> ## Documentation Index
> Fetch the complete documentation index at: https://agent-compass.mintlify.site/llms.txt
> Use this file to discover all available pages before exploring further.

# Configure an Evaluation

Understand the parts of an evaluation request and find the authoritative documentation for each setting.

Use this page after completing the [Quick Start](/en/get_started/quick_start) to decide whether to adjust the model,
benchmark, harness, environment, or [run-level controls](#run-level-controls). See [`agentcompass run`](/en/user_guide/using_agentcompass/cli/run#parameter-reference)
for complete CLI parameters and defaults, and [`agentcompass config`](/en/user_guide/using_agentcompass/cli/config) for
configuration files and precedence.

This page describes one `agentcompass run` request. To coordinate multiple models, benchmarks, harnesses, or
environments with one scheduler, use [`agentcompass launch`](/en/user_guide/using_agentcompass/cli/launch).

## Evaluation Structure

Every run selects four core components:

```bash theme={"system"}
agentcompass run <benchmark> <harness> <model> --env <environment>
```

Each part owns a different set of decisions:

| Part                                                        | What it controls                                                                                                                                                                  | Where to configure it                              |
| ----------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | -------------------------------------------------- |
| [Model](/en/user_guide/modules/models/overview)             | Model ID, API endpoint, credentials, protocol, and inference parameters                                                                                                           | Third positional argument and `--model-*` flags    |
| [Benchmark](/en/user_guide/modules/benchmarks/overview)     | Dataset, task selection, task preparation, scoring, and aggregate metrics                                                                                                         | First positional argument and `--benchmark-params` |
| [Harness](/en/user_guide/modules/harnesses/overview)        | Agent loop, framework integration, tool behavior, and harness execution limits                                                                                                    | Second positional argument and `--harness-params`  |
| [Environment](/en/user_guide/modules/environments/overview) | The environment provider—the backend that creates and manages task environments, such as Docker, Daytona, or Modal—plus workspace behavior, network policy, images, and resources | `--env` and `--env-params`                         |

<Info>
  Keep parameters with the component that owns the behavior. For example, task filters belong in `--benchmark-params`, agent step limits belong in `--harness-params`, and container resources belong in `--env-params`.
</Info>

## Run-Level Controls

An evaluation request also contains controls that do not belong to any one of the four core components, including concurrency, timeouts, retries, result naming, historical run reuse, environment retention, and logging. See [Run Controls](/en/user_guide/using_agentcompass/run_controls) for how these settings work together, and the [`agentcompass run` parameter reference](/en/user_guide/using_agentcompass/cli/run#parameter-reference) for complete signatures and defaults.

## Next Actions

<CardGroup cols={2}>
  <Card title="Adjust Run Controls" icon="sliders-horizontal" href="/en/user_guide/using_agentcompass/run_controls">
    Configure concurrency, timeouts, retries, output reuse, debugging state, and logging.
  </Card>

  <Card title="Run Multiple Evaluations" icon="list-tree" href="/en/user_guide/using_agentcompass/cli/launch">
    Coordinate requests that use different models, benchmarks, harnesses, or environments with one global scheduler.
  </Card>

  <Card title="Manage Configuration" icon="file-cog" href="/en/user_guide/using_agentcompass/cli/config">
    Inspect merged file values and component fields, and understand configuration-file loading order.
  </Card>

  <Card title="Inspect Results" icon="chart-no-axes-combined" href="/en/user_guide/other_features/results">
    Read per-task details and summaries, preserve reusable evidence, and analyze failures.
  </Card>

  <Card title="Troubleshoot a Run" icon="wrench" href="/en/user_guide/other_features/troubleshooting">
    Reduce a failure to one task, locate its lifecycle phase, and apply the correct fix.
  </Card>
</CardGroup>