Argusic: Does Real Software Run?

Field notes

The archive evaluates repositories from a pinned commit inside a disposable machine, asking an agent to install, launch, debug, and verify each project without human assistance. Public run records distinguish pass_full, pass_mock, and fail outcomes and attach numerical scores, including a 36.67 result for Nest alongside multiple 100.00 results. The homepage reports five subjects and twelve recorded runs, while linking repository metadata to GitHub and licensing its data under CC BY 4.0.