My way to solve some issues

Google Antigravity
,
1

Hi there.

Sometimes I run into the problems that no model can solve.
And the reason I decided to write this post is that I realized that every time I have these situations, the only way I solve the problem is by making the agent scan the entire app and find files/codes that are blocking the fix or change. I basically told the agent, “Scan the entire app, all folders and files, and find the code that is blocking your changes.”
And some issues that take you hours to solve will be solved in minutes, or even seconds.
I realized that all models are missing this approach for some reason.
They are trying to solve the problem by writing new codes, but they are never really scanning the app for what’s blocking their fix or change.
I don’t know; maybe I am saying something stupid, but it works for me all the time.
I hope that my post is going to help someone.

Have a great day. :relieved_face:

1 Like
2

Great insight. What you are describing is fundamentally Prompt Engineering. By explicitly instructing the agent to ‘scan’ first, you are effectively overriding its default bias towards immediate generation and enforcing a Chain of Thought that prioritizes global context analysis and dependency checking before implementation. You are guiding the model’s reasoning process rather than just requesting a raw output.

2 Likes
3

Thank you for sharing insights! It is always valuable to see the creative prompting strategies that help unblock difficult tasks.
Explicitly instructing the agent to ‘scan the entire app for blocking code’ is a great strategy. It forces the model to utilize its context window to identify dependencies or conflicting logic that might be preventing the fix, rather than just attempting to overwrite them.