Test Runner
Run request → bundle harness + tests → isolate execution → synchronized results
A built-in test harness runs in isolated runtimes and publishes structured result updates in real time.
In-browser test execution with a built-in describe/it/expect harness
Test Execution
Result Broadcasting
Notes
Tests use a built-in describe/it/expect harness — no Vitest or Jest dependency. Each test file is bundled with the harness and executed in an isolated WorkerLoader V8 sandbox (max 30s timeout). The AI agent's test_run tool shares the same runner. Partial re-runs merge results at the suite level via mergeTestRunResults().