artifact-evaluation

Evaluating Copilot artifacts against a structured rubric involves applying specific assessment criteria to prompts, skills, and workflows, then calculating weighted scores to ensure quality and adherence to predefined standards.

Chain-of-Verification methodology for artifact evaluation incorporates self-critique and evidence-based validation to identify potential issues, revising Copilot artifacts based on feedback before calculating final weighted scores.

Yes, you can use this evaluation process for Copilot prompts and workflows. The Skill supports the assessment of prompts, skills, and workflows by applying rubric criteria to measure their effectiveness.

You specify the artifact path for Copilot evaluation by providing a designated file location, such as the .github/prompts directory, which is required for the Skill to execute the rubric assessment.

Artifact evaluation operates using dependencies like Jest, Babel, and preset-env. These frameworks support the underlying scripts that automate the rigorous evaluation of Copilot artifacts against predefined rubrics.