跳至正文
OpenAgentSkill
浏览已开放 · 目录记录日期 2026-11-21 · 自动更新和账户服务暂时关闭。
浏览器与自动化

skill-folder-name

One sentence stating what this skill produces and per which standard or character. Use when the user [concrete trigger situations, the actual phrases they'd type, and the contexts where this beats default prose].

surendranb · 已有静态检查记录

安装与使用

运行命令前先阅读来源与依赖说明。复制命令不会自动安装或执行。

npx skills add surendranb/writing-skills --skill skill-folder-name

技能说明


name: skill-folder-name description: One sentence stating what this skill produces and per which standard or character. Use when the user [concrete trigger situations, the actual phrases they'd type, and the contexts where this beats default prose].

Skill Title

One line: the essence of this style in a sentence fragment.

The core rule

The single bolded sentence that IS this skill. One or two sentences of sharpening — what this style is really doing and the failure it prevents.

Workflow: first move → second move → third move → run the Verify checklist.

Mechanics

  1. Named rule. Concrete and checkable — a cap, a ban list, a required shape.
  2. Named rule. "X not Y" with a real example inline.
  3. Named rule. Numbers where possible (sentence length, ratios, counts).
  4. Named rule. For voices: include a rate-limit that prevents caricature.

Verify

  • Checkable statement (an agent can pass/fail this mechanically)
  • Checkable statement
  • Checkable statement

Do not

  • The tempting mistake this style invites
  • The failure mode that looks like the style but isn't

Example transformations

Before: "A realistic 'default prose' input."

After: "The same content, transformed by the mechanics above."

Before: "A second, different-shaped input."

After: "Its transformation."