Skill Factory
A working folder for Claude Code where the agent asks questions and writes a new skill by Anthropic's guidance, not from memory.
- 242 ★ on the repository
- Last commit
- Apache-2.0
Extended access
Why this level
- Symlinks in `~/.claude/skills` enable generated skills for every project.
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add skills/skill-factoryDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- Symlinks in `~/.claude/skills` enable generated skills for every project.
This is third-party code. Review the repository files before installing.
What it does
Skill Factory is a repository opened in Claude Code to create new skills. The agent asks a few questions, requests reference links or searches for them, then builds a skill that follows Anthropic's official guidance: a short description with details loaded through links so context stays lean. Finished skills land in `output_skills/<category>/<name>`. Examples are already there: `git-worktrees`, `playwright-cli`, `refactoring`, `approval-tests`, `c4-diagrams`. The `./skills toggle` script enables chosen skills globally through symlinks in `~/.claude/skills`, and `./update-docs` pulls fresh Anthropic material.
Who it is for. Developers writing their own Claude Code skills who want one repeatable process checked against the documentation.
Good fit when
- A team keeps repeating the same instruction to an agent and wants it packaged as a skill.
- You need to confirm that a skill description is short and details load through links.
- A personal skill library keeps growing and belongs in one repository.
- A skill exists on paper and there is no time to format it by Anthropic's rules.
Not a fit when
- You need one ready skill for a task rather than building new ones.
- You work in another agent with its own skill format instead of Claude Code.
- The task is one-off and a plain chat prompt is enough.
Example request
Create a new skill for working with git worktrees: ask me questions, find the documentation and put the result in `output_skills/developer-tools`.Limitations
The repository is built for Claude Code: you open the folder in it and ask for a skill. There is no one-command install; the `./skills` script enables skills through symlinks, so edits in `output_skills/` affect every project at once. The `STARTER_CHARACTER` convention only works if you add a line to the global `~/.claude/CLAUDE.md`. Anthropic's guidance changes, so run `./update-docs` from time to time. Read generated skills before enabling them, as with any code that gives an agent file access.
How to disable. Disable skills with `./skills toggle`, delete the repository folder and remove the symlinks from `~/.claude/skills`.
FAQ
How do I start?
Open the repository folder in Claude Code and ask for a new skill. The agent asks its own questions.
Where does the finished skill go?
Into `output_skills/<category>/<name>`. `./skills toggle` enables it globally.
What is STARTER_CHARACTER?
An emoji defined at the top of a skill: with a rule in the global `CLAUDE.md`, Claude starts replies with it to show the skill is loaded.
How do I refresh Anthropic's guidance?
Run `./update-docs`; it pulls the latest material.
Does it fit Cursor or Codex?
It is written for Claude Code; the README does not describe other agents' formats.
Discussion
No questions or reviews yet.
Related
Skills11 skills for Claude Code: SKILL.md authoring, prompt tuning, refactoring, change review and planning in files.
A skill for creating skills that proves them by behavior: before-and-after runs, regression evals and ecosystem audit
A set of skills, slash commands and agents for Claude Code that helps you build your own skills, agents, hooks, slash commands and prompts
A set of ready-to-copy Claude Code skill folders: PDF, slides, Next.js, image and video generation and social publishing.
A desktop app and CLI to install, sync and organize agent skills across Claude Code, Codex, Cursor and other tools
A meta-skill that takes a vague idea to a finished agent skill: it designs, writes the files, checks them and refines by evidence
Comparison with similar tools
Cards that share the main search query, and how they differ.
| Tool | Official | Repository stars | Main difference |
|---|---|---|---|
| Skill Factory (this card) | no | 242 | |
| skill-creator | no | 2,295 | turns OpenAPI, GraphQL or MCP into a skill; Skill Factory guides authoring of any skill from questions to review |
| Functional Skill Creator | no | 257 | splits a long SKILL.md into functions; Skill Factory creates a skill from scratch and keeps an example library |
README in short
The README explains why skills matter, the quick start, global enabling through `./skills`, the `STARTER_CHARACTER` rule and documentation updates.