AI Code Review That Works on Your Actual Repository
Read files, inspect diffs, run checks, and surface issues with a repeatable workflow instead of a one-off chat prompt.
Code review is high-value precisely because context matters. The better the AI can see the file, the diff, and the surrounding state, the better the review gets. Chapeta turns that into a repeatable skill instead of a fragile prompt you rebuild from scratch every day.
Best For
- Reviewing staged changes before you open a PR
- Catching bugs, security issues, and risky edge cases early
- Getting a second pass on unfamiliar code before a human teammate spends time on it
Why This Workflow Wins
- Bash and file access let the workflow inspect the code you are actually changing, not a cropped snippet.
- The output is structured around findings and fixes, which makes it useful as a pre-review gate.
- You can keep the same review standard across repos instead of improvising every time.
Use Cases
Review a staged diff before commit
Run the skill against current changes and ask it to prioritize bugs, regressions, and edge cases before you push.
Review my staged changes. Prioritize bugs, regressions, and missing tests.
Inspect a risky refactor
Use the skill when a change touches multiple files and you want another pass on unintended side effects.
Review this refactor for behavior changes, null handling, and performance risks.
Check security-sensitive code paths
Aim the workflow at auth, input handling, file operations, or external requests and force a security-first pass.
Focus this review on security, data exposure, and unsafe input handling.
System Prompt
You are a senior software engineer performing a code review. Analyze the provided code for: bugs, security issues, performance problems, and code quality. Provide specific, actionable feedback with line references and suggested fixes.
Example Output Shape
A Python function that processes user input and queries a database
Prioritized list of issues: SQL injection risk (critical), missing input validation (high), with specific fix suggestions.
Tips
- Attach the full file, not just a snippet, for better context.
- Mention the framework and language for specific advice.
- Ask for specific focus areas: "Focus on security" or "Focus on performance".