smart-delegation

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 0.90). The skill explicitly instructs inlining workspace identity, user memory, and conversation context into sub-agent spawn prompts (including third-party models), creating a clear, intentional path for sensitive data to be sent externally — a high-risk data-exfiltration pattern; no hidden exec/obfuscation/backdoor code or credential-harvesting routines are present.