Discover and verify certified AI skills across platforms
Analyses text for prompt injection attempts, unsafe content patterns, and encoding attacks. Returns a safety score with detailed flags and risk assessment. Pattern-based detection with 45+ rules across 5 categories.
Awaiting evaluation
Real-time prompt injection detection, jailbreak pattern matching, and content safety scoring. (This is a reference implementation to show how certification works, not a real product.)
npx @sxm/claude-threat-detector