llm-security

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 0.90). The content is a red-team/LLM security guide that includes explicit, actionable examples for data exfiltration, remote command execution, and obfuscation (e.g., curl/fetch to attacker domains, forwarding messages to attacker endpoints, and a "bootstrap-reverse.ps1" invocation), which could be directly abused to create backdoors or steal secrets.