Ranking from our own tests
Only projects we have run ourselves enter, with their test published in this repository. Anything merely read does not score: it is listed below as pending.
| # | Project | Cases passed | Coverage | Problems found | Date | Test |
|---|---|---|---|---|---|---|
| 1 | suraj-phanindra/wellposed | 18/18 | 58% (7/12) | 3 | 2026-10-31 | run.json |
| 2 | RiskAverseTech/toolgate | 4/4 | 57% (4/7) | none | 2026-10-30 | run.json |
| 3 | ursuciprian/reflex | 28/28 | 54% (7/13) | 3 | 2026-10-30 | run.json |
| 4 | Koushik890/jev-firewall | 15/15 | 50% (5/10) | 3 | 2026-10-30 | run.json |
| 5 | ismaelsoilet/jev-harness | 18/21 | 54% (7/13) | 7 | 2026-10-31 | run.json |
120 assessed projects still lack a test of our own and do not score.
How it is ordered
- First, fewer failed cases. Then, more coverage: what was measured versus what was declared unmeasured. On a tie, more cases exercised.
- It is not a quality grade. We write the cases ourselves for each output contract, so it compares how much of each project has been checked, not which one is better.
- A test with every case green can still leave the important part out: the coverage column says so, and the detail of what was not measured is in the linked test.