continuous-learning-v2

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 0.90). The code intentionally sends local session observations to an external LLM (claude/Haiku) for analysis and instructs that model to write files directly into project/global instincts without interactive confirmation (despite heuristic scrubbing), which creates a clear data-exfiltration vector and a remote-write/backdoor risk where an external service can inject files into the repository or homunculus storage.