Within a ClearML task, is it possible to checkout another task code ?

Context:
I have a bit of a special use-case: we train our model using ClearML. Our code prep the data: download images (we are not using clearml dataset) and bundle labels into a dataframe (df) then run training on it (computer vision model). We did not have inference code in the repo in the past.

I want now to be able to automatically run inference with the same data that been trained in the past. My plan is:

1. start a ClearML task with another ClearML ID as parameter.
2. in the task, given a ClearML ID, somehow "checkout" the corresponding code
3. with the new python code, from current task, in a new file that does not exist in the checkout code, run the step that download image and bundle label to df. But instead of running the training code, this time do inference. Get the results and upload it as artifact to current task.

Step 2 is the one I am wondering if there are any existing API/function for that ? Or can you suggest me how can I hack/build one ?