Adding example for OpenAI compliant Agent URL

[matzew]

Run this, with:

# Set your model credentials
export MODEL_BASE_URL='https://your-api-endpoint.com/v1'
export MODEL_KEY='your-api-key'
export MODEL_NAME='your-model-name'

# Run the test
./gevals run examples/kube-mcp-server/openai-agent/eval.yaml

[Cali0707]

@matzew I think the best way to use the eval framework in this case would be to have the same task files and just change the agent/eval. So, instead of copying the tasks we would just leave them. WDYT?

[matzew]

I think the best way to use the eval framework in this case would be to have the same task files and just change the agent/eval. So, instead of copying the tasks we would just leave them. WDYT?

yeah, that sounds good - let me fiddle

[Cali0707]

LGTM - this works for me!

Thanks @matzew