# skillaudit.sh

> Security audit for LLM skill files (SKILL.md, .cursorrules, AGENTS.md) in GitHub repositories. Paste a repo URL to get an instant report on prompt injection, data exfiltration, privilege escalation, and supply-chain risks.

skillaudit.sh scans skill files for Cursor, Windsurf, Claude CLI, and other AI-native IDEs. Supported by CodeThreat.

## Core

- [Home](https://skillaudit.sh/): Paste a GitHub repo to audit
- [Security Checks](https://skillaudit.sh/checks): List of rules we scan for
- [Glossary](https://skillaudit.sh/glossary): Security terms explained

## Optional

- [Prompt Injection](https://skillaudit.sh/glossary/prompt-injection): What it is and how we detect it
- [Data Exfiltration](https://skillaudit.sh/glossary/data-exfiltration): Outbound data risks
- [Privilege Escalation](https://skillaudit.sh/glossary/privilege-escalation): Command and system access risks
- [Supply Chain Attack](https://skillaudit.sh/glossary/supply-chain-attack): npm/PyPI package verification
- [Obfuscation](https://skillaudit.sh/glossary/obfuscation): Hidden or encoded malicious content
- [Skill File](https://skillaudit.sh/glossary/skill-file): SKILL.md, .cursorrules, AGENTS.md explained