using-superpowers

To ensure relevant skills are invoked before a response, establish a mandatory protocol that discovers and applies applicable skills at the start of every conversation. This enforces priority: user instructions, then superpowers skills, then default system behavior.

The correct priority order for applying conversation skills enforces user instructions first, followed by superpowers skills, and finally falls back to default system behavior. This ensures specific user guidance always overrides automated tool selection.

Yes, you can use skill invocation protocols with Copilot and Gemini tools. The system supports integration with these platforms via platform-appropriate mapping, facilitating consistent and safe responses across different AI environments.

To structure a SKILL.md file for discovery, include a frontmatter with the skill name and description. You can also optionally add references, scripts, and assets to support the skill's execution and provide deeper context.

Safety guards to prevent prompt injection and data exfiltration are required while loading skill content. These protections ensure that external content or malicious references cannot manipulate the AI's context or compromise the conversation.

Yes, you need to invoke skills for brainstorming and clarification prompts. The protocol defines applicability across common prompt types, including task execution, and mandates tool usage guidance be applied before any response.