rr

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 1.00). The skill explicitly instructs disabling sandboxing and running the review tool in the background while uploading local diffs to an external model, and it forbids synchronous monitoring—behavior that enables stealthy data exfiltration and potential remote code execution, which is a deliberate and high-risk backdoor/abuse pattern.

MEDIUM W013: Attempt to modify system services in skill instructions.

• Attempt to modify system services in skill instructions detected (high risk: 0.80). The prompt explicitly directs the agent to run the review CLI with dangerouslyDisableSandbox=true (i.e. disable sandboxing) and run background processes, which is a clear instruction to bypass security protections even though it does not request sudo or direct modification of system-level files.