When building out your tests in the cloud, the mabl agent now replays steps to confirm they execute correctly and self-corrects test steps when something doesn’t work. With agent self-correction, you can spend less time manually cleaning up agent-generated tests.
These improvements come from two key changes to how the agent creates tests:
- Step replay and self-correction: The agent can now replay steps it has already created to verify they still work. If a step fails during replay, the agent undoes it and retries with a different approach, iterating multiple times if needed before moving on.
- Updated testing best practices: The agent follows upgraded guidance for writing reliable mabl tests, producing better assertions, cleaner step organization, and tests that don’t require manual cleanup after creation.
No action is required. These improvements apply automatically to cloud-based agentic test authoring.