Add AgentOps support for Microsoft Agent Framework evaluations #41

Objective

Add support in AgentOps Toolkit for evaluation workflows based on Microsoft Agent Framework.

Scope

Enable AgentOps to work with agent and workflow evaluations (including multi-turn and tool-aware scenarios) defined by the framework, while remaining provider-agnostic.

Tasks

• Map Agent Framework concepts to AgentOps (agents, workflows, conversations, tool calls)
• Define how run.yaml represents these evaluation scenarios
• Design input handling for conversations, expected outputs, and tool calls
• Define normalized output structure for results and reporting
• Identify minimal integration points (single agent, workflows, no re-run scenarios)

Acceptance Criteria

• High-level integration design is defined
• Supports agent and workflow evaluation scenarios
• Maintains provider-agnostic execution model
• Output format is defined for CI/reporting

References:

https://github.com/microsoft/agent-framework/blob/main/docs/decisions/0020-foundry-evals-integration.md
https://github.com/microsoft/agent-framework/tree/main/python/samples/05-end-to-end/workflow_evaluation