SPLIT interest. Past log completion. Better list handling. I'm going this now it and it works better. (also helpful for failing agents)

[bbarclay]

What problem or use case are you trying to solve?
Agent doesn't have task list to work on, or reference back to. In the even it crashes. Or other agents want to work on the task at the same time.

• Currently, I'm opening multiple browsers to have threaded agents working on app conversion.

Describe the UX of the solution you'd like

• When the agent is prompted, It creates a folder of tasks and subtasks.
• What can is the general prompt for this task?
• What are the subtasks that can be distributed at the same time, or if they can't label them.
• Create a directory of these, so they can be checked out by multiple agents. To work on them at the same time.
• Smaller tasks and more of them, = Smaller context windows.

Do you have thoughts on the technical implementation?

Basically before I prompt anything now. I say, here's the tasks, create a folder with a list, or multiple lists, of what the tasks is, so the next agent when you fail at 50,000 characters another agent can pick up on what you haven't finished. Put a check in the list. 1 meaning it's being worked on, and 2 meaning you finished it.

• AI will then write a prompt. and these tasks to complete.
• Then I can spin up multiple browsers and tell it to work on the same thing. In the case that it can work on this simultaneously.

Describe alternatives you've considered
I'm using the alternative suggested above to solve this. But what happens, is when the agents fail. the majority of the time. They have no point of reference where to start up again.

• Even during their process of working. They could be creating more subtasks. LIke. I finished this, but I recommend testing.

So they can add to another lists of @todo's .

If the lists isn't completed, or a test hasn't passed, or something. Then the agents don't move on, and you keep deploying them, or tell them to move on from that task.

Additional context

In today's ever evolving world of technology, we delve into. JK. JK. Additional context: The additional context is this. When the application is going through it's processes, having a list of tasks, and subtasks, and being able to deploy threaded agents towards these tasks is super helpful.

[github-actions]

This issue is stale because it has been open for 30 days with no activity. Remove stale label or comment or this will be closed in 7 days.

[github-actions]

This issue was closed because it has been stalled for over 30 days with no activity.