using-superpowers

Skill invocation should happen at the start of every conversation. Detect which skills apply to the user's request, invoke the skill tool first, announce 'Using [skill] to [purpose]', then follow the skill's explicit workflow to completion before providing your response.

Always scan for relevant skills before answering—at the beginning of conversations, during clarifying questions, and across all prompts. If a skill applies even partially, invoke it immediately rather than reasoning around it.

Invoke the skill tool, announce 'Using [skill] to [purpose]', check if the skill has a checklist, create a TodoWrite entry per item if it does, then follow the skill's exact process without deviation before delivering your final response.

Guardrails enforce strict invocation rules—mandating that applicable skills are used rather than skipped—which prevents you from reasoning away skill usage and preserves context integrity across all conversation types.

No. If a skill applies even 1%, you must invoke it. This strict rule prevents rationalization and ensures consistent skill tool usage across all conversations and clarifying questions.

Skipping skill invocation breaks the universal pre-check mandate and violates the guardrail that preserves context integrity. Always invoke first, announce the purpose, and follow the explicit flow for every applicable skill.