- Home
- Skills
- Code Quality & Review
- ReviewPilot — Pre-PR AI Code Reviewer & Auto-Fixer
ReviewPilot — Pre-PR AI Code Reviewer & Auto-Fixer
Catch security bugs, leaked secrets, edge cases, and performance regressions in your git diff—and auto-fix them inside Cursor, Claude Code, and Antigravity before opening a PR.
$3
Works with the AI Agents you already use
ReviewPilot — Pre-PR AI Code Reviewer & Auto-Fixer
Example session with this skill installed
Run a pre-PR check on my staged changes and look for security or performance issues.
- Read your context and instructions
- Compiled the reviewpilot
I have analyzed the staged changes in your repository. I found 1 critical security risk and 2 performance warnings.
ReviewPilot Scorecard
| Severity | Category | File | Finding | Recommendation |
|---|---|---|---|---|
| 🚨 Critical | Security | src/auth/service.ts:42 | Hardcoded JWT secret found in environment fallback. | Move secret to environment variables. |
| ⚠️ Warning | Performance | src/db/repo.ts:118 | N+1 query detected in getUserPosts loop. | Use a single JOIN or WHERE IN query. |
| ⚠️ Warning | Logic | src/utils/parser.ts:12 | Unhandled null from JSON.parse result. | Add a null check before accessing properties. |
| 💡 Suggestion | Clean Code | src/api/routes.ts:88 | Leftover console.log(data) debug statement. | Remove debug logging. |
Proposed Fix for src/auth/service.ts
// Current
const secret = process.env.JWT_SECRET || "dev-secret-123";
// Fixed
const secret = process.env.JWT_SECRET;
if (!secret) {
throw new Error("JWT_SECRET environment variable is not defined");
}
Next steps
- Would you like me to apply the JWT secret fix and remove the console logs?
- Should I refactor the
getUserPostsmethod to resolve the N+1 query? - Once fixed, I can generate your GitHub PR title and description.
Connects securely to your tools. The creator never sees your data.
About this skill
The Problem
Every developer using AI coding agents knows the embarrassment of pushing a PR to GitHub, only to have CodeRabbit or their Tech Lead comment 15 times on unhandled nulls, leaked environment variables, missing permissions, and N+1 database queries.
Cloud code reviewers are great, but they have major flaws:
- ❌ They catch bugs too late: After you’ve already pushed code and opened a public PR.
- ❌ They don't fix the code: They leave comments, forcing you to manually rewrite lines.
- ❌ Expensive subscriptions: Charging $15–$24/user/month for cloud services reading private code.
The Solution: ReviewPilot (Your Local, Private Pre-PR Gate)
ReviewPilot runs 100% locally inside your favorite coding agent (Cursor, Claude Code, Antigravity). It inspects your uncommitted changes (git diff) before you push, and performs an adversarial 4-pillar review:
- 🛡️ Security & Secret Leak Detection:
- Detects hardcoded API keys, tokens, and credentials.
- Flags SQL/NoSQL injection, missing authorization guards, and multi-tenant data leaks.
- Audits unsanitized user inputs and dangerous deserialization.
- ⚡ Logic & Edge Cases:
- Detects floating promises, missing error boundaries, and unhandled null/undefined.
- Identifies async race conditions and boundary condition errors.
- 🚀 Performance & Database Safety:
- Flags N+1 queries in ORMs (Prisma, TypeORM, Drizzle, SQLAlchemy).
- Identifies unbounded array operations, missing database indexes, and memory leaks.
- 🔧 One-Click Interactive Auto-Fixing (The Game Changer):
- Unlike cloud bots that only complain, ReviewPilot directly patches your files with approved fixes in a single step!
- 📝 Executive PR Description Generator:
- Generates a ready-to-paste PR title, bulleted feature summary, and test plan for GitHub/GitLab.
What’s Included in this Package (.zip)
- Universal Agent Skill (Antigravity, Claude Code, Windsurf)
- Cursor Rules Pack (.cursor/rules/review-pilot.mdc)
- Claude Code Configuration (CLAUDE.md)
- Complete Multi-Agent Setup Guide
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.
Requires local git repository. Large diffs may hit context limits. Fixes require user approval before file write.
How to use
- 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
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 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 9 days ago
- Accounts
Gmail
GitHub
Discord
- Software
Git
Permissions
Allowed Hosts