User expects iterate-pr workflow — fix CI, resolve all bot+human comments, repeat

Iterate-PR workflow: When asked to keep running on a PR, loop: (1) Push fixes. (2) Wait for CI — including 'Sentry Seer' and 'Cursor BugBot' jobs (they may be NEUTRAL/pending; wait until they complete). (3) Get failing logs via gh run view --log-failed --job $(gh pr checks $PR_NO --json state,link -q '.[] | select(.state == "FAILURE").link | split("/")[-1]'). (4) Fix failures. (5) Get unresolved PR comments via GraphQL (replace owner/name/number). (6) Address each — fix code or reply with rationale. (7) Resolve threads. (8) Repeat until zero CI failures and zero unresolved bot+human comments.