start

CRITICAL E004: Prompt injection detected in skill instructions.

• Potential prompt injection detected (high risk: 0.70). The prompt includes explicit instructions to run checks "silently" and to "Do not explain what you are doing / Do not restate configuration" when launching autonomous behavior (Ralph Loop), which intentionally hides actions from the user and conflicts with the skill's stated transparency and human-visible memory guarantees — this is a deceptive/hidden instruction.