This is an example of a "Task List" thunk pattern.
As the name suggests, this kind of thunk doesn't really have much upfront planning. Each task is heterogenous and different. You add tasks to it and your AI agent starts working on them. It's as simple as that.
Here is a link to the sample and here's a video describing how the sample works.
Creation
The goal could be completely generic (eg: "Manage my work list") or something more specific (eg: "Plan a product launch"). In all cases, the plan is trivial -- it simply tells your AI agent to do each task based on the description of the task.
The meaningful part of the thunk therefore is in the project data. Each entry corresponds to a task to do and its description.
Execution
In this thunk, the AI agent is involved in at least the following activities:
The generation of appropriate tasks based on the goal
The execution of individual tasks. These tasks could involve searching the web, reading or creating files, searching for content in images, or many other things that your AI agent is capable of doing or reasoning about.
There are many variations of the "task list" thunk pattern as well. You can:
Import tasks automatically from email
Add data validation checking (eg: to ensure that inputs or results are sensible)
Add AI formulas (eg: to automatically deduce priorities)
Incorporate team policies to guide the work done
Interact with your AI agent on each task to provide an arbitrary level of detail or guidance
How to use it
There are three ways to use this sample:
Browse the sample and explore what it does
Make a copy of the sample so that you can interact with it, add your own tasks, etc
Create a new thunk, give it your own goal, choose the Task List category, and apply what you have learned from the sample