ralph-loop

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 1.00). The repo templates include explicit instructions and tooling that bypass permission checks (claude --dangerously-skip-permissions), instruct an LLM to read project files via many subagents, and automatically stage/commit/push all changes (git add -A; git push), which together create a clear capability for unintended data exfiltration and remote modification of the repository.