pstack: watch-pr --status-only retries for minutes on a repo with no checks #750

On a PR in a repository that reports no checks, watch-pr --status-only never prints a status. It prints

RETRY: GitHub status query failed; retrying in 60s
detail=could not read PR checks: fast path exit=1; GraphQL rollup was empty; no checks reported on the '<branch>' branch

and retries after 60, 120 and 240 s before giving up. Babysit's check mode calls --status-only and expects one status line, so the check stalls for minutes.

Repro. Open a PR in any repository with no CI or check apps, then run:

skills/poteto-mode/scripts/watch-pr/watch-pr --owner <owner> --repo <repo> --pr <n> --status-only

Cause. When both gh pr checks and the GraphQL rollup are empty, resolveChecks throws a retryable ChecksUnavailable (github.ts L630). readSnapshot has no state for a PR without checks (policy.ts L71). runSimple sends --status-only through pollUntilTerminal with the full --max-query-errors budget (policy.ts L513), so the error is retried with backoff.

A fix. A downstream fork fixed it in two small PRs, with tests:

• fix(watch-pr): make --status-only read once and report a PR with no checks MaxFabian25/pstack-harness-adaptations#2 adds a ci-none CI state ("no checks reported") that only --status-only produces, and gives --status-only a one-error budget so it never sleeps.
• watch-pr: no review tick without checks; say --status-only reads once MaxFabian25/pstack-harness-adaptations#5 shows "no review checks" instead of ✅ in the --pretty Review column for such a PR, and says in the help that --status-only reads GitHub once.

Polling mode is deliberately unchanged. A head that was just pushed has no checks yet either, so treating "no checks" as ready while polling would report READY too early.

Seen on pstack 0.15.25. policy.ts and github.ts are unchanged on main at 190d287.