dashboard

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 1.00). The server intentionally delegates generation to a Claude agent with allowedTools (Bash/Write/Read), feeds it local vector-store content, and spawns child processes inheriting full environment — this design grants any invoked model/process the ability to execute arbitrary shell commands and access/export local files and environment variables, enabling remote code execution and data exfiltration.