I've been testing the agent in a pretty large project (12 CMS) 70+ pages and 200 components.

Credit usage is reasonable with 200+ credits used per prompts.

The agent struggles a lot when asked to batch edit items. I asked it to reorganize the project components into clear folders but the agent failed to do so. I also asked it to think of a good naming convention for the components and to rename them accordingly to the naming convention it chose. It failed again.

It would be nice if the agent had a larger context window (even if it means consuming more credits) for it to be useful when doing batch edits on large projects (which is the only use case i've found for the agent personally yet).

Did some of you ran into the same problem ?

I've been testing the agent in a pretty large project (12 CMS) 70+ pages and 200 components.

Credit usage is reasonable with 200+ credits used per prompts.

The agent struggles a lot when asked to batch edit items. I asked it to reorganize the project components into clear folders but the agent failed to do so. I also asked it to think of a good naming convention for the components and to rename them accordingly to the naming convention it chose. It failed again.

It would be nice if the agent had a larger context window (even if it means consuming more credits) for it to be useful when doing batch edits on large projects (which is the only use case i've found for the agent personally yet).

Did some of you ran into the same problem ?