we have a similar conditional WF that has similar logic as yours and it is working

With is_true() and with changing my task/dynamic to a subflow, it works very well! It still requires me to use an “else”, otherwise the conditional seems to be ignored, but the “else” can be a minimal dummy task that immediately exits, and as soon as the conditional’s result is decided it expands the graph to show my subflow tasks. Working example (where

shape_fit_flow

is a sub @workflow):

conditional("should_run_fit")
        .if_(run_fit.is_true())
        .then(shape_fit_flow(
            fitting_repo=fitting_repo,
            core_algo_repo=core_algo_repo,
            smplx_version=smplx_version,
            debug_fit=debug_fit,
            shape_input_path=parsed_session_info.shape_input_path,
            scan_bucket=parsed_session_info.scan_bucket,
            is_ipad=parsed_session_info.is_ipad
        ).else_().then(dummy_task())

Broken example (I removed the else then with the dummy task on the last line):

conditional("should_run_fit")
        .if_(run_fit.is_true())
        .then(shape_fit_flow(
            fitting_repo=fitting_repo,
            core_algo_repo=core_algo_repo,
            smplx_version=smplx_version,
            debug_fit=debug_fit,
            shape_input_path=parsed_session_info.shape_input_path,
            scan_bucket=parsed_session_info.scan_bucket,
            is_ipad=parsed_session_info.is_ipad
        )

The docs didn’t show an example without else/then, and it’s not working for me without it, so I’ll assume it’s not supported. Not too much of a problem, because using dummy task is an acceptable workaround in my case. Thanks!

yeah probably not supported, but i think it works for dynamic task though