How Penny is tested
Every change to Penny ships through a pull request. TestSprite runs the test suite against that pull request's own preview, and GitHub refuses to merge it until the check passes.
When TestSprite fails one of these pull requests, a locked-down job asks Claude Code for a fix, a script with no AI in it checks the change before it is pushed, and TestSprite tests the fix again. A person reads and merges every fixed pull request.
Pull requests titled "Study" are planted bugs from a catch-rate study: each adds one deliberate bug, TestSprite tests it, and it is closed unmerged. The results are in the docs.
Reading the latest checks...