Run a Test Case
Launch a single test case through an agent harness and watch it to completion.
This quickstart uses the CLI, the most direct path
for scripting and batch sweeps. You can also launch and watch runs interactively
in the Tauri desktop app or the
web console. All three enqueue the run at the
backend, which executes it server-side as a per-run
driver Job. For the full walkthrough,
prerequisites, and platform notes see First Time Setup.
Prerequisites
Section titled “Prerequisites”A working setup: a reachable backend (TCAB_BACKEND_URL) whose run queue an
in-cluster dispatcher is draining — for local
development, the k3d service stack brought up and
forwarded — plus a logged-in account (tcab login). tcab needs no container
runtime of its own; the cluster supplies the harness credentials to the run. See
First Time Setup if any of those are missing.
Run it
Section titled “Run it”tcab run \ --test-case carom --version v1.0.0 --variant base \ --harness claude --model claude-opus-4-8From a source checkout, substitute cargo run -p test-cabinet-cli -- run … for
tcab run ….
--variantis required: a run targets exactly one variant.--modelis passed to the harness unchanged; it is opaque to The Test Cabinet.--max-runtime <hours>overrides the case’smax_runtime_hoursfor this invocation only (fractional hours allowed, for example0.5).--out-dir runsis optional: it writes the fetched run record JSON locally (otherwise nothing is written — the backend holds the artifacts).
tcab run enqueues the run on the backend’s queue, prints the queued job id,
streams the live event stream as the driver executes
it (seeding a fresh repository, driving the harness in a sandbox pod, then
validating), and reads the produced
run record back to print its summary.
Inspect inputs without a run
Section titled “Inspect inputs without a run”tcab prompt --test-case carom --version v1.0.0 --variant base # the rendered prompttcab seed --test-case carom --version v1.0.0 --variant base # the seeded repo, on disktcab harnesses # harness availabilitySee the CLI overview for every subcommand.
Next steps
Section titled “Next steps”- Review a Run once it finishes.
- Reviewing Test Run Results for the full review workflow.