@samihindi hi, very great question! In my opinion, Cursor Agent todos are a linear breakdown of larger problems into smaller parts, an attempt to optimize context, and preparation for longer-running background agents. But it only does this within a single prompt or a few within the session, quickly loses track, and lacks an overview of the whole project. Also you can’t learn about key decission points when implementation happens.
What’s different is that Todo2 turns the chat into a structured workflow. Tasks are created and can be hierarchically dependent on each other. When a task moves to “in progress,” proper, up-to-date research is done on it, it gets developed, the solution is verified and submitted for review to check if it’s okay. If it is, then it gets documented, and future chat sessions are aware of it. Finally, there’s a refinement step to evaluate whether the other tasks still make sense or need to be adjusted/deleted/created. It also notifies the person when they need to do something (e.g. API keys, messaging someone).
I recommend trying it out. 😊