agentic-eval

What problem does it solve?

This Skill addresses the challenge of ensuring AI-generated outputs meet specific quality standards by enabling iterative refinement through self-evaluation and correction.

Core Features & Use Cases

• Self-Critique Loops: Agents can evaluate their own responses against defined criteria and identify areas for improvement.
• Evaluator-Optimizer Pipelines: Separates the generation and evaluation processes for clearer workflow management and targeted refinement.
• Code-Specific Reflection: Utilizes unit tests to drive code generation and bug fixing.
• Use Case: When generating a complex report, an agent can first produce a draft, then critically assess it for accuracy, clarity, and completeness, and finally revise it based on the identified shortcomings.

Quick Start

Use the agentic-eval skill to refine the generated code for the specified task by running it through a self-critique and improvement loop.