Feature hasn't been suggested before.
Describe the enhancement you want to request
I spent the last week working on improvements of the agentic orchestrator, and was able to find major improvements, severely limiting (compared to Claude code 2.1.92, at the very least by a 10x magnitude) the amount of tokens consumed. I am currently working on implementing my changes in cheetahclaws, but would also be very pleased to contribute to opencode afterwards. Here is a link to my repository, the readme covers the reasoning behind the main ideas: https://github.com/Simon-Free/bouzecode. Here is another discussion i had where some further clarifications were made: SafeRL-Lab/cheetahclaws#43. I would be very happy to answer any other questions !
Feature hasn't been suggested before.
Describe the enhancement you want to request
I spent the last week working on improvements of the agentic orchestrator, and was able to find major improvements, severely limiting (compared to Claude code 2.1.92, at the very least by a 10x magnitude) the amount of tokens consumed. I am currently working on implementing my changes in cheetahclaws, but would also be very pleased to contribute to opencode afterwards. Here is a link to my repository, the readme covers the reasoning behind the main ideas: https://github.com/Simon-Free/bouzecode. Here is another discussion i had where some further clarifications were made: SafeRL-Lab/cheetahclaws#43. I would be very happy to answer any other questions !