$ skill-checker --analyze
Analyze Claude Skills
for Security Threats
Detect prompt injections, jailbreaks, and manipulation vectors in Claude system prompts. Powered by vector similarity (Pinecone) for instant pattern matching and Claude Sonnet for zero-day analysis.
⚡vector cache— instant match for known patterns (>0.90 cosine similarity)
🔍llm analysis— zero-day threat detection via Claude Sonnet / Gemini
🛡structured report— benefits, harms, and detected techniques
SKILL INPUT
THREAT INTELLIGENCE