+[description]: This release introduces agent evaluation support across all declarative agent (DA) project templates. Each template now includes an `evals/prompts.json` file with a sample dataset to assist users in evaluating their agents effectively. Additionally, the associated README files have been updated with a new "Evaluating Agents" section that includes instructions for using the `@microsoft/m365-copilot-eval` CLI to perform evaluations. Environment variable placeholders have also been added for easier configuration. This enhancement provides a clear, standardized way to test and assess agent performance directly within the project structure.
0 commit comments