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

# Earnie for AI Agents

> Check your own work before handing code back: a deterministic verdict, grounded in evidence a human can inspect.

If you're a coding agent reading this, this page is for you.

## Check Your Own Work Before Handing Code Back

Call [`earnie mcp review`](/en/latest/earnie/using-earnie/earnie-cli#review-hooks-and-self-checks) against your staged changes, a diff, or specific files:

```bash theme={null}
earnie mcp review --staged
earnie mcp review --diff <base>
earnie mcp review <file> [<file>...]
```

This is a [Self-Check](/en/latest/earnie/getting-started/first-scan#self-checks-the-dry-run-before-the-gate), not a Scan, so it costs nothing and moves no posture. If a project has end-of-turn review hooks enabled, this already happens automatically when you reach the end of a turn, and a blocking result gives you one correction turn before the review ends.

## What You May Assume

* **Same input, same answer, every time.** Findings are deterministic, built from what the code actually matches, not model guesswork.
* **Every verdict traces to evidence.** A human can inspect the match, side by side, line by line, in the [Review Workspace](/en/latest/earnie/using-earnie/triaging-findings).
* **Read-only policy commands** ([`earnie policy list/show/violations`](/en/latest/earnie/using-earnie/earnie-cli#policies-read-only)) let you check what a project actually enforces before proposing a change, rather than guessing.

If a turn genuinely doesn't need review, `EARNIE_SKIP_REVIEW=1` records that as a visible, deliberate skip instead of silently omitting the call.

## What's Next

Read the [Review Hooks and Self-Checks](/en/latest/earnie/using-earnie/earnie-cli#review-hooks-and-self-checks) section for the full setup, or [Earnie for Developers](/en/latest/earnie/by-role/developers) for the human side of the same workflow.
