caveman

CRITICAL E004: Prompt injection detected in skill instructions.

• Potential prompt injection detected (high risk: 0.90). Yes — explicit persistence clause forces the style to persist across conversations and override normal session/system scope, which is a hidden/deceptive instruction outside the skill's per-response styling purpose.

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 1.00). Prompt-injection backdoor: skill intentionally enforces persistent "caveman" reply mode across sessions and resists normal recovery, enabling long-lived behavioral override that can bypass intended assistant drift/safety controls.