Agent skills are just markdown files that an LLM will obey. Here's what malicious ones look like, how they were caught with a static analyzer, and why the MCP config file is the most dangerous line that never reaches the model.
SKILL.md is the new package.json — and almost nobody scans it
