.

Progressive Disclosure

For an under-50-line skill the single-bundle-file structure is ideal: scripts/doctor.mjs exists in the bundle and is referenced exactly where needed via its repo path, and external pointers (`.agents/pstack/rules/land.md`, the `plan-page` skill) are one level deep and clearly signaled, matching the simple-skill anchor-5 case.

Specificity

Two concrete, precisely scoped actions ("Bring a branch that predates next's agent rules up to date once", "fix the bodies of your open PRs into next") sit above anchor 3's generic 1-2 actions but short of anchor 4's 'several specific actions'; the second-person "your open PRs" triggers the rubric's -1 voice penalty, landing at 3.

Completeness

Explicitly answers both: what ("Bring a branch ... up to date once, then fix the bodies of your open PRs") and when ("Use when the PR body check fails or when an agent opened a PR into next by hand") with two concrete trigger phrases, matching anchor 5 rather than 4 since the 'when' needs no added specificity.

Trigger Term Quality

"PR body check fails" and "agent opened a PR into next" are natural phrases a user in this workflow would say, matching anchor 4's good-but-incomplete coverage; anchor 5 is not reached because synonyms like "pull request" (vs. only "PR") or "update branch" are absent.

Distinctiveness Conflict Risk

"next's agent rules", "PRs into next", and "PR body check" define a narrow repo-specific niche with distinct triggers and virtually no overlap risk with other skills, matching anchor 5.