EffortlessMetrics

RIPR finds what your tests run but don't actually check. Meta's agents fix those gaps, RIPR checks whether they're closed. I built RIPR because I needed cheaper, earlier evidence that tests actually constrain the behaviour being changed. It's a static mutation exposure analyzer that finds weaknesses in tests without needing to execute thousands of mutations. For the hackathon, we built a native Muse Code plugin that puts RIPR's findings directly into the coding loop. Muse Code can inspect a finding, repair the test, and get fresh RIPR feedback. We demonstrated the integration on a boundary-test repair. Muse Code improved the test, RIPR independently reassessed it, and an executed check confirmed that the repaired test rejected the incorrect implementation. We also tested an incorrect repair that looked useful but didn't actually distinguish the right behaviour from the wrong behaviour. That doesn't count as an improvement. Better tests, less wasted compute.