Expert at writing unified documentation for LangChain products using Mintlify MDX format with language-specific code fences and snippet handling
This skill helps you write and maintain documentation for LangChain's unified documentation repository, following the project's established guidelines and the Google Developer Documentation Style Guide.
**These instructions apply to manually authored documentation only. They do NOT apply to:**
1. Document just enough for user success - not too much, not too little
2. Prioritize accuracy and usability of information
3. Make content evergreen when possible
4. Search for existing information before adding new content to avoid duplication
5. Check existing patterns for consistency
6. Reference existing content when possible rather than duplicating
Every MDX file must include:
```yaml
---
title: Clear, descriptive, concise page title
description: Concise summary for SEO/navigation
---
```
Follow the [Google Developer Documentation Style Guide](https://developers.google.com/style):
Use custom code language fences to tag language-specific content:
**Python content:**
```
:::python
Python-specific content here
:::
```
**JavaScript/TypeScript content:**
```
:::js
JavaScript/TypeScript-specific content here
:::
```
When these fences exist on a page at `/concepts/foo.mdx`, two outputs are created:
When adding new navigation groups, ensure the root `index.mdx` is included:
```json
{
"group": "New group",
"pages": ["new-group/index", "new-group/other-page"]
}
```
Note: Include the trailing `/index` (without extension) to avoid parser warnings.
Refer to the [docs.json schema](https://mintlify.com/docs.json) for complete configuration options.
Snippet files in `src/snippets/` are reusable MDX content that can be imported into multiple pages.
**Important:** When writing links in snippets, be careful about path segments. Snippets undergo special link preprocessing that converts absolute `/oss/` links to relative paths. Read the implementation in `pipeline/core/builder.py` method `_process_snippet_markdown_file` (lines 807-872) to understand the preprocessing behavior.
```mdx
---
title: Getting Started with Chains
description: Learn how to create and use chains in LangChain
---
You can create chains to link multiple operations together.
:::python
from langchain.chains import LLMChain
chain = LLMChain(llm=llm, prompt=prompt)
result = chain.run("input text")
:::
:::js
import { LLMChain } from "langchain/chains";
const chain = new LLMChain({ llm, prompt });
const result = await chain.call({ input: "input text" });
:::
```
For questions, refer to the [Mintlify documentation](https://docs.mintlify.com/docs/introduction) or use the Mintlify MCP if available.
Leave a review
No reviews yet. Be the first to review this skill!
# Download SKILL.md from killerskills.ai/api/skills/langchain-documentation-writer-nkrucl/raw