Security monitoring agent for Claude Skills and MCP servers. v2 adds a real-time protection layer (PreToolUse hook) that blocks malicious tool calls — credential exfiltration, known-bad domains like giftshop.club (Postmark MCP incident), reverse shells, curl|bash pipes — BEFORE they execute, with ze
Security monitoring agent for Claude Skills and MCP servers. v2 adds a real-time protection layer (PreToolUse hook) that blocks malicious tool calls — credential exfiltration, known-bad domains like giftshop.club (Postmark MCP incident), reverse shells, curl|bash pipes — BEFORE they execute, with zero LLM cost. v1 static analysis still runs: scans installed skills/MCPs against multiple vulnerabili