Token导航 LogoToken导航

Agent Skill · 研究检索

blog-drafter(博客起草者)

blog-drafter 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。

blog-drafter 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 73,674 个 Skill 里排第 56,650。

总安装 447周安装 19GitHub Stars 4下载 157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:blog-drafter(博客起草者)
来源仓库:https://github.com/petekp/agent-skills
仓库路径:skills/blog-drafter
安装命令:
npx skills add https://github.com/petekp/agent-skills --skill blog-drafter
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/petekp/agent-skills --skill blog-drafter

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。

用于访谈式内容整理与初稿生成,提取作者独特视角形成论证框架。适合深度专题报道或专家观点整合,避免泛泛而谈缺乏个性。通过系列提问锁定核心论点,输出带参考文献建议的完整草稿。安装方式:GitHub 仓库,使用 npx skills add 命令添加;需多轮对话完成信息整理。

注意:初稿仅为雏形,重大事实陈述务必交叉验证来源可靠性。

来源仓库内路径:skills/blog-drafter

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/petekp/agent-skills --skill blog-drafter;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

blog-drafter 怎么安装?

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

blog-drafter 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。

blog-drafter 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。

blog-drafter 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。

blog-drafter 用的人多吗?

公开数据是累计安装 447 次,近一周安装 19 次,下载 157 次,GitHub 4 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。

blog-drafter 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
blog-drafter

博客起草者 · 本页

4474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

Blog Drafter

Interview the user to extract their unique perspective, then produce a structured draft with thesis, outline, and research suggestions.

Process Overview

Phase 1: Discovery Interview → Structured Draft + Research
Phase 2: Prose Refinement Interview (after user approves draft)

Phase 1: Discovery Interview

Opening

Ask what topic they want to write about. If they've already stated it, acknowledge and move directly to the interview.

Interview Strategy

Use AskUserQuestion for structured choices. Use regular follow-up questions for open-ended exploration. Aim for 4-6 question rounds total.

Round 1: Core Thesis

AskUserQuestion:
  question: "What's the single most important thing you want readers to take away?"
  options:
    - "A specific insight or realization"
    - "A call to change behavior or practice"
    - "A framework or mental model"
    - "A contrarian or non-obvious take"

Then probe: "Can you state that in one sentence?"

Round 2: The "So What"

Ask directly: "Why should a PM, designer, or engineer care about this right now? What pain or opportunity does this address?"

Round 3: Evidence & Experience

AskUserQuestion:
  question: "What's your strongest evidence for this thesis?"
  options:
    - "Personal experience or case study"
    - "Data or research I've seen"
    - "Pattern I've observed across projects/companies"
    - "Logical argument from first principles"

Follow up: "Walk me through the specific example or evidence."

Round 4: Anticipated Resistance

Ask: "What's the strongest objection someone might raise? What would a skeptic say?"

Round 5: Unique Angle

AskUserQuestion:
  question: "What makes your perspective different from what's already written on this topic?"
  options:
    - "I have direct experience others don't"
    - "I'm connecting ideas that aren't usually connected"
    - "I disagree with conventional wisdom"
    - "I have a specific framework or process"

Round 6: Scope & Format

AskUserQuestion:
  question: "What length and depth feels right?"
  options:
    - "Short and punchy (800-1200 words)"
    - "Standard blog post (1500-2500 words)"
    - "Deep dive (3000+ words)"

Interview Principles

  • Listen for contradictions—they often reveal the real insight
  • When answers are abstract, ask for concrete examples
  • If the thesis sounds generic, push: "What would make someone disagree with this?"
  • Capture specific phrases and terminology the user employs

Phase 1 Output: Structured Draft

After the interview, produce:

1. Thesis Statement

One clear sentence stating the core argument.

2. Draft Structure

## [Working Title]

**Hook**: [Opening that creates tension or curiosity]

**Thesis**: [Core argument, stated directly]

### Section 1: [Setup/Context]
- Key point
- Key point

### Section 2: [Core Argument/Evidence]
- Key point with specific example from interview
- Key point

### Section 3: [Addressing Objections]
- Anticipated resistance
- Response

### Section 4: [Implications/Call to Action]
- What readers should do differently
- Why it matters

**Closing**: [Callback to hook or forward-looking statement]

3. Research Suggestions

Provide 3-5 specific suggestions:

  • Relevant studies, books, or articles to cite
  • Data points that would strengthen arguments
  • Examples from well-known companies/products that illustrate points
  • Experts or practitioners whose work relates to the thesis

Format as actionable items:

## Suggested Research

- [ ] Look for data on [specific metric/phenomenon] to support Section 2
- [ ] Reference [Author]'s work on [topic] for theoretical grounding
- [ ] Find a counter-example from [domain] to strengthen the objection response
- [ ] Check if [Company] has published anything on their approach to [topic]

4. Open Questions

Note 2-3 areas where more depth or clarity would strengthen the piece.


After presenting the draft, ask: "Does this structure capture what you want to say? Any sections that feel wrong or missing?"

Phase 2: Prose Refinement

Trigger Phase 2 only after user approves the structure.

Refinement Interview

Round 1: Tone

AskUserQuestion:
  question: "What tone fits this piece?"
  options:
    - "Conversational and accessible"
    - "Authoritative and direct"
    - "Provocative and opinionated"
    - "Thoughtful and nuanced"

Round 2: Opening Style

AskUserQuestion:
  question: "How do you like to open posts?"
  options:
    - "Start with a story or anecdote"
    - "Lead with the controversial claim"
    - "Open with a question"
    - "Set up a problem or tension"

Round 3: Technical Depth

Ask: "How much should I explain? Are readers already familiar with [key concepts from interview], or do they need context?"

Round 4: Specific Preferences

Ask: "Any writing patterns you like or hate? (e.g., 'I never use bullet points' or 'I always include code examples')"

Refinement Output

Expand the structure into full prose, incorporating:

  • The chosen tone throughout
  • The selected opening style
  • Appropriate technical depth
  • User's stated preferences

Mark areas where user's voice is needed:

[VOICE: Add your personal take on why this matters to you]
[EXAMPLE: Insert specific story from your experience here]

Remind user: "This is a starting point for your voice. The final pass is yours."

平台分布

Codex

34.84%
按下载量换算55

Claude

33.97%
按下载量换算53

Cursor

17.88%
按下载量换算28

Gemini CLI

10.28%
按下载量换算16

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