Paul Stack on Mythryl: “We test our code but we also test the thing users download. Separate…”

We test our code but we also test the thing users download. Separate repo, real compiled binary, CLI tests plus adversarial security and benchmarks. Three potential regressions that were caught in minutes before any user saw them. The next part of the machine: stack72.dev/the-gate-bet...

stack72.devThe Gate Between Our Agent Code and Our UsersEvery release binary gets dispatched to a separate repo running CLI tests, adversarial security tests, and benchmarks. Three recent bugs — all with full unit test suites passing — never reached a user...

21/4/26, 5:35 pm0 reposts1 like

Post