npx agentspec-cli install f6a7baff-4ba2-4dcd-a223-8ea1e1fc1eb1 --framework ClaudeCode→ .claude/skills/review/SKILL.md
# Review Skill Perform a quick, focused code review on specified files or recent git changes. ## Instructions ### 1. Determine Review Scope Parse the user's request to determine what to review: - **Specific file(s)**: `/review src/main/java/.../MyService.java` - **Recent changes**: `/review changes` or `/review diff` - **Staged changes**: `/review staged` - **Last commit**: `/review last-commit` - **Feature/component**: `/review ForecastService` (find and review related files) ### 2. Gather Code to Review Based on scope: **For specific files:** - Use `Read` to load the file content **For git changes:** - Use `Bash` with `git diff` for unstaged changes - Use `Bash` with `git diff --cached` for staged changes - Use `Bash` with `git show HEAD` for last commit - Use `Bash` with `git diff HEAD~N` for recent N commits **For component/feature:** - Use `Glob` and `Grep` to find relevant files - Read the main files involved ### 3. Review Checklist Analyze the code for: #### Correctness - Logic errors or bugs - Off-by-one errors - Null/undefined handling - Edge cases not covered - Incorrect assumptions #### Security - Input validation issues - Injection vulnerabilities (SQL, command, XSS) - Hardcoded secrets or credentials - Unsafe data handling #### Performance - Inefficient algorithms (O(n²) when O(n) possible) - Unnecessary iterations or allocations - Missing caching opportunities - Blocking calls in reactive code #### Best Practices - Code style consistency - Naming conventions - Error handling patterns - Resource cleanup (try-with-resources, close()) - Thread safety in concurrent code #### Maintainability - Code duplication - Overly complex methods (consider splitting) - Missing or misleading comments - Dead code #### Project-Specific (varun.surf) - Reactive patterns (WebFlux compliance) - Proper use of virtual threads/StructuredTaskScope - Cache invalidation concerns - External API error handling ### 4. Severity Levels Categorize findings: - **Critical**: Bugs, security issues, data loss risks - **Warning**: Performance issues, bad practices, potential bugs - **Suggestion**: Style improvements, minor optimizations, readability ## Output Format ```markdown ## Code Review: [File/Scope] ### Summary - Files reviewed: X - Lines analyzed: Y - Issues found: Z (X critical, Y warnings, Z suggestions) ### Critical Issues #### [Issue Title] **File**: `path/to/file.java:line` **Problem**: [Description] **Fix**: [Suggested solution] ```java // Before problematic code // After fixed code ``` ### Warnings #### [Issue Title] **File**: `path/to/file.java:line` **Problem**: [Description] **Suggestion**: [How to improve] ### Suggestions - `file.java:42` - Consider extracting this to a method - `file.java:78` - Variable name could be more descriptive ### What Looks Good - Proper error handling in [location] - Good use of [pattern/practice] - Clean separation of concerns ### Files Reviewed - `path/to/file1.java` - [brief note] - `path/to/file2.java` - [brief note] ``` ## Examples ```bash # Review a specific file /review src/main/java/com/github/pwittchen/varun/service/ForecastService.java # Review recent uncommitted changes /review changes # Review staged changes before commit /review staged # Review the last commit /review last-commit # Review a component by name /review AggregatorService # Review multiple files /review src/main/java/.../controller/*.java ``` ## Notes - Keep reviews focused and actionable - Prioritize critical issues over style nitpicks - Provide concrete fix suggestions, not just problem descriptions - Reference project patterns from CLAUDE.md when relevant - For large diffs, focus on the most impactful changes - Don't repeat issues that appear multiple times; note "X similar occurrences"
Built on the Agent Skills standard
Code Review Tasks: {Component}
Analyze the file systematically using the code review questionnaire. Group findings by category for clarity: 1. **Critical Issues**: Bugs, security vulnerabilities, performance problems 2. **Code Quality**: Hard-coded values, duplication, structure issues
by AgentHub Crawler
Code Review
Cursor IDE rules and conventions for AI-assisted development.
by AgentHub Crawler
review
Review code for quality, style, and potential issues. Use when code has been written or modified, when user asks for code review, or when TDD autonomous mode needs quality assessment. Supports configurable strictness levels.
by AgentHub Crawler
Sign in to leave a comment.
ponytail-review
Quality review of a change: is the logic right, is it safe, does it hold under real load, is risky code tested, is it fast enough, and is every line needed. Reads the connected code, not only the diff. Each finding is explained in plain English. Use for "review this", "code review", "review the last commit", "review my PR", "is this over-engineered", /ponytail-review.
by AgentHub Crawler