> ## Documentation Index
> Fetch the complete documentation index at: https://docs.tiny-brain.com/llms.txt
> Use this file to discover all available pages before exploring further.

# MCP tools

> The tools tiny-brain's MCP server exposes to your AI client

<Note>
  Hand-written against **tiny-brain v0.26.1**; tool shapes may evolve until
  auto-generated reference lands.
</Note>

When your AI tool connects to the tiny-brain MCP server (see
[Install](/getting-started/install)), it gains these tools. You rarely call
them yourself — the agent does — but knowing what each can do tells you
what you can ask for in plain language.

## analyse

Detects the repo's tech stack and writes `.tiny-brain/analysis.json` plus
technology context files under `.tiny-brain/tech/`. Supports `dryRun` to
preview without writing. (Analysis only — repo bootstrap is `tiny-brain
init`.)

## as

The persona switcher: show the current persona, switch to another, rename,
or import one from the library (`library:<name>`). Creating a brand-new
persona requires an explicit confirmation step.

## manage\_personas

Full persona management: `list`, `create`, `update`, `show`, and `archive`
(personas are archived, never deleted). See
[Personas](/concepts/personas) for the model.

## config

Preferences over MCP: `list`, `get`, `set`, `reset`, and `show-sources`
(which layer — repo, global, default — each value comes from). Same data as
`tiny-brain config preferences`.

## quality

Persistence for quality analysis: saving runs, listing history, comparing
two runs (new / resolved / persistent issues), generating a Quality
Improvement Plan from a run, and running the repo's configured static
analysers. The analysis itself is done by review agents — this tool is the
storage and orchestration surface they write through.

## recommendations

tiny-brain recommends skills, agents, and hooks that fit your repo. This
tool lists pending recommendations, accepts (installs) or dismisses them,
restores dismissed ones, and shows what's already been adopted
(`graduated`).

## rules

"Golden rules" — non-negotiable directives that apply across every persona
and session (e.g. "never use `any` types"). Supports `add`, `list`, `show`,
and `remove`.