Right now the front-matter only specifies needing "model" which corresponds to the "Deployment Model" in Foundry (MAF). We should update it to accept an Endpoint option too.
Our runner currently specifies multiple options for agent harness in MAF - "openai", "foundry", "azure-openai". We should consider adding top-level models for agent-config (endpoint, model, temperature, etc) so that its a single place of truth for the customer.
Right now the front-matter only specifies needing "model" which corresponds to the "Deployment Model" in Foundry (MAF). We should update it to accept an Endpoint option too.
Our runner currently specifies multiple options for agent harness in MAF - "openai", "foundry", "azure-openai". We should consider adding top-level models for agent-config (endpoint, model, temperature, etc) so that its a single place of truth for the customer.