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

# AI Agent Harness

> Configure the context, tools, guardrails, runtime, and evaluations that make an AI agent dependable.

## What the harness controls

An AI agent harness is the operational layer around a model. It determines what an agent knows, what it can do, where it runs, and how you evaluate its results.

| Layer                | Purpose                                                               |
| -------------------- | --------------------------------------------------------------------- |
| Context              | Scope repositories, tickets, decisions, and domain rules for each run |
| Tools                | Expose typed actions with explicit permissions                        |
| Authority guardrails | Reserve sensitive or destructive decisions for people                 |
| Runtime              | Provide isolated, reproducible execution and an audit trail           |
| Evaluations          | Measure task outcome, boundary adherence, cost, and latency           |

## Configure an agent

1. Select the context sources the agent needs.
2. Grant only the tools required for its task.
3. Define actions that require approval or remain blocked.
4. Choose a model based on capability, latency, cost, security, and residency.
5. Create evaluation cases and acceptance criteria.
6. Review run outcomes before expanding access.

<Warning>
  Prompt instructions are not a security boundary. Enforce permissions and approval requirements in the runtime and tool layer.
</Warning>

## Model-agnostic operation

Keep context, tools, policies, and evaluations outside the model. This lets you compare or replace model providers without rebuilding the operational controls around an agent.

## Related

* [AI usage boundaries and controls](/docs/products/agents/ai-controls)
* [MCP Gateway](/docs/platform/mcp-gateway)
* [AI Agent Harness capability](https://www.scrums.com/platform/agent-harness)