AI Skills Security Scanner
Scan an AI Skill before installing it.
Claude Skills, Cursor Skills, custom prompts — any of them can contain hidden prompt injection, role confusion, or malicious payloads. Paste the code, get a verdict in seconds.
▶ Open Scanner
💉
Prompt Injection Detection
Catches "ignore previous instructions" and 20+ jailbreak variants.
🎭
Role Confusion
Detects skills that try to make the AI assume a different identity.
📦
Hidden Payloads
Finds Base64, Unicode, and zero-width character obfuscation.