My goal with the project is to test different agentic coding tools in a more organized setup.
The process will be:
- First, learn to use the tool, spend at most 1 hours learning about the tools and the best ways of using it
- Create a new repository from the template in bamorim/expensify-template
- Make sure I have the tool installed
- Initialize the tool (some tools provide with some sort of "init" process)
- Convert what's in the
.aidirectory to the tool specific configration (that is, creating a instructions file and mcp stuff) - Prompt the tool with the prompt below to get us started
- After that I might use custom prompts to fix issues or use some of the auxiliary prompts to get back on track
- I want to spend around 2 hours per tool on this phase
I keep adding tools to this list as I find them and I update the project links when I finish reviewing them.
Some of these tools can be used in multiple ways. For example, cursor started as it's own IDE but now there is Cursor CLI that you can even integrate to VSCode or your Jetbrains IDE. In the table below I try to classify the "main" way (and the one the creators try to sell as "the way to use it") it is used and it is how I tested. If you would want me to review a different variation of a specific tool, please let me know.
Let's develop an application for expense management for companies.
I wrote some PRDs in @docs/product/, but for now let's focus on on the @docs/product/prd-main.md
You should first plan the tasks we need to accomplish them and keep track of them in the @docs/tasks/
After that we can proceed on completing the tasks
Let's get back to work on coding this application, resume tasks in @docs/tasks/ while keeping in
mind to check @docs/product/ if the task goals are unclear. Remember that we are just focusing on @docs/product/prd-main.md
If needed, create more tasks to keep going.