writing-plans

CRITICAL E004: Prompt injection detected in skill instructions.

• Potential prompt injection detected (high risk: 1.00). The required plan header embeds an out-of-scope instruction directed at the model ("For Claude: REQUIRED SUB-SKILL: Use superpowers:executing-plans to implement this plan task-by-task"), which attempts to override the skill's stated purpose (produce implementation plans) by forcing execution behavior, so this is a prompt-injection-style directive hidden inside required output content.