> For the complete documentation index, see [llms.txt](https://docs.aloop.icustomer.ai/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.aloop.icustomer.ai/iharness/autonomy-levels.md).

# Autonomy levels

The three autonomy modes, Ask, Propose, and Auto: defined by what the system can spend, not by a marketing scale.

iHarness runs at one of three autonomy levels. The levels are not a mood setting; each one is defined by what the system is allowed to spend, credits, sends, budget, your audience's attention.

## The three modes

| Mode        | What it does                                             | Spend                                      |
| ----------- | -------------------------------------------------------- | ------------------------------------------ |
| **Ask**     | Reads, diagnoses, explains. Proposes nothing unprompted. | Spend verbs are absent                     |
| **Propose** | Raises cards, stages work, and waits for your click.     | Present, but gated on your approval        |
| **Auto**    | Completes an arc without a human.                        | Full, within caps and behind a kill switch |

## Absence is a stronger guarantee than a gate

The distinction that matters most is between Ask and everything else. In Ask mode the ability to spend is not gated behind an approval; it is not there at all. A gate can be misconfigured; an absent capability cannot fire. That is why diagnosis and explanation are always safe to ask for, at any level.

## Where the levels meet your day

**Propose** is where approvals live: staged work arrives as cards in [Pulse](/pulse/queue.md), and nothing moves until you click. **Auto** is bounded, not blind: caps limit what an arc can spend, the kill switch stops everything, and every decision still writes its [trace](/concepts/decisions-and-traces.md) before it runs. The per-tool controls that shape all three levels are covered in [Gates](/iharness/gates.md), and the practice of widening autonomy as judgment proves out is covered in [Autopilot vs approval-first](/pulse/autopilot-vs-approval-first.md).


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter:

```
GET https://docs.aloop.icustomer.ai/iharness/autonomy-levels.md?ask=<question>&goal=<endgoal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.