Repository adoption scorecard
Pin the exact revision, write down the intended use, and retain evidence for every score before starting a bounded pilot.
Save before closing or reloading. Drafts stay in this browser on this device; they are not uploaded or shared across devices. Avoid saving private details on shared computers.
A high score is not a safety verdict
Unknowns should remain unknown. Automated checks, repository activity, and documentation cannot prove secure code or production fitness.
Pin the decision context
A result without a target revision and intended use cannot be reproduced or compared later.
Score only recorded evidence
Use N/A only with a written reason; excluded checks always require manual review. A blank evidence note leaves the criterion incomplete.
License and redistribution
Can the intended use, modification, model or asset license, and redistribution boundary be identified?
Release and maintenance path
Do releases, changelogs, issue handling, and upgrade notes support the proposed maintenance window?
Reproducible installation
Can a clean environment reproduce the documented install with versions and artifacts pinned?
Permissions and secrets
Are tokens, filesystem access, network destinations, tool capabilities, and denial behavior bounded?
Exit and rollback
Can data, configuration, and workflows be exported or restored after a failed pilot or upgrade?
Maintainer and issue signals
Is maintenance distributed enough, and are high-impact issues triaged in a way the pilot can tolerate?
Waiting for evidence
- Complete criteria
- 0/6
- Points
- Waiting
- Normalized score
- Waiting
Any zero rating blocks a pilot recommendation. A complete context, executed installation, permission-denial check, and exercised rollback are required even when the average is high.
Files are created in this browser. This page does not upload repository names, evidence notes, or scores.