Summary:

• We need a dev setup which works with MCP tools to iterate on the AI assisted workflows.
• The lab resources available have API request limits (6 requests per minute) making development there unfeasible: https://developer.models.corp.redhat.com
• We have no unrestricted LLM resources which work with lightspeed-stack. Many teams are using OpenAI or Google vertext AI.
• Many of us do have Gemini API keys.

Goal:

• Look into getting lightspeed-stack to support gemini via the responses API which can be used to iterate on goose workflows for AI assisted workflows.
• https://github.com/dprince/openstack-lscore-deploy

TimeBox:

• less than a week

NOTE: Gemini likely isn't a realistic customer environment but we have no other viable resources at the moment.