LLMs.txt Generator Skill
Generate llms.txt and llms-full.txt files following the llmstxt.org protocol specification to make projects more accessible to Large Language Models.
What is llms.txt?
The /llms.txt protocol is a standard for providing structured information about a website/project to LLMs. It helps LLMs quickly understand the project structure and access key documentation without parsing complex HTML pages.
Protocol Benefits
-
Better LLM Interaction: LLMs can quickly understand project capabilities
-
Accurate Responses: Provides authoritative documentation links
-
Developer Experience: Improves AI-assisted development
-
Discoverability: Makes projects accessible to AI coding assistants
File Structure Requirements
llms.txt Format (Mandatory Structure)
The file MUST follow this exact order:
Project Title
Brief project summary in a blockquote (1-2 sentences)
Additional project details (optional paragraphs/lists, no H2-H6 headers here)
Section Name
- Link Title: Description of the link
- Another Link
Optional
- Secondary Info: Links that can be skipped when context is limited
Two File Types
File Purpose Size
llms.txt
Concise overview with essential links ~100-200 lines
llms-full.txt
Comprehensive documentation with code examples ~500-1000 lines
Generation Workflow
Step 1: Gather Project Information
To generate accurate llms.txt files, collect the following information:
-
Project README: Read README.md for project overview, features, and quick start
-
Documentation Structure: List files in docs/ directory to identify available guides
-
API Reference: Find API documentation files
-
Examples: List examples in examples/ directory
-
Package Info: Read package.json , pyproject.toml , or Cargo.toml for metadata
Step 2: Generate llms.txt (Concise Version)
Create llms.txt with:
-
H1 Title: Project name
-
Blockquote: One-sentence project summary
-
Key Info:
-
Package installation command
-
Supported platforms/versions
-
License
-
Repository URL
-
Key Features: Bullet list of main features (5-10 items)
-
Documentation Section (## Documentation ):
-
Getting started guide
-
Installation
-
Core concepts
-
Architecture overview
-
API Reference Section (## API Reference ):
-
Main API documentation links
-
Integration/Platform Sections (if applicable):
-
Platform-specific guides
-
Optional Section (## Optional ):
-
Advanced guides
-
Examples
-
Contributing guide
Step 3: Generate llms-full.txt (Comprehensive Version)
Create llms-full.txt with everything from llms.txt plus:
-
Extended Features: Complete feature list with descriptions
-
Technical Stack: Detailed technical information
-
Quick Start Code: Actual code examples for common use cases
-
Complete API Reference: All API classes and methods with signatures
-
All Documentation Links: Every guide and reference document
-
Code Examples: Inline code snippets demonstrating key functionality
-
RFCs/Architecture Docs: Technical design documents
-
Full Examples List: All example files with descriptions
Format Rules
Link Format
- Title: Brief description
Code Blocks in llms-full.txt
Include actual code examples:
Quick Start
```python from mypackage import MyClass
instance = MyClass() instance.do_something() ```
Section Guidelines
Section Required Content
H1 Title Yes Project name only
Blockquote Yes 1-2 sentence summary
Key Features Yes 5-10 main features
Documentation Yes Core documentation links
API Reference Yes API documentation links
Optional No Secondary/advanced content
Example Templates
llms.txt Template
ProjectName
Brief description of the project in one or two sentences.
ProjectName is a [type of project] that provides [main capability]. Built with [technology], it offers [key benefit].
- Package:
pip install projectname - Platforms: Windows, macOS, Linux
- License: MIT
- Repository: https://github.com/owner/projectname
Key Features
- Feature 1 description
- Feature 2 description
- Feature 3 description
Documentation
- Getting Started: Quick start guide
- Installation: Installation instructions
- Core Concepts: Core concepts
API Reference
- Main API: Main API reference
Optional
- Advanced Usage: Advanced features
- Examples: Code examples
llms-full.txt Template
ProjectName
Brief description of the project in one or two sentences.
ProjectName is a [type of project] that provides [main capability]. Built with [technology], it offers [key benefit].
- Package:
pip install projectname - Platforms: Windows, macOS, Linux
- License: MIT
- Repository: https://github.com/owner/projectname
- PyPI/NPM: https://pypi.org/project/projectname/
Key Features
- Feature 1: Detailed description of feature 1
- Feature 2: Detailed description of feature 2
- Feature 3: Detailed description of feature 3
Technical Stack
- Core: Technology 1, Technology 2
- Runtime: Platform requirements
- Packaging: Build system details
Quick Start
Installation
```bash pip install projectname ```
Basic Usage
```python from projectname import MainClass
instance = MainClass(param="value") result = instance.method() print(result) ```
Core API
MainClass
```python MainClass( param1: str = "default", param2: int = 100, ) ```
Methods
```python instance.method1() # Description instance.method2(arg) # Description ```
Documentation
- Getting Started: Quick start guide
- Installation: Installation instructions
- Core Concepts: Core concepts
- Architecture: System architecture
API Reference
- Main API: Main API reference
- Secondary API: Secondary API reference
Advanced Guides
- Advanced Topic 1: Description
- Advanced Topic 2: Description
Examples
Update Workflow
To update existing llms.txt files:
-
Read Current Files: Load existing llms.txt and llms-full.txt
-
Identify Changes: Compare with current documentation structure
-
Update Sections: Add new documentation links, update descriptions
-
Maintain Format: Preserve the protocol structure
-
Validate: Ensure all links are valid and descriptions are accurate
Validation Checklist
Before finalizing, verify:
-
H1 title is present and matches project name
-
Blockquote summary is present
-
Key information (package, platforms, license, repo) is included
-
All links use full GitHub URLs (not relative paths)
-
Links include : description format
-
Optional
section contains secondary content
-
No H2-H6 headers before the first ## Section
-
Code examples in llms-full.txt are syntactically correct
-
All referenced files exist in the repository
Common Patterns
For Python Projects
- Package:
pip install packagename - Python: 3.7+
For Rust Projects
- Crate:
cargo add cratename - Rust: 1.70+
For Node.js Projects
- Package:
npm install packagename - Node.js: 18+
For Multi-Language Projects
- Python Package:
pip install packagename - Rust Core: Built with Rust 1.75+
- Python: 3.7+