agentleFS
littlebearappslittlebearapps/pitchdocs/.claude/skills/llms-txt/SKILL.md

llms-txt

Skill8 starsChanged 8 months agoMITView on GitHub

Generates llms.txt and llms-full.txt files following the llmstxt.org specification. Provides LLM-friendly content curation for AI coding assistants (Cursor, Windsurf, Claude Code) and AI search engines. Use when generating or updating llms.txt for a repository.

14 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

8 rules. 0 are common, 8 are distinctive, and 0 are very specific to this project.

  • 0

    Common

    Most of the 14 edited versions we read keep these.

  • 8

    Distinctive

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    No other heading levels are used: Only H1 (one, at the top) and H2 (for sections).

    line 31
  2. Distinctive:

    Read the primary manifest for the project name and description:

    line 44
  3. Distinctive:

    Check for these files and directories:

    line 56
  4. Distinctive:

    Use the feature-benefits approach: describe what the reader gains from reading that file.

    line 89
  5. Distinctive:

    After adding or removing documentation files

    line 188
  6. Distinctive:

    After major version releases

    line 189
  7. Distinctive:

    After restructuring the docs directory

    line 190
  8. Distinctive:

    Add to the release checklist alongside CHANGELOG updates

    line 191

“Common” is measured: kept by at least 6 in 10 of the 14 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

llms.txt Generator

Generate structured, LLM-friendly content indexes following the llmstxt.org specification.

Background

llms.txt was proposed by Jeremy Howard (Answer.AI) in September 2024. It provides a curated Markdown file that gives LLMs a structured map of a project's most important content — solving the problem of context windows being too small to process entire websites or repositories.

Adopted by: Anthropic, Cloudflare, Stripe, Vercel, Cursor, Mintlify, GitBook, Fern.

Used by: Cursor, Windsurf, Context7 MCP, Claude Code (reading local files), AI search engines.

Specification (llmstxt.org)

An llms.txt file is a Markdown document with sections in this exact order:

  1. H1 heading (required) — the name of the project or site
  2. Blockquote (optional) — short summary with key information for understanding the rest of the file
  3. Body text (optional) — zero or more Markdown sections of any type except headings
  4. H2 sections with file lists (optional) — each contains a Markdown list where every item has:
    • A required hyperlink: [name](url)
    • Optionally a : followed by notes about the file
  5. ## Optional section (special) — URLs here can be skipped when shorter context is needed

No other heading levels are used. Only H1 (one, at the top) and H2 (for sections).

Two Output Files

FileContentSize TargetUse Case
llms.txtIndex with links and descriptionsUnder 10K tokensReal-time AI assistants navigating quickly
llms-full.txtConcatenated Markdown of all referenced filesVaries (can be 100K+ tokens)RAG ingestion, IDE indexing, full-context tools

Generation Workflow

Step 1: Gather Project Metadata

Read the primary manifest for the project name and description:

FileName FieldDescription Field
package.jsonnamedescription
pyproject.toml[project].name[project].description
Cargo.toml[package].name[package].description
go.modmodule pathFirst line of README
.claude-plugin/plugin.jsonnamedescription

Step 2: Scan for Documentation Files

Check for these files and directories:

Primary docs:

  • README.md
  • docs/ directory (hub page, guides, API reference)
  • examples/ directory

Supporting docs:

  • CONTRIBUTING.md
  • CHANGELOG.md
  • SECURITY.md
  • CODE_OF_CONDUCT.md
  • ROADMAP.md
  • LICENSE

Code entry points (include only if the project is a library/framework):

  • src/index.* or lib/index.*
  • Config files with schema documentation

Step 3: Write Descriptive Annotations

For each file, write a benefit-focused description — not just the file name:

Good:

- [Getting Started](./docs/guides/getting-started.md): Install, configure, and deploy your first worker in under 5 minutes

Bad:

- [Getting Started](./docs/guides/getting-started.md): Getting started guide

Use the feature-benefits approach: describe what the reader gains from reading that file.

Step 4: Assemble llms.txt

For repositories (local paths):

# [Project Name]

> [Description from manifest or README first paragraph]

[Optional body text: language, framework, key technical context]

## Docs

- [README](./README.md): Project overview, value proposition, and quick start
- [API Reference](./docs/api.md): Complete endpoint documentation with authentication and error codes
- [Configuration](./docs/configuration.md): All config options with defaults and examples

## Guides

- [Getting Started](./docs/guides/getting-started.md): Install, configure, and run your first example in under 5 minutes
- [Deployment](./docs/guides/deployment.md): Production deployment to Docker, AWS Lambda, and Cloudflare Workers

## Examples

- [Basic Usage](./examples/basic/): Minimal working examples for common use cases
- [Advanced Patterns](./examples/advanced/): Complex integrations and performance optimisation

## Optional

- [Changelog](./CHANGELOG.md): Version history with user-facing change descriptions
- [Contributing](./CONTRIBUTING.md): Development setup, coding standards, and PR workflow
- [Code of Conduct](./CODE_OF_CONDUCT.md): Community behaviour standards (Contributor Covenant v3.0)
- [Security](./SECURITY.md): Vulnerability reporting process and response timeline
- [License](./LICENSE): MIT license terms

For documentation sites (full URLs):

# [Project Name]

> [Description]

## Docs

- [Getting Started](https://docs.example.com/getting-started): Installation and first steps
- [API Reference](https://docs.example.com/api): Complete API documentation

## Optional

- [Changelog](https://docs.example.com/changelog): Version history
- [Contributing](https://github.com/org/repo/blob/main/CONTRIBUTING.md): How to contribute

Step 5: Generate llms-full.txt (if requested)

Concatenate all referenced files in the same order as llms.txt, with clear separators:

# [Project Name] — Full Documentation

> Complete documentation content for LLM ingestion.

---

## README.md

[Full contents of README.md]

---

## docs/api.md

[Full contents of docs/api.md]

---

[Continue for all referenced files, excluding the Optional section unless specifically requested]

Size management:

  • Skip binary files (images, PDFs)
  • For very large files (>50K tokens), include only the first section or a summary
  • Note the total token count at the end of the file

When to Generate

Project Typellms.txtllms-full.txt
Public repo with docs siteAlwaysAlways (host on docs site)
Public GitHub repoRecommendedOptional (large repos benefit)
Claude Code pluginRecommendedOptional
Small utility / internal toolOptionalSkip

Regeneration

llms.txt should be updated when documentation changes significantly:

  • After adding or removing documentation files
  • After major version releases
  • After restructuring the docs directory
  • Add to the release checklist alongside CHANGELOG updates

Real-World Examples

Projectllms.txtllms-full.txtNotable Pattern
Anthropicdocs.anthropic.com/llms.txt (~8K tokens)481K tokensOrganised by product area
Cloudflaredevelopers.cloudflare.com/llms.txtPer-product files (~3.7M total)Product-specific full files
Stripedocs.stripe.com/llms.txtYesUses Optional for niche products
Vercelvercel.com/docs/llms.txt~400K wordsMulti-product structure

Specification Reference

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.