using-superpowers

To ensure an AI agent checks for relevant skills before responding, enforce a mandatory skill invocation rule at the start of conversations and during planning workflows. This forces the agent to query a platform skill tool for applicability, reducing ad-hoc behavior and ensuring consistent, safe agent workflows.

The best way to map skill invocation across Claude, Copilot, and Gemini is to use a cross-platform tool-mapping reference. This documents equivalent tool names and invocation methods for each environment, ensuring agent workflows remain compatible and skills load correctly regardless of the underlying platform.

Yes, a mandatory skill check integrates directly into debugging and implementation workflows. It applies during planning phases to load relevant process skills like TDD, brainstorming, or subagent dispatching, ensuring the agent follows structured checklists exactly rather than generating ad-hoc or unsafe solutions.

Yes, explicit user instructions override skill directives during agent conversations. The workflow priority rules establish that user instructions take precedence over superpowers skills and default system prompts, allowing you to maintain control over the agent's behavior while still benefiting from automated skill discovery checks.

Your AI agent overlooks applicable skills during bugfix planning because it lacks an enforced invocation check at the start of its workflow. Without a mandatory rule to search for specialized skills before acting, the agent defaults to its base system prompt, leading to ad-hoc behavior and missed process checklists.