Evaluate AI agent performance using systematic testing and multi-dimensional rubrics.
npx skills add https://github.com/boazcstrike/opencode --skill evaluation-boazcstrike
Please help me install this Agent Skill. Skill: evaluation Source: https://github.com/boazcstrike/opencode/tree/main/skills/evaluation Command: npx skills add https://github.com/boazcstrike/opencode --skill evaluation-boazcstrike
This Skill addresses the challenge of systematically testing and validating the performance of AI agents, ensuring their reliability and effectiveness.
Use the evaluation skill to build a test framework for your agent.
High-intent search queries and answers about installing and using this skill.
Build a test framework using multi-dimensional rubrics and detailed metrics to measure accuracy, completeness, and efficiency across a suite of test cases for your AI agents.
Testing non-deterministic agents requires a comprehensive evaluation framework that accommodates multiple valid execution paths, utilizing both LLM-as-judge and human evaluation methodologies to ensure thorough quality assurance.
Run your updated agent against a suite of test cases to verify that quality assurance metrics such as accuracy, completeness, and efficiency have improved or at least not degraded.
Yes, LLM-as-judge is supported as an evaluation methodology alongside human evaluation, providing a scalable approach to thoroughly assess agent outputs against defined multi-dimensional rubrics.
Build a robust evaluation framework by defining and applying multi-dimensional rubrics to systematically test agent systems, ensuring their reliability and effectiveness through detailed metrics.
NicolasPadovamDevelop, evaluate, and optimize custom agent skills within the Opensquad framework.
slowman2084Analyze prompt histories to detect anti-patterns and generate retrospective reports.
flutterOrchestrate automated evaluation of agent skills using isolated subagents and rubrics.
guanyangEvaluates agent performance using multidimensional rubrics and systematic testing.
