zlacker

I bet it's pretty easy. Capture your own mouse movements from one place to the next as denoted by clicks. Then train a model on reproducing those movements, using your captured data of movement from points A to B. It would probably generalize well enough to pass the verifications. Humans are very unpredictable, so I assume those are mostly looking for superhuman speed and accuracy.

>>helloj+(OP)
https://github.com/vincentbavitz/bezmouse

> BezMouse is a lightweight tool written in Python to simulate human-like mouse movements with Bézier curves. Some applications might include:

> BezMouse was originally written for a RuneScape color bot and has never triggered macro detection in over 400 hours of continuous use.

:)