using-superpowers

The using-superpowers skill acts as the foundational governance layer for software development agents utilizing the Superpowers methodology. It is specifically designed to enforce a disciplined, systematic approach to coding by preventing agents from jumping prematurely into implementation. By establishing a mandatory protocol for skill invocation, it ensures that every interaction begins with an assessment of available capabilities, promoting the use of specialized skills like brainstorming, test-driven development, and subagent-driven workflows. This skill is intended for software engineers and development teams who demand high-integrity agentic behavior and reproducible coding processes. It effectively eliminates the tendency for agents to rationalize skipping established workflows or ignoring project context, ensuring that every user prompt is met with a rigorous, methodology-aligned response.

• Automatically triggers as a primary check at the start of every conversation, ensuring agent alignment with the Superpowers framework.

• Mandates the use of the Skill tool for discovery, preventing the agent from bypassing predefined methodologies.

• Establishes a strict priority hierarchy, ensuring user instructions always supersede system defaults while maintaining workflow discipline.

• Prevents common pitfalls like ad-hoc coding, undocumented experimentation, and inconsistent testing practices.

• Facilitates the integration of secondary skills including git-worktree management, systematic debugging, and parallel agent dispatching.

• This skill should be invoked before any clarifying questions or task execution to maintain context and methodological rigor.

• It treats every user request as a potential task, ensuring that even simple queries are evaluated for applicable skill sets.

• Agents are instructed to follow the 1% rule: if there is even a minimal chance a skill applies, it must be invoked.

• Operates as a meta-skill, providing the necessary operational discipline for Claude Code, Copilot CLI, and Gemini CLI environments.

• Does not replace user intent; instead, it provides the structured path required to execute user-defined goals with maximum engineering quality and reliability.