Three Open-Source Tools Let AI Coding Agents Check Security, Browser QA, and UI Quality
AI coding agents such as Claude Code, Cursor, and OpenAI Codex can generate full applications quickly, but their output still requires manual checking for security flaws, runtime errors, and design problems. A developer has released three free, local open-source tools to automate this verification: vibe-shield for security and supply-chain scanning, ghost-qa for Playwright-based browser testing, and unslop for UI craft and contrast checks. vibe-shield runs a 92-rule static scan to catch issues like leaked secret keys, missing webhook signature verification, and poisoned agent instruction files. ghost-qa tests a running local or staging app across multiple viewports and checks for console errors, failed API calls, accessibility gaps, and performance problems. The tools are designed to let coding agents verify their own work before each commit.
This is an AI-generated summary. ShortSingh links to the original source for the complete article.
Discussion (0)
Log in to join the discussion and vote.
Log in