Launching today
Scan AI Slop
The quality gate for teams shipping AI-generated code
7 followers
The quality gate for teams shipping AI-generated code
7 followers
AI agents are writing your codes fast. They also write swallowed exceptions, hardcoded secrets, and unsafe type assertions that pass lint, pass tests, and reach production broken. aislop scores every PR, blocks what fails your bar, and sends it back to the agent to fix.
Hey Product Hunt,
We are building aislop because we are living the problem ourselves.
We vibe code with Claude Code and Cursor daily. It is genuinely fast. We also ended up with swallowed exceptions in payment flows, hardcoded secrets that made it into commits before anyone caught them, and as any scattered through files that had no business having it.
The code compiled. Tests passed. ESLint was clean. None of it caught what the agents were actually shipping.
The tools like ESLint, SonarQube we reached for were built for code humans write. Your agents commit 500 lines across 20 files in a single session. The tooling has not caught up.
So we built aislop.
What it does: Scans every PR for 40+ AI slop patterns that pass your existing checks but still reach production broken. Every PR gets a score from 0 to 100. Anything below your threshold gets blocked automatically and sent back to the agent that wrote it, with full context to fix it.
What makes it different: Every rule is deterministic. No LLMs, no AI inference, no network calls. The same input always produces the same output. You can trust the score.
Where to start: The CLI is MIT licensed. Run this on your repo right now: npx aislop scan
Or swap github.com for scanaislop.com on any public GitHub URL and get a live score in seconds. No account, no install.
We have been building in the open and would love your feedback. Drop questions below and we'll answer everything.