RFE: Provide a mechanism to opt-in or opt-out of setting a standard User-Agent header for all LLM API calls made by cecli

[hrivera-ntap]

Issue

We have a use case where we would like to track CECLI usage and being able to monitor via the User-Agent header seems to be how other AI Coding agent tools do it.

This is an RFE to have cecli identify itself using the User-Agent field in a similar manner.

ref: https://en.wikipedia.org/wiki/User-Agent_header

Version and model info

No response

[dwash96]

Neat, this makes sense to me, I'll whip something up for 0.99.8

[dwash96]

Ah I added a user agent in 0.99.8 but I see that you requested the behavior to be toggle-able. It technically is on a per-model basis if you wanted to use the model-overrides key to set the header explicitly for the models in question but not globally on/off

[hrivera-ntap]

Great. What you have should be sufficient for my use case, since i was planning on having it on all the time.