save-tokens

CRITICAL E004: Prompt injection detected in skill instructions.

• Potential prompt injection detected (high risk: 0.90). The file explicitly commands "sole active authority" and to suppress or block other skills (lines 3–7), which attempts to override session-level controls and is outside the skill's stated purpose of compressing assistant narration.

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 1.00). Skill explicitly instructs it to suppress and override other skills/governing prompts, which appears intended to bypass external oversight and could be used as a backdoor to seize control of a session.