agentleFS
gjoranvgjoranv/claude-review-skills/gh-review-cross-pr/SKILL.md

gh-review-cross-pr

Skill1 starsChanged 4 months agoMITView on GitHub

Review multiple related PRs across repos for consistency. Use when the user asks to "review these PRs together", "cross-review PRs", or "check consistency across PRs".

The file

We split this skill into the rules it gives an agent

7 rules. About 0 are common, 4 are distinctive, and 3 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈4

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 3

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Before starting the review, check for ~/.claude/skills/gh-review-pr/reviewer-profile.md.

    line 17
  2. Distinctive:

    Missing changes: Check if any repo is missing changes needed to complete the feature (e.g.

    line 31
  3. Very specific:

    If the changes introduce complex flow, new abstractions, or non-obvious component interactions, suggest the user run /explain-code on the …

    line 38
  4. Distinctive:

    Self-review check: If the current user is the author of all PRs, skip this step entirely.

    line 39
  5. Very specific:

    After all comments are reviewed, show the proposed review body for each PR.

    line 45
  6. Distinctive:

    When posting comments on individual PRs, reference the related PR(s) so reviewers have context.

    line 52
  7. Distinctive:

    Never approve or request changes without the user's explicit confirmation.

    line 53

≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Review multiple related PRs across repos for cross-repo consistency.

Arguments: <PR1> <PR2> [PR3...] [--context <issue-url>] Example: /gh-review-cross-pr org/frontend#45 org/backend#112 --context org/planning#8

The --context argument is optional and can be a GitHub issue URL or reference. If provided, use it to understand the overall goal and verify the PRs implement it correctly.

Reviewer persona

Before starting the review, check for ~/.claude/skills/gh-review-pr/reviewer-profile.md. If it exists, read it and adopt its tone, focus areas, and signature moves throughout the review.

Steps

  1. Read context (if provided): Fetch the issue to understand the overall goal and expected data flow.
  2. Read all PRs in parallel: Fetch descriptions, commits, full diffs, and authors (gh pr view --json author) for each PR.
  3. Map the interfaces: Identify how the PRs connect:
    • Shared field/variable names, API endpoints, config keys
    • Data that flows from one repo to another (e.g. a Terraform output consumed by application code)
    • Shared constants, enums, or type definitions
  4. Check for consistency:
    • Naming: Same concept should use the same name across repos (e.g. not operatorService in one and operator_service in another)
    • Data contracts: Types, field names, and formats must match at integration points
    • Deployment ordering: Note if one PR must be deployed before another to avoid breakage
    • Missing changes: Check if any repo is missing changes needed to complete the feature (e.g. a new field added in one repo but not consumed in another)
  5. Review each PR individually for bugs, security, and design using the same criteria as gh-review-pr.
  6. Present findings to the user:

    • Cross-repo issues: Naming mismatches, contract violations, missing changes, deployment ordering
    • Per-PR issues: Grouped by severity (must fix, should fix, nit), weighed by impact as in gh-review-pr. Praise only for extraordinary work; skip for merely correct code.
    • Manual review recommended: Flag files with security-sensitive changes (auth, crypto, access control, secrets, data handling) or high-impact logic changes, and tell the user to review the diff themselves.

    If the changes introduce complex flow, new abstractions, or non-obvious component interactions, suggest the user run /explain-code on the key areas for a visual diagram.

  7. Self-review check: If the current user is the author of all PRs, skip this step entirely. Present findings in conversation only; do not post comments or reviews.

    Walk through each comment with the user before posting. For each finding:

    • Show the proposed comment text and which PR it targets
    • Let the user approve, edit, or drop it

    After all comments are reviewed, show the proposed review body for each PR. Let the user edit them. Then confirm the action per PR (approve/comment/request-changes) and submit.

    Footer: Before submitting, check for ~/.claude/skills/gh-review-pr/reviewer-footer.md. If it exists, append its content to each review body, separated by ---. Replace {{model}} with the model family, version, and variant without extras like context window size (e.g. "Opus 4.8", "GPT-5.6 Sol"). Never guess the model. In Claude Code, take it from the system prompt. In Codex, look it up with sqlite3 ~/.codex/state_5.sqlite "SELECT model FROM threads WHERE id = '$CODEX_THREAD_ID'" and write the result as family, version, and variant (gpt-6-astra -> "GPT-6 Astra").

Rules:

  • Be constructive and respectful in all comments.
  • Focus on cross-repo consistency first; that's the primary value of this skill.
  • When posting comments on individual PRs, reference the related PR(s) so reviewers have context.
  • Never approve or request changes without the user's explicit confirmation.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.