Token导航 LogoToken导航

Agent Skill · 研究检索

written-communication(书面沟通)

written-communication 用于查找、检索和筛选相关信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要根据关键词、任务场景或来源线索快速定位候选结果时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

written-communication 是研究检索类 Agent Skill,可用宿主为 Codex、Claude、Cursor、Gemini CLI,通过GitHub 仓库获取,权限为只读;许可证信息还不明确,第三方扫描已通过,按安装量在研究检索类的 73,675 个 Skill 里排第 58,252。

总安装 419周安装 18GitHub Stars 3下载 147核验 2026-05-01

安装说明

本站只整理中文说明,不托管安装包,也不代用户安装。
GitHub
许可证
unknown

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:written-communication(书面沟通)
来源仓库:https://github.com/oldwinter/skills
仓库路径:skills/written-communication
安装命令:
npx skills add https://github.com/oldwinter/skills --skill written-communication
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 npx skills 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

skills.shnpx skills
npx skills add https://github.com/oldwinter/skills --skill written-communication

这个 Skill 是做什么的

研究检索类 Skill 主要解决「找资料、核对事实、把来源整理清楚」的问题,通常会让助手先检索再总结。

written-communication 用于查找、检索和筛选相关信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要根据关键词、任务场景或来源线索快速定位候选结果时使用。它可辅助信息聚合与初步筛选,提升 Agent 在知识探索阶段的效率。

通过 npx skills add 命令从指定仓库安装,需结合原始 README 核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。适用宿主包括 Codex、Claude、Cursor、Gemini CLI,接入前应确认版本、权限和运行环境要求。

来源仓库内路径:skills/written-communication

上手步骤

4 步
  1. 1

    确认自己的宿主工具

    公开资料标注支持 Codex、Claude、Cursor、Gemini CLI;先确认自己用的是其中之一,再按对应工具的 Skills 目录放置。

  2. 2

    拿到安装命令

    可以直接复制的命令是 npx skills add https://github.com/oldwinter/skills --skill written-communication;命令里的占位符换成自己的路径或密钥再执行。

  3. 3

    安装后重启宿主

    命令执行完要完全退出宿主工具再打开,让新装的 Skill 被重新扫描到,否则列表里可能看不到。

  4. 4

    按最小权限试用

    它标注为只读,先用只读任务试一次,确认行为符合预期再扩大使用范围。

常见问题

6 条

written-communication 怎么安装?

在终端执行 npx skills add https://github.com/oldwinter/skills --skill written-communication。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

written-communication 支持哪些 AI 工具?

公开资料标注支持 Codex、Claude、Cursor、Gemini CLI。其他工具只要能读取本地 Skill 目录,一般也能手动放进去,但要自己确认触发方式是否一致。

written-communication 需要什么权限?

权限级别是只读。该 Skill 主要提供规则、说明或参考内容,本身偏只读;真正读写文件、联网或执行命令仍取决于宿主 Agent 的任务。

written-communication 是免费的吗?

公开资料没有标注明确许可证,收费和商用限制都无法确认。正式使用前请到来源仓库核对授权说明。

written-communication 用的人多吗?

公开数据是累计安装 419 次,近一周安装 18 次,下载 147 次,GitHub 3 星。这些数字来自第三方目录统计,只能当作热度参考,不代表质量或安全性。

written-communication 安全吗?

第三方扫描结果为通过,但这只覆盖已知的静态检查项,不能替代自己读一遍源码和安装命令。

值得选的地方和要先想清楚的地方

值得选的地方

  • 同时适配 4 种工具,换客户端不用重新找替代品。
  • 第三方扫描已通过,常见的高风险行为没有命中。
  • 标注为只读权限,试用阶段不容易误改本地文件。

先想清楚的地方

  • 许可证没有明确标注,能不能商用、能不能再分发都不确定。
  • 第三方目录整理,版本更新和行为细节以来源仓库为准。

适不适合你

更适合

  • 已经在用 Codex,想直接扩展它能力的人。
  • 需要处理研究检索类任务、又不想每次都手写提示词的人。
  • 只想先试试效果、不希望它改动本地文件的人。

不太适合

  • 需要明确授权才能用到公司项目里的场景。
  • 需要厂商级技术支持和 SLA 的生产系统。

研究检索类 Skill 对比

6 个同类
Skill总安装GitHub 星权限许可证扫描
find-skills

寻找技能 · Codex、Claude

31,199,81316,486操作浏览器MIT有提醒
self-improving-agent

自我改进 Agent · OpenClaw

16,464,0023,396接触敏感数据MIT-0通过
soultrace

灵魂追踪 · Codex、Claude

7,588,94412需要联网未明确有提醒
self-improving

自我完善 · OpenClaw

6,829,2841,046需要联网MIT-0通过
microsoft-foundry

微软代工厂 · Codex、Claude

6,614,144764调用外部服务未明确有提醒
written-communication

书面沟通 · 本页

4193只读未明确通过

按安装量取同分类前几名;数字来自第三方目录统计,只作热度参考,不代表质量结论。

SKILL.md 原文

Written Communication

Scope

Covers

  • Turning messy notes into a clear email, memo, doc, or async update
  • Making the “how” explicit (what happens next, by whom, by when)
  • Editing for clarity at scale (scanability, definitions, single source of truth)
  • Creating/maintaining a canonical doc for an ongoing project

When to use

  • “Draft an email to stakeholders explaining a change and what I need from them.”
  • “Turn these bullets into a 1-page memo with a recommendation and next steps.”
  • “Rewrite this doc to be clearer, shorter, and more actionable.”
  • “Create a canonical doc as the source of truth for this project.”

When NOT to use

  • You need marketing/brand copy (landing pages, ads) more than internal/executive clarity.
  • You need a full product spec/PRD from scratch (use writing-prds or writing-specs-designs).
  • You’re writing legal/HR/regulated communications without expert review.
  • The real issue is alignment via facilitation (you may need a meeting/offsite plan, not a rewrite).

Inputs

Minimum required

  • Artifact type + channel (email / memo / doc / status update; where it will live)
  • Audience (roles/seniority) + what they care about
  • Goal + ask (inform/align/decide; what you want the reader to do, by when)
  • Key context (facts, constraints, timeline, links) + what must be avoided (sensitivities)
  • Source material (notes, existing draft, Slack threads, etc.)

Missing-info strategy

  • Ask up to 5 questions from references/INTAKE.md (3–5 at a time), then proceed.
  • If critical info remains missing, make explicit assumptions and offer 2–3 options (structure/tone/ask).

Outputs (deliverables)

Produce a Written Communication Pack in Markdown (in-chat; or as files if requested):

  1. Message brief (audience, goal, ask, constraints)
  2. Outline (TL;DR + key points + “how/next steps”)
  3. Draft artifact (email/memo/doc/status update) in final-ready format
  4. Canonical doc skeleton (optional; when the project needs a single source of truth)
  5. Risks / Open questions / Next steps (always)

Templates: references/TEMPLATES.md Expanded guidance: references/WORKFLOW.md

Workflow (8 steps)

1) Intake + choose the lightest artifact

  • Inputs: user request + references/INTAKE.md.
  • Actions: Clarify the channel and pick the smallest artifact that works (email vs memo vs doc vs status update vs canonical doc).
  • Outputs: Message brief (draft) + artifact selection.
  • Checks: You can answer: “Who is this for, and what should they do after reading?”

2) Lock the reader outcome + ask (one sentence)

  • Inputs: brief.
  • Actions: Write one sentence: “After reading, the audience will ____.” Make the ask explicit (decision/options, approval, feedback, or FYI) and include a deadline if relevant.
  • Outputs: Outcome/ask line + decision/feedback request.
  • Checks: The ask is unambiguous and doesn’t require a meeting to interpret.

3) Convert “what/why” into “how” (actionable next steps)

  • Inputs: source material + outcome/ask.
  • Actions: Identify the 3–7 concrete steps, responsibilities, and dependencies. If proposing a change, include what changes, what stays the same, and what happens next.
  • Outputs: “How / Next steps” bullets (owner + date where possible).
  • Checks: A reader could execute without asking “so what do you want me to do?”

4) Structure for skim (clarity at scale)

  • Inputs: brief + next steps.
  • Actions: Create a TL;DR, then headings in the order readers scan: Ask → Context → Details → Next steps. Use bullets, short paragraphs, and explicit labels.
  • Outputs: Outline with headings.
  • Checks: A skim-reader can capture the point in < 60 seconds.

5) Draft the artifact (write to be forwarded)

  • Inputs: outline + templates.
  • Actions: Draft in plain language; avoid jargon; put key numbers and decisions in writing. If this is ongoing work, link to (or create) the canonical doc.
  • Outputs: Draft email/memo/doc/status update.
  • Checks: The draft is safe to forward; it stands alone without verbal context.

6) “Letter to yourself” clarity pass (then rewrite for the audience)

  • Inputs: draft.
  • Actions: If the content is fuzzy, write a quick internal version (“what am I actually saying?”), then rewrite in the audience’s language and incentives.
  • Outputs: Clarified rewrite with cleaner logic.
  • Checks: The message has a single through-line; no contradictions or buried ledes.

7) Canonical doc check (single source of truth)

  • Inputs: draft + project context.
  • Actions: If readers will keep asking “where is the latest?”, create/update a canonical doc (links, owners, last updated, decisions, next update cadence).
  • Outputs: Canonical doc skeleton or link section.
  • Checks: There is one obvious place to find the current state and decisions.

8) Quality gate + finalize

  • Inputs: full pack.
  • Actions: Run references/CHECKLISTS.md and score with references/RUBRIC.md. Add Risks/Open questions/Next steps.
  • Outputs: Final Written Communication Pack.
  • Checks: Clarity, actionability, and ownership meet the bar (≥ 3 on each rubric dimension).

Quality gate (required)

Examples

Example 1 (stakeholder email): “Draft an email to exec stakeholders: the launch is slipping 2 weeks; we need approval to cut scope and a decision by Friday.” Expected: TL;DR + explicit ask/options + what changes + next steps with owners.

Example 2 (project memo + canonical doc): “Turn these notes into a 1-page memo that aligns the team on the new onboarding approach, and create a canonical doc outline for ongoing updates.” Expected: memo with recommendation + tradeoffs + next steps, plus a source-of-truth doc skeleton.

Boundary example: “Write a legal/HR disciplinary notice.” Response: decline to fabricate legal/HR guidance; request expert review; offer to help with neutral structure, tone, and clarity if the user provides approved language.

平台分布

Codex

37.06%
按下载量换算54

Claude

33.45%
按下载量换算49

Cursor

17.77%
按下载量换算26

Gemini CLI

8.55%
按下载量换算13

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

只读

该 Skill 主要提供规则、说明或参考内容,本身偏只读;真正读写文件、联网或执行命令仍取决于宿主 Agent 的任务。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。目前只有一个公开来源,建议同时核对源仓库。

继续浏览同类 Skills