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

# Onboarding

> The guided checklist that takes a new workspace from empty to a working, connected, governed agent.

A new workspace in the Komodor Agentic Operation Platform (KAOP) opens with a checklist rather than a blank screen. It names the handful of steps
that take you from nothing to an agent doing real work, tracks which are done, and lets the
assistant do several of them for you. This page explains what the checklist covers, how progress is
decided, and how to get it out of your way when you are finished.

## Where you see it

The checklist follows you rather than living on one page:

* A **progress button** in the header, showing how many steps are left.
* A **card on the workspace overview**, which you can dismiss once you no longer want it there.
* **Empty states inside a module**, pointing at the step that would fill them.
* The **onboarding page** itself, which is the full list.

Dismissing the overview card only hides that card. The header button and the page stay.

## What the steps cover

The exact list adapts to your workspace and your permissions, but it follows one arc — connect
something, get an agent, watch it work, then make it trustworthy:

| Step                                      | What it means                                                       |
| ----------------------------------------- | ------------------------------------------------------------------- |
| **Connect Datadog / GitHub / Kubernetes** | Give agents something real to observe and act on                    |
| **Bring your own model provider**         | Point KAOP at the model provider you intend to use                  |
| **Add your first agent**                  | Install one from the marketplace, build one here, or bring your own |
| **Watch it run once**                     | Confirm the loop works end to end and read the evidence             |
| **Triage incidents automatically**        | Turn alerts into investigated incidents                             |
| **Review pull requests automatically**    | Put an agent on your pull requests                                  |
| **Set a quality standard**                | Start grading what your agents produce                              |

Some steps nest: *Add your first agent* has child steps for the three ways to get one, and
finishing any of them satisfies the parent.

<Note>
  You only see steps your role lets you act on. A step you cannot see is absent from your counts
  rather than sitting there permanently incomplete.
</Note>

## How a step gets completed

Each step has an **in-app destination** that completes it — following the link and doing the thing
is what moves it. A step is in one of four states:

| State           | Meaning                                                  |
| --------------- | -------------------------------------------------------- |
| **To do**       | Not started                                              |
| **In progress** | Under way — for example an agent created but not yet run |
| **Done**        | Completed                                                |
| **Skipped**     | You chose to set it aside                                |

**Done** and **Skipped** both stop a step asking for attention, so the checklist finishes even if
some steps do not apply to you.

## Let the assistant do it

Several steps can be handed to the chat assistant as work rather than instructions. Where that is
possible, the step offers to open chat with the request already framed, and the assistant carries
it out with you watching. Anything that changes your workspace still runs under your own
permissions and is recorded like any other action.

## Finishing

Once every visible step is done or skipped, the header button stops showing a count. Nothing is
locked before then — the checklist is a guide, not a gate, and you can build, run, and govern
agents in any order you like.

## Next steps

<CardGroup cols={2}>
  <Card title="Build your first agent" href="/manage-your-agents/build/how-it-works">
    The hands-on version of the biggest step.
  </Card>

  <Card title="Integrations overview" href="/manage-your-agents/build/integrations-overview">
    Connect the first system.
  </Card>

  <Card title="Modules & workflows" href="/modules-and-workflows/overview">
    Switch on a prebuilt operational outcome.
  </Card>

  <Card title="Runs & evidence" href="/manage-your-agents/run/runs-evidence">
    Read what an agent actually did.
  </Card>
</CardGroup>


## Related topics

- [Production Readiness & Standards](/modules-and-workflows/ops-automation/production-readiness.md)
- [MCP server](/developer-tools/mcp-server.md)
- [Install](/get-started/on-prem/install.md)
- [Before you begin](/get-started/before-you-begin.md)
- [Model Evaluation](/manage-your-agents/optimize/model-evaluation.md)