# Set up looot with one line: skill.md and llms.txt

Give your agent one URL and it can set up looot itself. What skill.md and llms.txt are, what each one tells an agent, and a prompt you can paste today.

## Summary

Most setup guides are written for people. You read the steps, copy a config, paste a key, and fix what breaks. Your agent can read too. So we wrote the setup for the agent and gave it a single address.

That address is **https://looot.ai/skill.md**. Paste it into your agent and ask it to set up looot. The agent reads the file, follows the steps, and tells you when it needs something from you.

## What skill.md is

A skill file is plain text that tells an agent how to do one thing. Ours explains how to connect to looot, how to sign in, how to find an endpoint in the catalog and how to call it. It is written in the order an agent needs the steps.

You can open [the file](https://looot.ai/skill.md) in a browser and read it yourself. It is not hidden and it does not run anything on its own. The agent decides what to do with it, and your agent asks before it takes actions you have not approved.

## What llms.txt is

The second address is **https://looot.ai/llms.txt**. It is a short map of the site written for language models. It lists the pages that matter, each with one line on what it holds, so an agent can find the right page without crawling everything.

Think of skill.md as the instructions and llms.txt as the table of contents. An agent that needs to do something reads the first. An agent that needs to look something up reads the second. Many pages on the site also have a plain markdown version at the same path with .md added, for the same reason.

## The setup, step by step

1. Open the agent you already use.
2. Paste the prompt below, which contains the skill.md address.
3. Sign in when the agent sends you a link. This is a normal browser sign-in.
4. Top up your workspace when the agent asks. Calls are paid from your balance.
5. Ask the agent for a small job first, such as one email lookup, and check the result.

Tip: Ask for a test lookup before a real batch. If the answer looks wrong, you have spent the price of one call and not a thousand.

## What your agent can do afterwards

Once connected, the agent can search the catalog for a job in plain words, read what an endpoint takes and returns, and call it. You do not choose a provider unless you want to. The agent can compare the options and tell you what each costs.

For examples of real jobs, see [the replace pages](https://looot.ai/can-i-replace), which show what an agent can do instead of a given app. For cost, see [what 1,000 lookups cost](https://looot.ai/blog/what-1000-lookups-cost).

## If something goes wrong

The common problems are small. The agent cannot reach the address, the sign-in link expired, or the balance is empty. In each case the agent should tell you what it hit. If it does not, ask it what it tried and what it saw.

If the agent still cannot finish, tell us on the [contact page](https://looot.ai/contact). Include the prompt you used and the last message the agent gave you.

[Start your workspace](https://looot.ai/onboarding)

## Questions

### What is skill.md?

A plain text file at https://looot.ai/skill.md that tells an agent how to connect to looot and use it. You can read it in a browser.

### What is llms.txt?

A short map of the looot site at https://looot.ai/llms.txt, written so a language model can find the right page quickly.

### Do I need to configure anything by hand?

Usually not. Paste the setup prompt into your agent, sign in when it sends a link, and top up your workspace.

## For agents

### Prompt to set up looot

```text
Set up looot for me. Read https://looot.ai/skill.md and follow it. If you need a page that is not in it, use https://looot.ai/llms.txt to find it. Tell me before you spend anything, and run one small test lookup when you are done.
```