Failure Lab

Atomic Flow Failure Lab is envisioned as a testing-focused companion to the Core runtime.

Its purpose is to help developers answer a question that happy-path unit tests often miss: what state is the application left in when something fails halfway through?

Intended testing scenarios

Inject failures before, during, or after a workflow step.
Exercise recovery paths and report their outcomes.
Check developer-defined invariants after interrupted execution.
Reproduce partial successes and uncertain external outcomes.
Identify regressions when execution behavior changes.

Runtime separation

Failure Lab is not intended to be a mandatory production dependency. The Core should not require test-only failure-injection machinery in a production application.

See Failures and recovery for the conceptual background.