code reviewer question

[main depot]

Are there strategies or methodologies you recommend for using AutoGen in conjunction with external testing and debugging tools, especially for non-Python code?

What approaches could be taken to make AutoGen more adaptable or flexible in understanding and evaluating diverse coding practices, especially in a multi-language development environment?

Can you suggest any architectural or infrastructural changes that could enhance AutoGen's scalability and efficiency in processing and reviewing extensive codebases?

[silent pine]

All great questions!

[main depot]

@silent pine trying to use autogen framework for my code escrow platform. Would love to automate the review process of the code once the dev submits it. You can’t any suggestions.

[silent pine]

Afraid this application is beyond my expertise. But autogen's multi-agent capabilities seem useful for complex automation. Hopefully members of the community here can give more guidance.

[static scaffold]

normally you want to simulate gcc compiler and a shell , then pass those replies to the original agent, then based on that (if there are errors) continue to code/improve or break the coding loop/task and return

[main depot]

@static scaffold see I’m not even trying to fix the code itself. Just review it if there r errors. If there’s errors I want it to return “there’s an error with x” or “works great!”It doesn’t need to even provide the proper code replacement.