Alex Hancock
Alex Hancock
@codefromthecrypt of course - let us know whenever ready for a review! It's a big feature and will be huge to get it right
> it should just be integrated directly into the MCP server overview can you clarify the suggestion here @DOsinga?
I think there are (at least) two questions to consider 1. Should the registry do any form of scanning/sanitizing of the input if we do choose to enable the `template`...
@katzdave The failure was actually unrelated to the model here. I think this should be more reliable. If it proves not to be once in main + everyone's PRs we...
@jamadeo @DOsinga Updated it to always uses the agent's current provider via a shared reference. Verified by the `llm_request` logs that it uses different models for sampling after switching providers.
Draft of a couple different approaches to integration testing for sampling https://github.com/block/goose/pull/5456 Wanted to make a separate PR to get feedback on which one seems most valuable @jamadeo @DOsinga
Thanks for the work on this. Going to close this out per https://github.com/block/goose/pull/4834#issuecomment-3418837151 being three weeks ago, but welcome to re-open a new PR.
https://github.com/block/goose/issues/5270 will resolve this. Looking into it
Sounds good @DOsinga @katzdave I don't believe they have a machine readable public spec for the Responses API but I did get some advice from an OpenAI engineer on this...
@kgeis I am also curious for the answer's to @jamadeo's Qs above. Also, is there any MCP server that repros this without setting up a jenkins instance?