rspress-description-generator — Claude Code skill for Rspress docs
A skill that fills in the description frontmatter field on Rspress pages that lack one. Useful when the docs site has grown and meta descriptions are missing in places.
npx skills add rstackjs/agent-skills --skill rspress-description-generator -a claude-codeWhat rspress-description-generator does
The skill globs **/*.md and **/*.mdx under the docs root and inspects each page's frontmatter. Pages that already carry a non-empty description are skipped. For the rest it writes a 50–160 character plain-text description in the language of the content. Home pages with pageType: home get their text from hero.text, hero.tagline or the features list instead of the body.
It starts by locating the Rspress config (rspress.config.ts, .js, .mjs or .cjs) and reading the root option. When root is absent, the docs folder next to the config is used. i18n projects are handled per language directory, so en/ and zh/ pages get descriptions in their own language. New fields go after title; files without a frontmatter block get one created. Descriptions containing colons or quotes are wrapped in double quotes, otherwise the YAML breaks.
Excluded from the scan: node_modules, build output (doc_build, .rspress, dist), _meta.json and _nav.json, plus **/shared/** folders, which hold snippets rather than pages. On larger sites the skill suggests parallel agent calls grouped by directory, then a second pass to catch anything missed.
What it helps with
- fill in descriptions across a large Rspress docs site
- refresh SEO metadata for every page
- add frontmatter to a new md file that has none
- keep descriptions in the right language per i18n directory
- leave pages with existing descriptions untouched
Example prompts for the agent
- scan the Rspress docs and add description where it's missing
- update meta descriptions for every page under docs/guide
- list the pages in docs that still have no description
How to install rspress-description-generator
1Skills CLI (Vercel)
npx skills add rstackjs/agent-skills --skill rspress-description-generator -a claude-code2Manually: put the files in place
~/.claude/skills/rspress-description-generator/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/rspress-description-generator/
1Skills CLI (Vercel)
npx skills add rstackjs/agent-skills --skill rspress-description-generator -a codex2Manually: put the files in place
.agents/skills/rspress-description-generator/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add rstackjs/agent-skills --skill rspress-description-generator -a cursor2Manually: put the files in place
.cursor/skills/rspress-description-generator/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install rstackjs/agent-skills rspress-description-generator2Manually: put the files in place
.github/skills/rspress-description-generator/SKILL.md1Skills CLI (Vercel)
npx skills add rstackjs/agent-skills --skill rspress-description-generator -a gemini-cli2Platform CLI
gemini skills install https://github.com/rstackjs/agent-skills --path skills/rspress-description-generatorManually: put the files in place
.devin/skills/rspress-description-generator/SKILL.mdThe legacy .windsurf/skills/ path still works
Skills CLI (Vercel)
npx skills add rstackjs/agent-skills --skill rspress-description-generator -a opencodeClaude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about rspress-description-generator
How do I install rspress-description-generator in Claude?
Run npx skills add rstackjs/agent-skills --skill rspress-description-generator -a claude-code. The skill file lands at ~/.claude/skills/rspress-description-generator/SKILL.md.
What does rspress-description-generator do?
It generates missing description frontmatter for Rspress Markdown and MDX pages and inserts it into the frontmatter block. Pages that already have a description are left alone.
How does the skill find the docs root?
It reads the root option from the Rspress config. Without it, the docs folder next to the config is used, falling back to doc.
Source text
SKILL.md · MIT
---
name: rspress-description-generator
description: Generate missing description frontmatter for Rspress Markdown/MDX pages, including new docs pages and site-wide SEO metadata updates.
---
# Rspress description generator
The `description` field in Rspress frontmatter generates `<meta name="description" content="...">` tags, which are used for search engine snippets, social media previews, and AI-oriented formats like llms.txt.
## Step 1 — locate the docs root
1. Find the Rspress config file. Search for `rspress.config.ts`, `.js`, `.mjs`, or `.cjs`. It may be at the project root or inside a subdirectory like `website/`.
2. Read the config and extract the `root` option.
- The value might be a plain string (`root: 'docs'`) or a JS expression (`root: path.join(__dirname, 'docs')`). In either case, determine the resolved directory path.
- If `root` is set, resolve it relative to the config file's directory.
- If `root` is not set, default to `docs` relative to the config file's directory.
3. Confirm the directory exists. If neither `docs` nor the configured root exists, check for `doc` as a fallback.
## Step 2 — detect i18n structure
Rspress i18n projects place language subdirectories (e.g., `en/`, `zh/`) directly under the docs root:
```
docs/
├── en/
│ ├── guide/
│ └── index.md
└── zh/
├── guide/
└── index.md
```
Check if the docs root contains language subdirectories (two-letter codes like `en`, `zh`, `ja`, `ko`, etc.). If so, process each language directory separately — the description language should match the content language.
If there are no language subdirectories, treat the entire docs root as a single-language site.
## Step 3 — scan and process files
Glob for `**/*.md` and `**/*.mdx` under the docs root. Exclude:
- `node_modules`, build output (`doc_build`, `.rspress`, `dist`)
- `_meta.json` / `_nav.json` (sidebar/nav config files, not doc pages)
- `**/shared/**` directories (reusable snippets included via `@import`, not standalone pages)
For each file:
1. **Read the file.**
2. **Check for existing `description` in frontmatter.** If it exists and is non-empty, skip.
3. **Check `pageType` in frontmatter.** For `home` pages, derive the description from the `hero.text` / `hero.tagline` fields or the features list, not from body content.
4. **Generate a description** following the writing guidelines below.
5. **Insert `description` into frontmatter:**
- If the file has frontmatter with a `title` field, insert `description` on the line after `title`.
- If the file has frontmatter without `title`, insert `description` as the first field.
- If the file has no frontmatter block, add one:
```yaml
---
description: Your generated description here
---
```
### YAML formatting
Most descriptions can be bare YAML strings:
```yaml
description: Step-by-step guide to setting up your first Rspress site
```
If the description contains colons, quotes, or other special YAML characters, wrap in double quotes:
```yaml
description: 'API reference for Rspress configuration: plugins, themes, and build options'
```
## Step 4 — batch processing
For sites with many files, use parallel agent calls to process independent files simultaneously. Group by directory (e.g., all files in `guide/`, then all in `api/`) to maintain focus and consistency within each section.
After processing all files, do a quick scan to ensure no files were missed — re-glob and check for any remaining files without `description`.
## Description writing guidelines
The description serves three audiences: search engines (Google snippet), AI systems (llms.txt, summarization), and humans (scanning search results). A good description helps all three.
### Rules
- **Length**: 50–160 characters. Under 50 is too vague for search engines; over 160 gets truncated in snippets.
- **Language**: Match the document content. Chinese docs get Chinese descriptions, English docs get English descriptions.
- **Be direct**: State what the page covers. Avoid starting with "This document", "This page", "Learn about" — jump straight to the substance.
- **Be specific**: Mention concrete technologies, APIs, or concepts the page covers. "Configure Rspress plugins for search, analytics, and internationalization" beats "How to use plugins."
- **No markdown**: Plain text only, no formatting syntax.
### Examples
**Good:**
| Content | Description |
| -------------------------- | ---------------------------------------------------------------------------- |
| Plugin development guide | Create custom Rspress plugins using the Node.js plugin API and runtime hooks |
| MDX component usage | Import and use React components in MDX documentation files |
| Rspress 快速开始 | 从安装到本地预览,搭建 Rspress 文档站点的完整流程 |
| 主题配置 | 自定义 Rspress 主题的导航栏、侧边栏、页脚和暗色模式 |
| Home page (pageType: home) | Rspress documentation framework — fast, MDX-powered static site generator |
**Bad:**
| Description | Why |
| ------------------------------------------------------- | ------------------------------------------------ |
| "About plugins" | Too vague — which plugins? what about them? |
| "This page explains how to configure the Rspress theme" | Wastes characters on "This page explains how to" |
| "Learn everything about Rspress!" | Marketing fluff, says nothing specific |
## Documentation
- Frontmatter fields: <https://rspress.rs/api/config/config-frontmatter>
- Basic config (`root` option): <https://rspress.rs/api/config/config-basic>
- Full Rspress docs: <https://rspress.rs/llms.txt>
© rstackjs · MIT · rstackjs/agent-skills