菜单
Design and run an evaluation for an AI or agent feature using representative tasks, explicit success criteria, failure taxonomy, baselines, and reproducible evidence.
默认使用会先检查来源的 Prompt；你也可以切换为直接命令，或下载本地副本。
决定是否安装前，请先阅读 SKILL.md，以及 SkillsMP 当前展示的配套文件。
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
直接命令不会经过审查 Prompt；运行前请先检查来源。
npx skills add https://github.com/Dadmin88/hermes-agency --skill llm-evaluation命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地？可下载 SkillsMP 当前能够提供的文件。
Perform a structured accessibility audit of a product experience or implementation and produce prioritized remediation guidance.
Review product design artifacts before implementation for accessibility of information, interaction, focus, input, error recovery, responsive/reflow behavior, motion, contrast, content alternatives, and assistive-technology semantics.
Validate a user flow with keyboard or equivalent non-pointer input by checking reachability, order, focus visibility, activation, escape, shortcuts, custom widgets, and recovery from dynamic UI changes.
基于 SOC 职业分类
正在显示 SKILL.md
| name | llm-evaluation |
| description | Design and run an evaluation for an AI or agent feature using representative tasks, explicit success criteria, failure taxonomy, baselines, and reproducible evidence. |
Use before relying on a model, prompt, retrieval flow, or agent behavior in a consequential product workflow.
A model change is better only when evidence shows improvement on the intended task without unacceptable regression, cost, latency, or safety tradeoffs.
