Summary
When using parallel execution for grouped blocks in MindStudio, the first Generate Text block does not allow mock data configuration from a previous run, while subsequent blocks do.
Steps to Reproduce
- Create a new AI Agent.
- Add a Generate Text block:
- Prompt: “Pick a color.”
- Save output to variable:
color
- Add another Generate Text block:
- Prompt: “Pick a number.”
- Save output to variable:
number
- Group both Generate Text blocks.
- Set the group execution mode to Parallel.
- Run the agent in Draft Mode.
- After execution:
- Select the first Generate Text block.
- Open Debugger → Configure Mock Data.
- Select the previous run from the list.
Actual Behavior
- The system displays the message: “Block was not executed in this run.”
- This prevents configuring mock data for the first Generate Text block.
- Interestingly, the second Generate Text block works as expected, allowing mock data configuration using the same previous run.
Expected Behavior
- Both Generate Text blocks (regardless of order in the parallel group) should:
- Be recognized as executed in the selected run.
- Allow mock data configuration from that run in the Debugger.
- Parallel execution should not affect the ability to access or reuse execution data for debugging.
Impact
- Prevents consistent debugging workflows when using parallel execution.
- Creates confusion since identical blocks behave differently.
- Slows down development and testing, especially for more complex agents using parallel logic.
Additional Notes
- The issue appears to be order-dependent within parallel groups.
- May indicate a problem with how execution metadata is stored or referenced for parallelized blocks.