菜单
Delegate simple sub-tasks to a fast model (GLM-5). Use for fetch, extract, query, and tool-heavy work that needs minimal reasoning.
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/Enriquefft/zeroclaw --skill fast-run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。
基于 SOC 职业分类
正在显示 SKILL.md
| name | fast-run |
| description | Delegate simple sub-tasks to a fast model (GLM-5). Use for fetch, extract, query, and tool-heavy work that needs minimal reasoning. |
Delegate simple, self-contained sub-tasks to GLM-5. The fast model has access to all tools (browser, email, calendar, sqlite, shell) but runs at lower reasoning effort.
Run a single self-contained task via GLM-5.
The task string must be complete and self-contained. Include:
Output: JSON to stdout with success and output fields.
fast_run_cli "Search Wellfound for founding engineer roles. Return JSON array: [{title, company, url, salary, remote}]. Limit 20."
fast_run_cli "Read ~/.zeroclaw/documents/LORE.md and extract the job boards list as a JSON array of strings."
fast_run_cli "Query state.db: SELECT url FROM job_applications WHERE status='new'. Return the URLs as a newline-separated list."