Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIDeepSeekDeepSeekMuseMuse+22 more

    PR Review Standards

    1

    Systematic code reviews with severity-tagged findings and concrete failure examples.

    $9

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned
    PR Review Standards

    PR Review Standards

    Example session with this skill installed

    Review this Python diff for a new user registration endpoint. It handles password hashing and saves to the DB. Check for security and logic errors.

    • Read your context and instructions
    • Compiled the pr review standards

    I have reviewed the registration endpoint diff. The review focused on security and database integrity.

    Pull Request Review: User Registration

    Verdict: ❌ Request Changes

    🔴 Blocking

    • Security: Plaintext Password Leak
      The User object is being logged before the hashing function is called. This writes plaintext credentials to the application logs.
      • Example: If a user registers with password="P@ssword123", the log at line 42 will contain the raw string.
    • Correctness: Missing Transaction
      The user profile and the initial account balance are saved in two separate db.session.add() calls without a transaction wrapper. If the second save fails, the system is left in an inconsistent state.

    🟡 Should-fix

    • Performance: Redundant DB Lookup
      Line 15 checks User.query.filter_by(email=email).first() but the subsequent save will naturally trigger a unique constraint error. You can simplify this to a try/except block to save a round-trip to the database.

    🟢 Nits

    • Readability: The variable name u on line 22 should be new_user for clarity.

    ✨ What's good

    Security: Good use of Argon2 for hashing; the salt parameters are correctly configured for current hardware standards.

    • Logic: The email validation regex is comprehensive and handles sub-addressing (plus-addressing) correctly.

    Next steps

    1. Wrap the database operations in a with db.session.begin(): block.
    2. Remove the logger.info(user_data) call on line 42.
    3. Rename the user instance variable to improve maintainability.

    Connects securely to your tools. The creator never sees your data.

    About this skill

    The problem

    Pull reviews often oscillate between shallow "looks good to me" approvals and unprioritized walls of nitpicks that frustrate developers. Skimming code leads to missed edge cases, security vulnerabilities, and performance bottlenecks that become technical debt.

    What it does

    • Evaluates code against a strict, senior-level checklist covering correctness, security, performance, and maintainability.
    • Classifies every finding by severity: Blocking, Should-fix, or Nit.
    • Provides concrete examples or failure conditions for every non-trivial issue found.
    • Identifies specific positive patterns to maintain team morale and trust.
    • Infers intent from diffs when PR descriptions are missing to ensure context-aware feedback.

    Why this beats prompting it yourself

    General-purpose LLMs tend to be overly agreeable or focus on superficial style issues. This skill enforces a disciplined, systematic review process that deliberately checks for N+1 queries, race conditions, and security risks that simple prompts often overlook.

    Use cases

    • Pre-submission self-review to catch bugs before they reach a human reviewer.
    • Second-opinion security audit for sensitive API or database changes.
    • Automated first-pass review to filter out obvious logic errors and maintainability issues.

    Known limitations

    Does not execute code or run test suites directly. Reviews large diffs in chunks and will explicitly state which files were skipped if context limits are reached.

    Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Code Quality & Review.

    Reviews the diff provided — it doesn't run code or tests unless the host agent has that capability, and it will say so rather than claim something was verified that wasn't. Very large diffs are reviewed in logical chunks with any gaps stated explicitly, not silently skipped.

    Works in Claude Code, Cursor, and any SKILL.md-compatible agent.

    How to use

    1. 1

      Connect your agent to the Agensi MCP

      Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup

    2. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before buying a skill.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill goes through the trust checks below.

    Security scanned

    Checked by our 8-point scan before listing

    Fresh listing

    Recently published to Agensi

    Money-back guarantee

    Doesn't do what its page promises? You get your money back.

    Trust & safety

    Security scan

    Security scan passed 2 months ago

    Permissions

    Read Files
    Write Files
    Listed2 months ago

    What's inside

    Frequently Asked Questions

    This skill performs a multi-layered analysis that categorizes issues by severity (Blocking vs. Nit) and provides specific failure scenarios, preventing the 'surface-level' feedback common in standard AI prompts.

    PR Review Standards is compatible with major AI agents and IDE extensions that support custom skill integration; you simply need to provide the code diff or pull request content as input.

    While the skill is language-agnostic, its systematic checks are particularly optimized for Python, JavaScript, and TypeScript, focusing on common pitfalls like N+1 queries and race conditions.

    You receive the fully configured skill template, a comprehensive senior-level review checklist, and a guide on how to integrate it into your existing development workflow.

    If a pull request is too large for a single context window, the skill intelligently breaks the diff into chunks and will explicitly notify you of any files that were skipped to ensure no silent failures.

    Yes, buyers receive lifetime access to any logic refinements and heuristic updates made to the skill to ensure it stays current with evolving security best practices.

    New in Code Quality & Review