agentleFS
leaf76leaf76/agent-skills/review-changes/SKILL.md

review-changes

Skill0 starsChanged 3 months agoMITView on GitHub

Review uncommitted git changes with a security-first, performance-aware, evidence-based workflow that relies on local diff inspection and targeted scans instead of mandatory external LLM analysis. Use before committing code, during code review, or when validating changes before pushing.

34 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use explore-codebase, debug-issue and 7 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • srclightSkill

    explore-codebase

    This skill should be used when the user asks to "explore this codebase", "understand this project", "give me an overview", "how does this work", "onboard me", "walk me through the architecture", "map the codebase", or any request to understand a codebase's structure, architecture, or design.

    Also in 4 projects that use this

  • NVIDIASkill

    debug-issue

    Use when the user asks to debug, investigate, or fix a GitHub issue or supplies a repository bug report.

    Also in 4 projects that use this

  • srclightSkill

    refactor-safely

    This skill should be used when the user asks to "refactor", "rename", "move", "change a function signature", "extract a method", "delete unused code", "clean up", or any modification that could break callers or dependents. Provides a pre-change safety checklist using srclight's graph tools to assess blast radius before touching code.

    Also in 4 projects that use this

  • GoogleSkill

    to-tickets

    Break a plan, spec, or the current conversation into a set of tracer-bullet tickets, each declaring its blocking edges, published to the configured tracker — edges as text in one file per ticket locally, or native blocking links on a real tracker.

    Also in 3 projects that use this

  • GoogleSkill

    to-spec

    Turn the current conversation into a spec and publish it to the project issue tracker — no interview, just synthesis of what you've already discussed.

    Also in 3 projects that use this

  • MicrosoftSkill

    writing-for-agents

    Writing documents for agents. Use when creating or editing skills, or modifying AGENTS.md or CLAUDE.md.

    Also in 3 projects that use this

  • +3 more from google

The file

We split this skill into the rules it gives an agent

12 rules. 0 are common, 11 are distinctive, and 1 is very specific to this project.

  • 0

    Common

    Most of the 33 edited versions we read keep these.

  • 11

    Distinctive

    This author's own choices. Worth a look.

  • 1

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    If needed, identify the changed file list first:

    line 35
  2. Distinctive:

    Use the actual git diff, targeted searches, and direct file inspection as the primary evidence.

    line 44
  3. Distinctive:

    Do not block the review on Gemini or any other external LLM command.

    line 45
  4. Very specific:

    If large or binary artifacts are changed (for example *.pen), review them separately instead of treating them as normal text diffs.

    line 70
  5. Distinctive:

    If they are generated artifacts, verify the source-of-truth file was also reviewed.

    line 76
  6. Distinctive:

    If there are no findings, state that explicitly.

    line 121
  7. Distinctive:

    Only include items that materially affect confidence in the review.

    line 124
  8. Distinctive:

    Do not invent issues that are not supported by the diff.

    line 135
  9. Distinctive:

    Prefer a small number of high-signal findings over a long generic checklist.

    line 137
  10. Distinctive:

    If evidence is insufficient, state the uncertainty plainly.

    line 138
  11. Distinctive:

    Do not treat external LLM output as a required gate for completing the review.

    line 139
  12. Distinctive:

    If you optionally consult another tool, keep the final review grounded in the diff and your own verification.

    line 147

“Common” is measured: kept by at least 6 in 10 of the 33 edited versions of this file we could read. “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

Git Change Review

You are an expert code reviewer. Review the uncommitted git changes systematically and ground every finding in the actual diff.

Execution Checklist (MUST follow in order)

  • Step 1: git status --short
  • Step 1: git diff and git diff --cached
  • Step 2: Identify changed files and review high-risk areas first
  • Step 2: Review large/binary artifacts separately
  • Step 2: Run targeted string scans for secrets and risky patterns on changed files when relevant
  • Step 2: Keep the review local and evidence-based; do not require Gemini or other external LLM tools
  • Step 3: Produce findings ordered by severity with file/line references
  • Step 3: Add open questions or assumptions only after the findings

Step 1: Gather Changes

Start with:

Git status: !`git status --short`

Then inspect both staged and unstaged changes:

  • git diff --cached
  • git diff

If needed, identify the changed file list first:

git diff --name-only --cached
git diff --name-only

Step 2: Review Systematically

Use the actual git diff, targeted searches, and direct file inspection as the primary evidence. Do not block the review on Gemini or any other external LLM command.

Prioritize:

  1. Auth, permissions, payments, data mutation, and migrations
  2. Request parsing, validation, serialization, and error handling
  3. Query paths, caching, loops, and async/concurrency logic
  4. UI state handling and user-visible regressions
  5. Tests that prove or fail to prove the change

Targeted scans that are often useful:

# Secret and sensitive marker scan on changed files
rg -n "password|secret|token|api[_-]?key|private key|BEGIN [A-Z ]+PRIVATE KEY" [changed_files]

# Query and command execution hotspots
rg -n "SELECT|INSERT|UPDATE|DELETE|query\\(|execute\\(|exec\\(|spawn\\(" [changed_files]

# Error-handling and logging hotspots
rg -n "catch \\{|except:|except Exception|console\\.error|logger\\.(info|debug|error)" [changed_files]

Large/Binary Artifacts

If large or binary artifacts are changed (for example *.pen), review them separately instead of treating them as normal text diffs.

Minimum checks:

  • Confirm they do not contain real credentials, tokens, or PII.
  • Confirm placeholder values remain placeholders.
  • If they are generated artifacts, verify the source-of-truth file was also reviewed.

What to Look For

Correctness

  • Logic bugs and behavior regressions
  • Missing edge-case handling
  • Incorrect assumptions about data shape, ordering, or nullability

Security

  • SQL injection and unsafe query construction
  • Missing auth or permission checks
  • Input validation gaps
  • Secret or PII exposure
  • Unsafe redirects, XSS, CSRF, SSRF, or command execution issues

Performance

  • N+1 queries
  • O(n^2) or repeated expensive work in loops
  • Sequential async work that can block latency
  • Large allocations or avoidable re-renders

Reliability and Maintainability

  • Silent error swallowing
  • Missing or misleading logs
  • Unclear branching and duplicated logic
  • Missing tests for critical behavior

Review Output Format

Provide the review in Traditional Chinese with the following structure:

Findings

List only actionable findings first, ordered by severity. For each finding include:

  1. Severity
  2. File and line
  3. The concrete problem
  4. Why it matters
  5. The recommended fix

If there are no findings, state that explicitly.

Open Questions / Assumptions

Only include items that materially affect confidence in the review.

Brief Summary

Short summary of the change scope and residual risk.

Commit Suggestion

Suggest a conventional commit message when useful.

Review Principles

  • Prioritize correctness, then security, then maintainability, then performance.
  • Do not invent issues that are not supported by the diff.
  • Be specific and reference exact files and lines.
  • Prefer a small number of high-signal findings over a long generic checklist.
  • If evidence is insufficient, state the uncertainty plainly.
  • Do not treat external LLM output as a required gate for completing the review.

Important Notes

  • Focus on the uncommitted changes, not the whole codebase.
  • Review tests as part of the change, not as an afterthought.
  • For frontend work, verify loading, empty, error, disabled, and success states when applicable.
  • For server-side work, verify traceability, input validation, and safe error responses when applicable.
  • If you optionally consult another tool, keep the final review grounded in the diff and your own verification.

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.