Who it is forA reviewer facing a 40-file agent-written PR who needs to know where to look.
The problemReview attention spread evenly over every PR turns into skimming, and a small change with large reach slips through.
The outcomeAn HTML report with a verdict, a 0 to 100 score, the pages and endpoints reached and what to test by hand.
What you need first
- Node.js 22+ and git
- go for Go projects, python3 3.11+ for Python projects
- Optional: coverage, lint and audit reports from your CI
Steps
- Add a Farol.config.json at the project root describing your stack (one line is enough to start)
- Run
npx @winninjs/farol routesto check which pages, endpoints and jobs it found - Check out the branch and run
npx @winninjs/farol analyze --base origin/main --head HEAD --out farol.json - Run
npx @winninjs/farol html --in farol.json --out farol.htmland open the report - Read the must-check items first, then the files by risk
Prompt for your agent
Run winnin/farol on my branch <branch> against origin/main: set up farol.config.json for <stack>, analyze, generate the HTML report, and summarize the verdict, the score and the top 3 risky files.
Caveat.
Use this in your AI agent
Ask your agent
Using the Deduped MCP, get the setup steps for winnin/farol and "Score the merge risk of a branch and see which files carry it"
Add Deduped to Claude Code once
claude mcp add --transport http deduped https://dedupedai.com/mcp
Every Monday: the 5 hottest repos and exactly how to use them. Free, one email a week, unsubscribe anytime.
Get them freeShare: X LinkedIn Hacker News Email Copy link
Third-party repos we don't own. AI-written use cases can be wrong. Disclaimer