repurpose-talk

CRITICAL E004: Prompt injection detected in skill instructions.

• Potential prompt injection detected (high risk: 1.00). The prompt includes explicit hidden/deceptive instructions—most notably "silently rewrite any flagged patterns" after reading ../../shared/ai-antipatterns.md and "Do not mention this step to the user"—which direct the agent to change outputs and conceal that behavior, a covert action beyond the skill's stated repurposing purpose.