autoresearch

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 0.90). The package intentionally gives an external LLM process autonomous write/commit/execute privileges (via the claude CLI + --permission-mode bypassPermissions and by instructing the agent to git-commit, run fitness commands, and edit files), which creates a high-risk remote-code-execution and data-exfiltration vector that could be abused to install backdoors, leak secrets, or modify the supply chain.