Guide · Agent skill
How to use the Using Superpowers agent skill (using-superpowers)
Updated · written with AI from the project's own documentation and checked against our catalogue data
Using Superpowers (skill name: using-superpowers) is a meta-skill: a SKILL.md instruction file that tells an AI agent to check for and invoke other installed skills before it starts any task, including before it asks clarifying questions. This guide explains what the skill changes in agent behaviour, how to install it, how to trigger it, and what it needs to work. By the end you will know whether it fits your setup and how to test it.
Using SuperpowersjnMetaCode/superpowers-zh · MIT Download the skill Quick start
- 1Download the ZIP. Get the using-superpowers ZIP file from this page.
- 2Unzip the skill. Extract the folder so that SKILL.md and the references folder are present.
- 3Place it in your skills folder. Move the folder into ~/.claude/skills/ for personal use, or into .claude/skills/ inside a project.
- 4Or upload it in the Claude apps. Open Settings → Capabilities → Skills and upload the skill there instead.
- 5Install related skills. Make sure the other skills you want routed, such as brainstorming and systematic-debugging, are also installed.
- 6Test with a typical request. Ask the agent to build or fix something and check whether it names the skill it is using first.
What does the using-superpowers skill make the agent do?
The skill sets a rule for the start of a conversation. Before any response or action, including clarifying questions, exploring a codebase, or looking at files, the agent should call any relevant or requested skill. If a skill turns out not to fit the situation, the agent does not have to use it.
It also sets an order when several skills apply. Process skills come first, such as brainstorming or systematic debugging, and implementation skills such as frontend design come after them. For example, for "let's build X" the agent uses brainstorming before an implementation skill, and for "fix this bug" it uses systematic-debugging before a domain skill.
The file includes a table of common rationalizations, such as "this is just a simple question" or "I remember this skill", with the reply that the agent should check the skill anyway.
How do you install the using-superpowers skill?
Download the ZIP file from this page and unzip it. For a personal setup, place the unzipped folder in ~/.claude/skills/. For a single project, place it in that project's .claude/skills/ folder.
If you use the Claude apps, you can instead upload the skill under Settings → Capabilities → Skills.
- Personal install: ~/.claude/skills/
- Project install: .claude/skills/ inside your project
- Claude apps: Settings → Capabilities → Skills, then upload
How do you trigger the skill with natural requests?
You do not need a special command. The skill's description tells the agent to use it at the start of a conversation, so it is meant to run as a rule in the background. You can make this clearer by asking for work in plain language, for example asking the agent to build or fix something, and then watching whether it names the skill it is using before it starts.
The skill also says that when the agent starts a task it should announce which skill it is using and follow that skill's checklist, creating a to-do for each item if one exists.
- "Let's build a small booking form for our site." (expect brainstorming first)
- "Fix this bug where the login redirect loops." (expect systematic-debugging first)
- "Before you answer, check which skills apply to this request."
What does the skill need to run?
The skill needs an agent runtime that supports SKILL.md skill files, and it needs other installed skills for it to route to. On its own it only governs how other skills are chosen.
The listed files include reference notes for specific tools: Claude Code, Codex, Pi, Antigravity, Copilot CLI, Hermes Agent, Qoder, Muse, and a Gemini CLI mapping that loads through GEMINI.md. Read the note that matches your runtime if it is listed.
- An agent runtime that supports SKILL.md files
- Other installed skills to route to
- Optional: the platform reference file that matches your agent
Safety notes: scripts and network access
This skill runs no scripts and uses no network access. It is text instructions only, and no scripts are included in the listed files. The skill itself does not fetch data or run code.
The skill can still change behaviour. Because it asks the agent to check skills before almost everything, it can cause more skill invocations than you need, which adds delay. The file also says that the four chinese-* skills load only when the user explicitly asks for them, and that user instructions in files such as CLAUDE.md or AGENTS.md take priority over the skill.
- Runs scripts: no
- Uses network: no
- Chinese-language skills (chinese-code-review, chinese-git-workflow, chinese-documentation, chinese-commit-conventions) load only when you name them
Tips for getting the most out of it
Install it alongside the skills you already use. On its own it has nothing to route to, so its value depends on the skill collection around it.
If you find the agent invoking skills too often and slowing down simple questions, you can tell the agent to skip the workflow for a specific request, since the file says user instructions override skills. Test with a few typical tasks before you roll it out to a team.
Teams using superpowers-zh or similar collections can use it as a shared starting rule. Set it up in a new agent environment as a default workflow.
- Test on a few typical tasks before sharing with a team
- Keep your own CLAUDE.md or AGENTS.md instructions in place, since they take priority
- Check the platform reference file for your runtime
What people use it for
- Routing a request to brainstorming before an implementation skill when a developer asks for a new feature.
- Sending bug reports to systematic-debugging before any domain-specific fix.
- Giving a team a shared starting rule across a superpowers-zh skill collection.
- Setting a default workflow in a fresh agent environment so skills are checked before work begins.
- Adding a Claude Code or Codex reference note so the agent maps tool names correctly.
Frequently asked questions
Is the using-superpowers skill free?
The skill is released under the MIT licence, as stated in its SKILL.md file. The source does not mention any price, so check the licence text for terms.
What do I need before installing using-superpowers?
You need an agent runtime that supports SKILL.md skill files and other installed skills for it to route to. Without other skills, the file has little to govern.
Where do I put the using-superpowers skill folder?
Place it in ~/.claude/skills/ for a personal install, or in a project's .claude/skills/ folder. In the Claude apps you can upload it under Settings → Capabilities → Skills.
Does using-superpowers run code or fetch data?
No. It runs no scripts and uses no network access. It is text instructions only, and no scripts are included in the listed files.
Can I use using-superpowers without other skills?
It is not recommended. The file only governs how other skills are chosen, so on its own it is less useful and can make the agent invoke skills more often than needed.
Why isn't my agent using the skill?
The source does not describe specific troubleshooting steps. Check that the folder is in the right location, that your runtime supports SKILL.md files, and that the platform reference file for your agent is present if one is listed.
New agent skills, repos and AI tools — every week
The best new skills and open-source projects, the fastest-growing repos, and one free prompt. Five minutes to read. Unsubscribe anytime.
Our verdict
Use using-superpowers if you already have a set of agent skills and want your agent to check and apply them consistently, or if you are setting a shared default workflow for a team. Pick something else if you want one specific capability, or if you expect the skill to run code or fetch data, since it does neither and adds delay when used alone.
Download the skill