AUTOGOVERN / CONNECTED EVIDENCE

Discovery & Integrations

Discover declared AI dependencies. Bring traceable repository and CI evidence into your customer reviews.

Before you begin

1. Choose where evidence belongs

For a one-time scan, select a saved assessment or use Create a private assessment below. A registered AI system is not required.

2. Prepare repository access

Have the GitHub URL or owner/repository ready. Private repositories need a fine-grained token with Contents: read. Add Actions: read only if collecting CI metadata. Re-enter the token for each scan; it is cleared after submission.

3. For continuous sync

Sign in as an organization owner or admin, save the assessment in that organization, and register an AI system in the Workbench ↗. Return here and click Refresh sync status. Choose a schedule and authorize recurring reads; private tokens are stored encrypted when configured.

Review collected metadata in Private scan history before importing it. Customer sharing requires separate evidence-room approval.

GitHub · Read-only repository scan

Reads package.json, requirements.txt and pyproject.toml at the repository root. Optional CI collection reads up to 20 workflow runs for the captured commit. It does not request application source, secret files, logs or artifacts, or modify GitHub. Only extracted metadata is retained; raw manifests are not stored.

Loading saved assessments…

Create a private assessment

Create a named record to hold repository evidence. It starts without a risk rating or score; complete the assessment in the Workbench later.

Private scan history

Review each scan and select which items to import as private evidence.

Select an assessment above to view its scan history.

Continuous GitHub connections

Keep repository evidence and discovered dependency records up to date. Organization admins choose a system and authorize recurring reads. New evidence is collected privately and still requires review before customer disclosure.

Choose an assessment above.

Manage registered systems ↗