Test-run a task
Fire the task once as a test. The run does real research and produces real output, but every write-style tool refuses and every configured delivery is captured on the run instead of being sent. Poll the returned run, then read captured_deliveries for the exact message each channel would have received.
A test run never counts as a run of the task: it does not advance next_run_at, does not reset or trip the failure breaker, does not count toward autonomy graduation, and does not appear in the run history a later run reads. It does consume credits. 422 if the task already has a run in flight or the account is out of AI credits.
Authorizations
API key in the format: Bearer {api_token}
Headers
Makes this write safe to retry: a repeat with the same key replays the first response instead of creating a second resource. Scoped to the account, remembered for 24 hours; reusing a key with a different payload is an error.
255Path Parameters
Task id (task_...)
Response
Accepted — the test run executes in the background; poll GET /task_runs/{id}.
pending, running, completed, failed, needs_input, timed_out, unknown 
