The Agent Skills Directory

[SAFE]: The skill provides operational instructions and flowcharts to guide agent behavior when encountering tasks that may require specialized skills. It uses imperative language to enforce a 'search-first' mindset for skills, which is a design pattern for the 'superpowers' skill ecosystem.
[SAFE]: The documentation includes environment detection logic using standard Git commands (git rev-parse, git branch) to determine if the agent is operating within a specific workspace or branch. These commands are benign and used purely for context awareness.
[SAFE]: Tool mapping references for Codex, Copilot, and Gemini platforms translate Claude Code-specific tool names into their equivalents, facilitating consistent functionality across different AI agent environments without introducing any unauthorized access or command execution vectors.
[SAFE]: The skill explicitly defines a priority hierarchy that places user-specific instructions (e.g., in CLAUDE.md or direct requests) above the skill's own rules, ensuring the user remains in control of the agent's behavior at all times.

using-superpowers