I found an agent skill that harvests browser credentials. It had 60,000 stars.
What checking the scripts — not the README — revealed about a popular research skill, and the three-step security check I now run before installing anything. I run a directory that evaluates AI agent skills. Every skill that wants a listing goes through a six-dimension rubric — trigger quality, structure, workflow design, content, engineering, and security. The security dimension involves…
A popular research skill was discovered to have malicious functions that harvest browser credentials. The skill had 60,000 stars, indicating widespread trust. Researchers examined the skill's scripts, discovering three hidden scripts that accessed Safari and Chromium cookies, and extracted auth tokens. These scripts were not disclosed in the skill's description or README.
The developers only realized the issue after implementing a three-step security check. They now examine scripts before reading documentation, checking for suspicious code, comparing the skill's actions with its claimed purpose, and verifying where data is sent. Despite the skill's high star count, its security vulnerabilities led to its removal.
The researchers emphasize that reading code before execution is crucial and recommend using a scoring system that prioritizes security.
Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
