skills/gh-pr-review-fix/SKILL.md
Fix unresolved GitHub PR review threads with codex-dev: fetch hosted state, verify, patch, validate, commit, push, and close out.
npx skillsauth add bjornmelin/dev-skills gh-pr-review-fixInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
Use this as the sole active GitHub PR review remediation workflow. The canonical implementation is codex-dev; this skill is the operating procedure.
When the user invokes $gh-pr-review-fix, run the full closeout loop unless blocked:
AGENTS.md and inspect git status --short.codex-dev --json pr review start --repo <owner/repo> --number <pr> --freshcodex-dev pr review apply-suggestions reports an exact hunk match.references/closeout-and-commits.md.codex-dev --json pr review closeout --repo <owner/repo> --number <pr> --worklist <json> --expected-head-sha <pushed-head> --commit <sha> --validation-command <cmd> --applycodex-dev --json pr review start --fresh --repo <owner/repo> --number <pr> and continue until zero actionable unresolved items remain or a real blocker appears.references/closeout-and-commits.md before committing.codex-dev --json commit plan to inspect semantic groups in a mixed tree.codex-dev --json commit validate --subject "<subject>" before committing.codex-dev commit validate owns forbidden process-wording checks; do not maintain a second local phrase list here.$review-remediation.codex-dev --json pr readiness plus GitHub status checks.gh workflow-log remediation.codex-dev.pr-review-worklist.v1 worklistcodex-dev.pr-review-closeout.v1 thread-to-commit closeout evidencecompleted, blocked, or no-opreferences/closeout-and-commits.mddevelopment
Pre-PR multi-model review, parallel opus and codex exec adversarial lanes, then adversarial verification of merged findings. Read-only. Use before shipping nontrivial diffs.
tools
Independent gpt-5.6 diff review via the Codex CLI, normal or steerable adversarial with JSON findings. Use before shipping nontrivial changes.
development
Delegate implementation, investigation, or bulk work to gpt-5.6 codex via pinned codex exec. Use for clear-spec builds, migrations, debugging, or any task MODELS.md routes to codex.
development
Adversarial pre-mortem: imagine the plan failed, work backwards to surface risky assumptions + irreversible bets, then harden them. Proactively offer it (after the current request; confirm first) before a hard-to-reverse or one-way-door call (API, schema, framework, a hire), an all-upside plan, or unvalidated assumptions. Also on request.