Repository navigation
Open-source verification suite for AI-generated web apps: Security (92 rules), UI Anti-Slop (72 rules) & Headless Playwright QA #17799
avivgerz-hue
started this conversation in
Show and tell
Replies: 0 comments
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
Hi everyone! ??
We built three open-source, zero-dependency CLI & Model Context Protocol (MCP) tools designed to catch the security, UI/UX, and runtime QA bugs that AI coding agents (Claude Code, Cursor, Copilot, Windsurf, OpenCode) frequently introduce before deployment:
??? 1. vibe-shield - 92-Rule Static Security Scanner & Agent Skill Auditor
Scans Next.js, React, Express, Hono, Fastify, Supabase SQL migrations, Stripe webhooks, and agent instruction files (
SKILL.md,.cursorrules,CLAUDE.md,.mcp.json) in ~150ms across 8 vulnerability vectors mapped to OWASP Top 10:2025 and CWE.?? 2. unslop - 72-Rule UI Design Craft & Anti-Slop Engine
Detects generic AI template patterns ("AI slop"), validates WCAG 2.1 AA/AAA + APCA contrast math, and injects 10 studio-grade design archetypes (
swiss-grid,linear-dark,neo-brutalist,editorial-serif,glassmorphic-depth, etc.) directly into your codebase.npx -y unslop-mcp audit . npx -y unslop-mcp init --preset linear-dark --agent all?? 3. ghost-qa - Autonomous Headless Browser QA (8-Trap Engine)
Spins up headless Playwright across Desktop (
1440x900), Tablet (768x1024), and Mobile (375x812) viewports to catch hydration mismatches, uncaught runtime exceptions, failed network requests, horizontal overflow/element collisions, accessibility violations, and dark-mode contrast failures.npx -y ghost-qa test http://localhost:3000All three tools are MIT-licensed, run 100% locally offline, and work as standalone CLIs, native MCP servers, or Agent Skills. Feedback and stars are very welcome!
All reactions