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

# Code Governance

> Review AI authorship signals in pull requests and see where they overlap with active security findings

Code Governance shows where ZeroPath detects AI authorship signals in pull requests. Use it to see which repositories have detectable AI-assisted changes, inspect the evidence behind an attribution, and review how those changes overlap with active security findings.

Open [**Code Governance**](https://zeropath.com/app/code-governance) under **Automation**. The page is in early access and requires **View Custom Reports**. Metrics cover the repositories you can view.

## Review a repository or period

1. Select one or more repositories and a date range. **All repositories** uses the repositories you can view.
2. Use the trend, **AI tools detected**, and **Confidence of attribution** to see where signals appear and how strong they are.
3. In **Recent pull requests**, choose **Attributed** for PRs with signals or **All captured** for the wider set. Check the **Confidence / verdict** and **Evidence** columns, then open the PR in your code host to review the change.

| Metric | How to read it |
| - | - |
| **AI share of opened PRs** | The share of captured PR records in the selected period with a recognized signal. A PR that touches multiple monorepo partitions can contribute a record to each partition. |
| **AI share of merged PRs** | The share of recorded merges in the selected period that carried an AI signal. Each PR counts once across monorepo partitions. |
| **AI share of merged lines** | The share of added lines from attributed GitHub PRs among merges with known diff sizes. A PR with a mix of contributions counts all its added lines. |
| **Findings in AI code** | The share of active findings from code scans linked through an attributed commit or originating PR. The date range applies to PR activity. This card uses the current finding state. |

## Read the attribution evidence

The confidence reflects the strongest signal ZeroPath recognized.

| Confidence | Evidence |
| - | - |
| **High** | A recognized bot identity or email |
| **Medium** | A declared co-author trailer or PR marker |
| **Low** | A heuristic match on author metadata |
| **No AI signal** | ZeroPath found no recognized signal in the captured data |

The **≥** prefix marks a lower bound based on observable signals. Inline IDE completions and work committed under a personal account without a marker can be absent from this view. A finding's attribution identifies where to look for its origin. Review the finding evidence and the code change when deciding how to respond.

ZeroPath collects supported pull request events from GitHub, GitLab, Bitbucket, and Azure DevOps. Merged-line volume depends on GitHub diff sizes. Older or missed events can leave gaps in the history.

To review security findings before changes merge, see [Pull Request Scanning](https://zeropath.com/docs/scanning/pr-scanning.md). To see which AI SDKs, models, and tooling your applications include, see [AI Inventory](https://zeropath.com/docs/sca/ai-inventory.md).


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.