Guide · Agent skill
How to use the create-llms agent skill: build an llms.txt file
Updated · written with AI from the project's own documentation and checked against our catalogue data
The create-llms skill is a set of written instructions that teaches an AI agent to build an llms.txt file for a repository. An llms.txt file is a short Markdown index placed in the repository root that points language models to the project's most important documentation. This guide explains what the skill does, how to install it, how to trigger it, and where it falls short, so you can decide whether it fits your project.
Quick start
- 1Download the ZIP. Download the create-llms ZIP from this page and unzip it. The folder contains SKILL.md.
- 2Choose an install location. Place the skill in ~/.claude/skills/ for your user account, or in .claude/skills/ inside one project.
- 3Or upload in the Claude apps. In the Claude apps, go to Settings → Capabilities → Skills and upload the ZIP.
- 4Ask for an llms.txt file. Make a plain request such as asking for an llms.txt file for this repository based on its structure.
- 5Let the agent review the spec and repo. The agent reviews the llmstxt.org specification, scans the structure and docs, and plans the sections.
- 6Check the file. Open the generated llms.txt in the repository root, verify every relative link and the heading structure, and fix anything wrong.
What the create-llms skill makes the agent do
An agent skill is a folder containing a SKILL.md file with instructions that an AI agent reads when a task matches the skill's description. The create-llms skill covers one task: creating an llms.txt file from scratch based on a repository's structure, following the llms.txt specification at https://llmstxt.org/.
The skill walks the agent through four stages: reviewing the llms.txt spec, examining the repository structure and documentation, grouping key files into sections, and writing each link with a short description. The file is created in the repository root as llms.txt.
- Reviews the llms.txt specification at llmstxt.org before writing
- Scans the repository structure and finds READMEs, docs, specs, configuration and examples
- Groups key files into H2 sections such as Documentation, Specifications, Examples, Configuration and Optional
- Writes each entry as a Markdown link with a short description
What the llms.txt file looks like
The skill asks for a fixed layout: an H1 with the project name, an optional blockquote summary, optional context paragraphs without headings, and H2 sections containing lists of links. Each link follows the pattern [descriptive-name](relative-url): optional description.
The skill includes a template that shows this layout. Its Optional section has a special meaning in the spec: it marks files that can be skipped when a shorter context is needed.
# [Repository Name]
> [Concise description of the repository's purpose and scope]
## Documentation
- [Main README](README.md): Primary project documentation and getting started guide
## Optional
- [Architecture Documentation](docs/architecture.md): Detailed system architectureInstall the create-llms skill
Download the create-llms ZIP from this page and unzip it. The skill is a single file, SKILL.md. Put the unzipped skill folder in one of these locations:
- For your user account: ~/.claude/skills/
- For one project only: the .claude/skills/ folder inside that project
- In the Claude apps: upload the ZIP under Settings → Capabilities → Skills
Trigger the skill with a natural request
You do not need a special command. Ask in plain language for an llms.txt file, and the agent will match the request to the skill's description. Be specific about the repository you want covered and whether you want the file written to the root.
Once the agent creates the file, check the result yourself, because the skill relies on the agent to check links by hand.
- Create an llms.txt file for this repository based on its structure.
- Write an llms.txt in the repo root that points LLMs to our setup guide, API spec and examples.
- Review the docs folder and draft an llms.txt with an Optional section for design notes.
Review the output and fix links
The skill's own validation steps ask the agent to check that links are valid and that the file follows the spec. It also asks the agent to check the spec at llmstxt.org. The skill does not include a validator, and the agent may not be able to fetch the live spec, so compliance is not guaranteed.
Before you commit the file, open it and confirm each relative link points to a file that exists, the H1 and H2 structure is correct, and the descriptions are accurate.
- Click or open every relative link and confirm the target file exists
- Confirm there is exactly one H1 with the project name
- Remove or rename any section that does not match the spec
- Compare the file against the llmstxt.org specification yourself
Safety notes: scripts and network access
The skill is written instructions only. It contains no scripts, it does not require any listed tools, and it does not use the network according to its listing. The agent will still read your repository files and write a new llms.txt file in the root, so review the diff before committing.
The skill refers to llmstxt.org, but the skill itself does not fetch that page. Whether the agent can open it depends on the environment you run it in.
- Runs scripts: no
- Uses network: no
- Requires: nothing listed
- Review the generated file before committing it to version control
What people use it for
- Adding an llms.txt index to an open-source library so AI coding assistants find its setup guide and API docs quickly
- Helping an internal team's AI tools locate architecture notes, runbooks and configuration files
- Pointing language models at specs and examples in a documentation-heavy repository
- Drafting a first version of llms.txt that a maintainer then edits by hand
Frequently asked questions
What is an llms.txt file?
An llms.txt file is a Markdown file placed in a project's root that gives language models a short index of where the most useful documentation lives. The create-llms skill describes it as following the specification at llmstxt.org. It is meant to be readable by both people and machines.
How do I install the create-llms skill?
Download the ZIP from this page and unzip it. Put the folder in ~/.claude/skills/ for your user account or in .claude/skills/ inside a project. In the Claude apps, you can instead upload the ZIP under Settings → Capabilities → Skills.
Does the create-llms skill require anything else?
The listing shows no required tools or dependencies. The skill runs no scripts and uses no network access according to its listing. It consists of one SKILL.md file with written instructions.
Is create-llms free and what is its licence?
The skill is published under the MIT licence in the github/awesome-copilot repository. The listing does not mention a price, so check the repository for any terms that apply.
Does the skill validate that my llms.txt follows the spec?
No. The skill has no validator. It tells the agent to check the spec and to check links manually, but the agent may not be able to fetch llmstxt.org, so you should review the file yourself.
What is the difference between this skill and a dedicated llms.txt generator?
The skill relies on the agent to read your repository and write the file from written instructions. A dedicated tool that generates and validates the file automatically would be a better fit if you need verified output on every build. This skill suits one-off or occasional drafting, followed by manual review.
New agent skills, repos and AI tools — every week
The best new skills and open-source projects, the fastest-growing repos, and one free prompt. Five minutes to read. Unsubscribe anytime.
Our verdict
Use create-llms if you maintain a repository and want an AI agent to draft a clear llms.txt index that you then review and edit. Pick a different option if you need a fully automated, validated generator or a guaranteed check against the live llmstxt.org specification.
Download the skill