What it is
We prepare tasks where the model must find and fix a weakness in your code, such as an injection, a missing permission check or unsafe file handling. Each task has a test that tries the weakness; the fix must stop it without breaking your other tests.
What you give us
Your repository, how to run its tests, and the kinds of weakness you care about most.
How a run is checked
A run passes when the test that tries the weakness no longer succeeds and your existing tests pass.
Request it
We prepare this eval from your repository; there is nothing to set up yourself. Request it from Prepare eval in the BoltzLabs desktop app: choose Security fixes, add your code and say what the eval should measure. We reply there when it is ready, and you then run it in Evals like any other eval.