Using Superpowers Skill

Using Superpowers is the bootstrap skill of Superpowers (skill source). It is loaded at the start of a session, and on some harnesses again after context compaction, and it establishes one rule: before any response or action, including a clarifying question, the agent checks whether a skill applies and invokes it.

Without this skill, the rest of the library is just documentation the agent might skip. With it, skills such as Brainstorming and Systematic Debugging fire when they should.

Key Features

• The 1% rule: if there is even a 1% chance a skill applies, the agent must invoke it; if it turns out wrong for the situation, it can drop it.
• Announce and follow: the agent says "Using [skill] to [purpose]" and turns any checklist into todos.
• Priority: process skills come before implementation skills. "Let's build X" starts with brainstorming; "fix this bug" starts with systematic debugging.
• Red-flag table: rebuttals to "this is just a simple question", "let me explore the codebase first" and "I remember this skill".
• User instructions win: CLAUDE.md, AGENTS.md, GEMINI.md and direct requests override skills.
• Subagent stop: a subagent dispatched for a specific task ignores this skill, so helpers do not re-run the whole process.
• Platform notes: reference files for Claude Code, Codex, Pi, Antigravity, Hermes Agent and Muse.

Use Cases

• Every Superpowers session; it is what makes the skills automatic.
• A model for anyone writing their own skill bootstrap.

Pricing

Free and open source under the MIT license. Prime Radiant, the company behind Superpowers, sells commercial support to enterprises. The practical cost is tokens: process skills add questions, reviews and subagent runs.

Getting Started

Superpowers installs as one plugin, so this skill arrives with the rest of the library (15 skills in v6.4.1). In Claude Code run /plugin install superpowers@claude-plugins-official. Cursor uses /add-plugin superpowers, Gemini CLI uses gemini extensions install https://github.com/obra/superpowers, and the README lists commands for Codex, GitHub Copilot CLI, OpenCode, Pi and other harnesses. Skills trigger on their own; you can also ask for one by name. The plugin injects this skill automatically; you do not call it yourself.

Limitation: the rule is deliberately aggressive, so agents sometimes invoke a skill for a trivial question and spend tokens reading it. If a session misbehaves, the README suggests asking the agent to run the diagnosing-superpowers skill on that session.

FAQ

Can I turn a skill off for one task?

Yes. Tell the agent explicitly; user instructions take precedence.

Does it work with non-Superpowers skills?

Yes. The rule applies to any installed skill.

Alternatives

• Claude Skills: Claude Code's built-in skill loading, without the forcing rule.
• OpenSkills: a universal skill loader across agents.
• Agent Skills: another lifecycle-enforcing skill pack.

Conclusion

The small skill that turns a folder of good advice into a workflow the agent actually follows. More in the skills hub.