--- title: "Agent Skills" description: "Publish Agent Skills from your Movk Nuxt Docs site, following the Cloudflare Agent Skills Discovery RFC and deployed automatically via the /.well-known/skills/ endpoint. Users can install your skills into AI assistants such as Claude Code and Cursor with a single command." canonical_url: "https://docs.mhaibaraai.cn/en/docs/getting-started/skills" --- # Agent Skills > Publish Agent Skills from your Movk Nuxt Docs site, following the Cloudflare Agent Skills Discovery RFC and deployed automatically via the /.well-known/skills/ endpoint. Users can install your skills into AI assistants such as Claude Code and Cursor with a single command. ## What are Agent Skills? Movk Nuxt Docs automatically discovers the skills in the `skills/` directory and, following the [Cloudflare Agent Skills Discovery RFC](https://github.com/cloudflare/agents/blob/main/docs/skills-discovery.md), deploys them under the `/.well-known/skills/` endpoint. With a single command, users can install your skills into AI assistants such as Claude Code and Cursor. Agent Skills is a lightweight open format for giving AI assistants dedicated knowledge and workflows. Each Skill is a `SKILL.md` file that describes what tasks the AI assistant can accomplish with your product and how to do them, and it can include any number of reference files. > [!NOTE] > See: https://docs.mhaibaraai.cn/.well-known/skills/index.json > > See the Skills published by this documentation site. ## Built-in Skills This site already publishes the following two ready-to-use Skills by default, which AI assistants can use immediately after installation: ### create-docs Generate a complete documentation site for any project, automatically including search, dark mode, MCP Server, and llms.txt integration. **Trigger words**: `create docs`, `add documentation`, `initialize docs`, `generate documentation`, `scaffold a documentation site`, and more. ### review-docs Review Markdown documentation for quality, clarity, SEO, and technical correctness, producing prioritized, actionable suggestions (Critical / Important / Suggested). Tuned for Movk Nuxt Docs / Nuxt Content but compatible with any Markdown docs. **Trigger words**: `review docs`, `check documentation`, `validate MDC syntax`, `check SEO issues`, and more. ### One-command install ```bash [Terminal] npx skills add https://docs.mhaibaraai.cn ``` The CLI automatically detects installed agents such as Claude Code, Cursor, and Windsurf, and installs both Skills together. ## Creating a Skill ### Create the skills directory Create a `skills/` directory in your project root, and inside it create a skill directory containing a `SKILL.md` file: ```bash my-docs/ └── skills/ └── my-skill/ └── SKILL.md ``` ### Write SKILL.md Follow the [agentskills.io specification](https://agentskills.io/specification). The `description` field in the frontmatter is required; if you omit the `name` field, the directory name is used as the Skill name by default. ```md [skills/my-skill/SKILL.md] --- name: my-skill description: | Describe what this Skill is for. List what users say that should trigger this Skill. --- # My Skill ## Workflow 1. Step one 2. Step two ``` ### Deploy Once deployed, Movk Nuxt Docs automatically publishes the Skills to `/.well-known/skills/`. ### Use the Skills Users add your published Skills to their AI assistant with the following command: ```bash [Terminal] npx skills add https://your-docs-site.com ``` The CLI detects installed agents (Claude Code, Cursor, Windsurf, and others) and installs the skill into all of them. ## Naming conventions The Skill name (both the directory name and the `name` field) must follow the [agent skills naming specification](https://agentskills.io/specification#name-field): - 1-64 characters - Only lowercase letters, numbers, and hyphens (a-z, 0-9, -) - Cannot start or end with a hyphen - Cannot contain consecutive hyphens (--) - The `name` field in the frontmatter must match the parent directory name exactly > [!NOTE] > > Skills that fail validation are skipped; check the build output for warnings. ## Directory structure A Skill directory can contain any number of supporting files, which the AI assistant fetches on demand when invoking the Skill: ```bash skills/ └── my-skill/ ├── SKILL.md ├── references/ │ ├── style-guide.md # Writing style guide │ └── checklist.md # Checklist └── assets/ └── template.md # Page template ``` All files are automatically listed in the `index.json` index and served at their respective paths under `/.well-known/skills/{skill-name}/`. > [!TIP] > > Keep the main > > SKILL.md > > file under 500 lines. Move detailed reference material into separate files in > > references/ > > —agents load these on demand, so smaller files mean less context usage. ## Publishing multiple Skills A single documentation site can publish multiple independent Skills: ```bash skills/ ├── create-docs/ # Create a documentation site ├── review-docs/ # Review documentation quality └── code-review/ # Code review ``` All Skills are listed in `index.json`, and the AI assistant selects automatically based on user intent. ## Comparison with llms.txt Both Agent Skills and `llms.txt` help AI tools use your documentation, but they serve different purposes:
llms.txt Agent Skills
Purpose A directory of all documentation pages A capability summary with actionable instructions
Content Page titles, descriptions, and links Step-by-step workflows, code examples, constraints
Loading time At discovery time On demand, when the skill is activated
Format Plain text with links Markdown with YAML frontmatter
Best for Helping agents find information Teaching agents how to use your product
> [!TIP] > > Use both together: > > llms.txt > > tells agents where to find information, while Skills tell agents what they can accomplish and how. ## Sitemap See the full [sitemap](https://docs.mhaibaraai.cn/sitemap.md) for all pages.