Commercial-Safe AI Stack
Permissive-license model, dataset, app, and eval records to review first for product work.
Audience: Commercial AI teams
- Filter to permissive licenses
- Inspect trust bundles and source policy
- Prefer scan-passed hosted files
- Record review gaps with reports or mirror requests
How to use this pack
This pack narrows the catalog to records with permissive licenses, such as MIT or Apache 2.0, so a product team has a shorter list to review first. A permissive label is a starting point. Read the license text on the record, check the trust bundle and the source policy, and prefer files that passed scans and are hosted with a SHA-256 hash. If something looks wrong, file a report or a mirror request from the record page.