# Slash commands

> The commands you can type in the message box - /summarize, /init and /review - and what each one does.

Type `/` at the start of the message box to see the commands. Use `↑` and `↓` to choose one and `Enter` to run it.

| Command | What it does |
|---|---|
| `/summarize` | Summarizes the older messages of this chat to free up [chat memory](https://rogenie.ai/docs/chat-memory). |
| `/init` | Drafts this place's [project notes](https://rogenie.ai/docs/memory#project-notes): the conventions and decisions every future chat should know. |
| `/review` | Runs a fresh-context security and quality [review](https://rogenie.ai/docs/code-review) of the scripts this chat wrote. |

## /summarize

Compacts the older turns of the open chat into a summary, so the model keeps working with a smaller, cheaper conversation. RoGenie also does this automatically when a chat's memory fills up. Summarizing makes one model call, which uses a little quota. See [Chat memory and summarizing](https://rogenie.ai/docs/chat-memory).

## /init

Asks RoGenie to inspect your place and draft its project notes, stored in the place itself at `ServerStorage.RoGenie.ProjectNotes`. You can add a focus after the command:

- `/init`
- `/init the combat system and its remotes`

The draft is a normal run: RoGenie reads your place, writes the notes, and you can ask for changes. Studio must be linked. See [Memory: instructions, notes and project notes](https://rogenie.ai/docs/memory#project-notes).

## /review

Starts a code review of the scripts this chat wrote. Add a focus to narrow it:

- `/review` reviews for both security and quality,
- `/review security` looks for exploits, such as a server that trusts a price sent by the client,
- `/review quality` looks for bugs, performance problems and maintainability.

The review runs separately from the chat and does not add a message to it. See [Code review](https://rogenie.ai/docs/code-review).