using-superpowers

CRITICAL E004: Prompt injection detected in skill instructions.

• Potential prompt injection detected (high risk: 0.90). The prompt contains explicit, coercive global rules that try to force the agent to always invoke and follow skills (and to never use certain tools), which overreach the (unstated) skill purpose and attempt to override normal agent behavior—this is a prompt-injection style instruction.

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 1.00). This content is a clear prompt-injection/backdoor: it coercively instructs an agent to always invoke and follow external "skills" (including bypassing read/audit steps), enabling unvetted code/content execution and creating a vector for supply-chain attacks or data exfiltration.