skills/github-review-workflow/SKILL.md
Export a PR's clean inline review comments, CodeRabbit outside-diff comments, and CodeRabbit nitpicks into local files, then triage review feedback through a stack-aware orchestrator workflow with durable reply drafts for follow-up PRs. Use when given a GitHub pull request URL and asked to work through review comments without relying on noisy raw API blobs.
npx skillsauth add rarestg/rarestg-skills github-review-workflowInstall 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.
Export inline PR review threads into a local queue, export CodeRabbit
outside-diff comments and nitpicks into separate local-only queues, then follow
this skill's stack-aware orchestration workflow in references/SOP.md.
gh installed and authenticatedpython3 available in the shellSet SKILL_DIR to this skill's installed or repo-local path, then run:
python3 "$SKILL_DIR/scripts/export_github_review_comments.py" <pr-url>
Useful export flags:
--include-resolved to reprocess already-resolved threads--include-ai-prompts to preserve embedded AI-agent prompt sections; they
are stripped by defaultRead the full bundle before editing:
$SKILL_DIR/references/SOP.md.github-review-workflow/pr-<number>-<slug>/README.md.github-review-workflow/pr-<number>-<slug>/context/01-coderabbit-walkthrough.md if it exists.github-review-workflow/pr-<number>-<slug>/todo/.github-review-workflow/pr-<number>-<slug>/outside-diff/ if it exists;
triage them as local-only review-summary items because GitHub could not post
them inline.github-review-workflow/pr-<number>-<slug>/nitpicks/ if it exists;
triage them as lower-priority local-only items unless the user asked to
handle them nowAddressed in <fix-pr-url>: <summary>.Not taking this change: <reason>.Queue accepted stacked-review fixes before the follow-up PR exists:
python3 "$SKILL_DIR/scripts/review_reply_queue.py" add-fixed <review-item-file> \
--summary '<summary>'
After creating the new top-of-stack PR, attach the PR URL and post pending replies:
python3 "$SKILL_DIR/scripts/review_reply_queue.py" post-pending \
--source-pr <reviewed-pr-number> \
--fix-pr-url https://github.com/<owner>/<repo>/pull/<fix-pr> \
--dry-run
python3 "$SKILL_DIR/scripts/review_reply_queue.py" post-pending \
--source-pr <reviewed-pr-number> \
--fix-pr-url https://github.com/<owner>/<repo>/pull/<fix-pr>
Declined comments can use the same queue. add-declined prints a draft ID; use
that ID for immediate posting:
python3 "$SKILL_DIR/scripts/review_reply_queue.py" add-declined <review-item-file> \
--reason '<reason>'
python3 "$SKILL_DIR/scripts/review_reply_queue.py" post <declined-draft-id> --dry-run
python3 "$SKILL_DIR/scripts/review_reply_queue.py" post <declined-draft-id>
Use post_github_review_followup.py only as a direct single-item escape hatch.
tools
Break large code changes into small, stacked pull requests using vanilla git and the gh CLI. Auto-trigger when implementing a feature or change that spans multiple logical steps, touches several files, or will exceed ~200 changed lines. Also trigger on "stack PRs", "break this into smaller PRs", "stacked diffs", or "create a PR stack". Do NOT trigger for single-file fixes, small bug fixes, or changes under ~200 lines that are a single logical unit.
development
Run reusable Graphify-led architecture analysis for codebases using semantic graphs, optional subagent extraction, graph synthesis, source-search validation, graph-shape review, and follow-up refactor planning. Use when asked to analyze repo architecture, god nodes, surprising edges, topology, module boundaries, or graph-derived cleanup/refactor opportunities.
testing
Run a meaningful coding ticket through a delegated delivery workflow: tighten the ticket, assign one ticket owner, delegate implementation, get an independent review, scan for high-value simplification, validate the result, and return a compact outcome packet. Use when the user wants structured agent execution with clear scope, ownership, and review rather than a single-pass implementation. Triggers on: "delegate this ticket", "use a sub-PM", "run this through worker and reviewer", "own this ticket end to end", "send this for independent review", or "close this ticket out with review". Skip trivial fixes and tasks that are still too vague to delegate.
data-ai
Operating protocol that uses Taskwarrior as the continuity layer across conversations. Grounds every interaction in real task state before acting. Activate on "morning" / "briefing" for daily overview, on "status" / "what's next" for current state, when capturing tasks ("remind me to", "add task", "I need to"), when recovering from stalls, or when executing structured work across projects, repos, and life domains.