Daytona
Give every build a clean start

Run isolated development environments, repeatable test jobs, and resource follow-through with Daytona and SPIRITT. Create sandboxes from deliberate configurations and verify readiness before executing work. Run commands, collect outputs, and carry failures into an owned engineering queue.

Yes. It can provision a sandbox, clone the selected repository, execute test or build commands, and download the output files. It ties failures to reproducible evidence in your engineering tracker and checks the next run before reporting a fix verified.

SPIRITT uses current sandbox, command, snapshot, and volume operations. It checks build and backup state before declaring completion, keeps important outputs outside disposable environments, and applies your resource and access rules.

For a GitHub change, SPIRITT can provision a Daytona sandbox, clone the repository, execute the agreed tests, and attach result files to the engineering review. Failed runs retain the reproduction command and source revision; SPIRITT reruns the corrected code and stops the environment after collecting its outputs.

SPIRITT can schedule repeatable test runs and separately inspect the sandbox fleet for idle environments. It compares sandbox labels and resource usage with your owner registry, applies approved idle-stop settings, and verifies the stopped state. Unowned environments and reproducible test failures stay in assigned engineering queues.

Open a SPIRITT workspace, connect Daytona, and hand over a concrete mission such as: build an isolated test runner. SPIRITT builds the workflow, connects the other tools it needs, and operates it around your rules.