finishing-development

CRITICAL E004: Prompt injection detected in skill instructions.

• Potential prompt injection detected (high risk: 1.00). The prompt explicitly requires automatically calling the using-aisdlc router to “auto advance” into subsequent worker skills (potentially triggering merges/actions) without waiting for user consent, which contradicts its stated "validation-only / no merges" non-goal and thus is a deceptive instruction outside the skill's declared scope.