# HAL for agents

The canon as data, the method, every image with its prompt, every page as Markdown, and a skill that holds all of it.

## Install the skill

```
curl -sL https://halhalai.com/skill/hal-skill.zip -o hal-skill.zip
unzip -o hal-skill.zip -d ~/.claude/skills/
```

## Read first

1. [The recipe](https://halhalai.com/wiki/recipe.md)
2. [The canon](https://halhalai.com/hal.md): paste the identity text word for word, once.
3. [His face sheet](https://halhalai.com/kit/hal-face-sheet.png): @Image2 of every take that shows his face.
4. Ask the person at the six checkpoints. Never spend on video before checkpoint 5.

## Endpoints

- `/llms.txt`: The site for language models: the index
- `/llms-full.txt`: The canon, the whole wiki and every film in one file
- `/hal.md`: The canon as Markdown
- `/skill/hal-skill.zip`: The agent skill: SKILL.md, the wiki, the films, the records, a fetch script
- `/kit/hal-kit.zip`: The kit: the scripts the films are made with (Python and Node)
- `/api/index.json`: The site, the endpoints and the item fields
- `/api/canon.json`: The character as data
- `/api/films.json`: The films, their numbers and documents
- `/api/films/{film}.json`: One film with every image and its prompt
- `/api/images.json`: Every image with its prompt
- `/api/images/{id}.json`: One image
- `/api/guides.json`: The wiki pages, with their Markdown addresses
- `/api/records.json`: The making-of documents

An MCP server is planned after the first film.
