菜单
Verify the actual material evidence behind a selected assertion, claim, finding, or final result. Use when the user explicitly invokes evidence or asks to substantiate, prove, check, or challenge something an agent reported.
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/minghinmatthewlam/agent-guards --skill evidence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。
Create a self-contained HTML explanation when the user requests one or when important, durable project knowledge is materially easier to understand visually than in concise chat. Use for significant architecture, research, decisions, incidents, audits, or changes that the user may revisit. Do not use for routine status, simple answers, or unverified findings.
Present concise plans, progress updates, findings, and final reports that keep the human oriented, informed, and able to learn how the project works. Use for complex agent coding work and focus on the most important points and decisions.
Run the bundled OpenClaw structured code-review helper for local changes, branches, commits, and PRs. Use for ordinary code-review requests, default closeout review, or an explicitly requested Codex/Claude review panel.
基于 SOC 职业分类
正在显示 SKILL.md
| name | evidence |
| description | Verify the actual material evidence behind a selected assertion, claim, finding, or final result. Use when the user explicitly invokes evidence or asks to substantiate, prove, check, or challenge something an agent reported. |
Check whether the selected claim is supported by evidence that actually proves it. This is an on-demand verification pass, not a requirement to expand every normal report.
