This ground truth study evaluated sample of 150 software repositories from
InnovationSprint2025
and a small sample from openAIRE from the original 1500 repositories.
All repositories contain a codemeta.json file and were examined across
multiple metadata sources including package manifests (e.g., setup.py,
package.json, DESCRIPTION), citation files
(CITATION.cff), licensing files, README documentation, and other
structured or semi-structured metadata artifacts.
| Repository | Commit ID | Pitfall / Warning Code | Description | Source File |
|---|