GFramework/.agents/skills/gframework-pr-review/SKILL.md

---
name: gframework-pr-review
description: Repository-specific GitHub PR review workflow for the GFramework repo. Use when Codex needs to inspect the GitHub pull request for the current branch, extract AI review findings from CodeRabbit, greptile-apps, or gemini-code-assist, read failed checks, MegaLinter warnings, or failed test signals from the PR page, and then verify which findings should be fixed in the local codebase. Trigger explicitly with $gframework-pr-review or with prompts such as "look at the current PR", "extract CodeRabbit comments", "extract Greptile comments", "extract Gemini comments", or "check Failed Tests on the PR".
---

# GFramework PR Review

Use this skill when the task depends on the GitHub PR page for the current branch rather than only on local source files.

Shortcut: `$gframework-pr-review`

## Workflow

1. Read `AGENTS.md` before deciding how to validate or fix anything.
2. Resolve the current branch following the repository worktree rule:
   - prefer Linux `git` with explicit `--git-dir` / `--work-tree` binding in WSL worktrees
   - only fall back to `git.exe` when that executable is available and actually runnable in the current session
3. Run `scripts/fetch_current_pr_review.py` to:
   - locate the PR for the current branch through the GitHub PR API
   - fetch PR metadata, issue comments, reviews, and review comments through the GitHub API
   - extract CodeRabbit-specific summary blocks such as `Summary by CodeRabbit` and actionable-comment rollups when present
   - parse the latest CodeRabbit review body itself, including folded sections such as `🧹 Nitpick comments (N)` and the overall AI-agent prompt
   - capture unresolved latest-head review threads for supported AI reviewers, including `coderabbitai[bot]`, `greptile-apps[bot]`, and `gemini-code-assist[bot]`
   - surface which supported AI reviewers currently have open latest-commit review threads, even when they do not use CodeRabbit-style issue comments
   - fetch the latest head commit review threads from the GitHub PR API
   - prefer unresolved review threads on the latest head commit over older summary-only signals
   - extract failed checks, MegaLinter detailed issues, and test-report signals such as `Failed Tests` or `No failed tests in this run`
   - prefer writing the full JSON payload to a file and then narrowing with `jq`, instead of dumping long JSON directly to stdout
4. Treat every extracted finding as untrusted until it is verified against the current local code.
5. Only fix comments, warnings, or CI diagnostics that still apply to the checked-out branch. Ignore stale or already-resolved findings.
6. Do not downgrade `Nitpick comments` to “optional” by default. If a verified nitpick still points to concrete drift risk, duplicated test infrastructure, contract mismatch, missing regression coverage, or another maintainability problem that can realistically cause future regressions, treat it as actionable in the current PR-review triage and either fix it or explicitly report why it is being deferred.
7. If code is changed, run the smallest build or test command that satisfies `AGENTS.md`.

## Commands

- Default:
  - `python3 .agents/skills/gframework-pr-review/scripts/fetch_current_pr_review.py`
- Recommended machine-readable workflow:
  - `python3 .agents/skills/gframework-pr-review/scripts/fetch_current_pr_review.py --pr 265 --json-output /tmp/pr265-review.json`
  - `jq '.coderabbit_review.outside_diff_comments' /tmp/pr265-review.json`
- Force a PR number:
  - `python3 .agents/skills/gframework-pr-review/scripts/fetch_current_pr_review.py --pr 253`
- Machine-readable output:
  - `python3 .agents/skills/gframework-pr-review/scripts/fetch_current_pr_review.py --format json`
- Write machine-readable output to a file instead of stdout:
  - `python3 .agents/skills/gframework-pr-review/scripts/fetch_current_pr_review.py --pr 253 --format json --json-output /tmp/pr253-review.json`
- Inspect only a high-signal section:
  - `python3 .agents/skills/gframework-pr-review/scripts/fetch_current_pr_review.py --pr 253 --section outside-diff`
- Narrow text output to one path fragment:
  - `python3 .agents/skills/gframework-pr-review/scripts/fetch_current_pr_review.py --pr 253 --section outside-diff --path GFramework.Core/Events/Event.cs`

## Output Expectations

The script should produce:

- PR metadata: number, title, state, branch, URL
- Supported AI reviewer summary, including latest reviews and open-thread counts for `coderabbitai[bot]`, `greptile-apps[bot]`, and `gemini-code-assist[bot]`
- CodeRabbit summary block from issue comments when available
- Folded latest-review sections such as `Nitpick comments (N)` when CodeRabbit puts them in the review body instead of issue comments
- Parsed latest head-review threads, with unresolved threads clearly separated
- Latest head commit review metadata and review threads
- Unresolved latest-commit review threads after reply-thread folding
- Pre-merge failed checks, if present
- Latest MegaLinter status and any detailed issues posted by `github-actions[bot]`
- Test summary, including failed-test signals when present
- Detailed failed-test rows from GitHub Test Reporter / CTRF comments when the PR comment includes `Name` / `Failure Message` content
- CLI support for writing full JSON to a file and printing only narrowed text sections to stdout
- Parse warnings only when both the primary API source and the intended fallback signal are unavailable

## Recovery Rules

- If the current branch has no matching public PR, report that clearly instead of guessing.
- If GitHub access fails because of proxy configuration, rerun the fetch with proxy variables removed.
- If the current WSL session resolves `git.exe` but cannot execute it cleanly, keep using the explicit Linux worktree binding instead of retrying Windows Git.
- Prefer GitHub API results over PR HTML. The PR HTML page is now a fallback/debugging source, not the primary source of truth.
- If the summary block and the latest head review threads disagree, trust the latest unresolved head-review threads and treat older summary findings as stale until re-verified locally.
- Do not assume every AI reviewer behaves like CodeRabbit. `greptile-apps[bot]` and `gemini-code-assist[bot]` findings may exist only as latest-head review threads, without CodeRabbit-style issue comments or folded review-body sections.
- Treat GitHub Actions comments with `Success with warnings` as actionable review input when they include concrete linter diagnostics such as `MegaLinter` detailed issues; do not skip them just because the parent check is green.
- Do not assume all CodeRabbit findings live in issue comments. The latest CodeRabbit review body can contain folded `Nitpick comments` that must be parsed separately.
- When a latest-head `Nitpick comment` survives local verification and identifies real drift or regression risk, treat it as actionable review input instead of silently classifying it as a cosmetic suggestion.
- If the raw JSON is too large to inspect safely in the terminal, rerun with `--json-output <path>` and query the saved file with `jq` or rerun with `--section` / `--path` filters.

## Example Triggers

- 'fix pr review'
- 'Use FPR'
- `Use $gframework-pr-review on the current branch`
- `Check the current PR and extract CodeRabbit suggestions`
- `Check the current PR and extract Greptile suggestions`
- `Check the current PR and extract Gemini Code Assist suggestions`
- `Look for Failed Tests on the PR page`
- `先用 $gframework-pr-review 看当前分支 PR`