Token导航 LogoToken导航TokenDH.com
研究检索只读github未标认证来源可访问许可证需确认审计提醒

draft-outreach外展草案

Agent Skill

draft-outreach 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。

总安装

26,928

周安装

1,108

GitHub Stars

11,693

下载量

8,624
CodexClaudeCursorGemini CLI

安装说明

本站只整理中文说明和来源信息,不托管安装包,也不代用户安装。

GitHub

来源数

2

许可证

unknown

最后核验

2026-05-01

来源状态

来源可访问

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:draft-outreach(外展草案)
来源仓库:https://github.com/anthropics/knowledge-work-plugins
仓库路径:skills/draft-outreach
安装命令:
npx skills add https://github.com/anthropics/knowledge-work-plugins --skill draft-outreach
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/anthropics/knowledge-work-plugins --skill draft-outreach

简介

研究潜在客户,然后根据调查结果起草个性化的推广活动。

  • 默认情况下,首先使用网络搜索来研究潜在客户,并通过可选的丰富工具(经过验证的联系信息、背景)和 CRM 数据(先前的关系背景)进行增强
  • 生成具有个性化开头、相关挂钩和清晰 CTA 的电子邮件草稿,以及电子邮件不可用时的 LinkedIn 替代方案
  • 支持三种发送模式:如果启用电子邮件连接器,则自动创建电子邮件草稿、手动复制的输出文本或 LinkedIn 连接/后续消息模板
  • 包括后续序列模板,并通过将每条消息锚定到特定的研究结果(新闻、招聘、内容、相互联系)来避免通用的开场白

SKILL.md

Draft Outreach

Research first, then draft. This skill never sends generic outreach - it always researches the prospect first to personalize the message. Works standalone with web search, supercharged when you connect your tools.

Connectors (Optional)

ConnectorWhat It Adds
EnrichmentVerified email, phone, background details
CRMPrior relationship context, existing contacts
EmailCreate draft directly in your inbox
No connectors? Web research works great. I'll output the email text for you to copy.

How It Works

+------------------------------------------------------------------+
|                      DRAFT OUTREACH                               |
|                                                                   |
|  Step 1: RESEARCH (always happens first)                         |
|  - Web search (default)                                           |
|  - + Enrichment (if enrichment tools connected)                  |
|  - + CRM (if CRM connected)                                      |
|                                                                   |
|  Step 2: DRAFT (based on research)                               |
|  - Personalized opening (from research)                          |
|  - Relevant hook (their priorities)                              |
|  - Clear CTA                                                      |
|                                                                   |
|  Step 3: DELIVER (based on connectors)                           |
|  - Email draft (if email connected)                              |
|  - Copy for LinkedIn (always)                                    |
|  - Output to user (always)                                        |
+------------------------------------------------------------------+

Output Format

# Outreach Draft: [Person] @ [Company]
**Generated:** [Date] | **Research Sources:** [Web, Enrichment, CRM]

---

## Research Summary

**Target:** [Name], [Title] at [Company]
**Hook:** [Why reaching out now - the personalized angle]
**Goal:** [What you want from this outreach]

---

## Email Draft

**To:** [email if known, or "find email" note]
**Subject:** [Personalized subject line]

---

[Email body]

---

**Subject Line Alternatives:**
1. [Option 2]
2. [Option 3]

---

## LinkedIn Message (if no email)

**Connection Request (< 300 chars):**
[Short, no-pitch connection request]

**Follow-up Message (after connected):**
[Value-first message]

---

## Why This Approach

| Element | Based On |
|---------|----------|
| Opening | [Research finding that makes it personal] |
| Hook | [Their priority/pain point] |
| Proof | [Relevant customer story] |
| CTA | [Low-friction ask] |

---

## Email Draft Status

[Draft created - check ~~email]
[Email not connected - copy email above]
[No email found - use LinkedIn approach]

---

## Follow-up Sequence (Optional)

**Day 3 - Follow-up 1:**
[Short, new angle]

**Day 7 - Follow-up 2:**
[Different value prop]

**Day 14 - Break-up:**
[Final attempt]

Execution Flow

Step 1: Parse Request

Input patterns:
- "draft outreach to John Smith at Acme" → Person + company
- "write cold email to Acme's CTO" → Role + company
- "reach out to sarah@acme.com" → Email provided
- "LinkedIn message to [LinkedIn URL]" → Profile provided

Step 2: Research First (Always)

Use research-prospect skill internally:

1. Web search for company + person
2. If Enrichment connected: Get verified contact info, background
3. If CRM connected: Check for prior relationship

Must find before drafting:

  • Who they are (title, background)
  • What the company does
  • Recent news or trigger
  • Personalization hook

Step 3: Identify Hook

Priority order for hooks:
1. Trigger event (funding, hiring, news) → Most timely
2. Mutual connection → Social proof
3. Their content (post, article, talk) → Shows you did research
4. Company initiative → Relevant to their priorities
5. Role-based pain point → Least personal but still relevant

Step 4: Draft Message

Email Structure (AIDA):

SUBJECT: [Personalized, <50 chars, no spam words]

[Opening: Personal hook - shows you researched them]

[Interest: Their problem/opportunity in 1-2 sentences]

[Desire: Brief proof point - similar company result]

[Action: Clear, low-friction CTA]

[Signature]

LinkedIn Connection Request (<300 chars):

Hi [Name], [Mutual connection/shared interest/genuine compliment].
Would love to connect. [No pitch]

LinkedIn Follow-up Message:

Thanks for connecting! [Value-first: insight, article, observation]

[Soft transition to why you reached out]

[Question, not pitch]

Step 5: Create Email Draft

If email connector available:
1. Create draft with to, subject, body
2. Return draft link
3. Note: "Draft created - review and send"

If not available:
1. Output email text
2. Note: "Copy to your email client"

Capability by Connector

CapabilityWeb Only+ Enrichment+ CRM+ Email
Personalized openingBasicDeepWith historySame
Verified emailNoYesYesYes
Background detailsPublic onlyFullFullFull
Prior relationshipNoNoYesYes
Auto-create draftNoNoNoYes

Message Templates by Scenario

Cold Outreach (No Prior Relationship)

Subject: [Their initiative] + [your angle]

Hi [Name],

[Personal hook based on research - news, content, mutual connection].

[1 sentence on their likely challenge based on role/company].

[Brief proof: "We helped [Similar Company] achieve [Result]".]

Worth a 15-min call to see if relevant?

[Signature]

Warm Outreach (Have Met / Mutual Connection)

Subject: Following up from [context]

Hi [Name],

[Reference to how you know them / who connected you].

[Why reaching out now - their trigger].

[Specific value you can offer].

[CTA]

Re-Engagement (Went Dark)

Subject: [Short, curiosity-driven]

Hi [Name],

[Acknowledge time passed without being guilt-trippy].

[New reason to reconnect - their news or your news].

[Simple question to re-open dialogue].

[Signature]

Post-Event Follow-up

Subject: Great meeting you at [Event]

Hi [Name],

[Specific memory from conversation].

[Value-add: article, intro, resource related to what you discussed].

[Soft CTA for next conversation].

Email Style Guidelines

  1. Be concise but informative — Get to the point quickly. Busy people skim.
  2. No markdown formatting — Never use asterisks, bold (text), or other markdown. Write plain text that looks natural in any email client.
  3. Short paragraphs — 2-3 sentences max per paragraph. White space is your friend.
  4. Simple lists — If listing items, use plain dashes. No fancy formatting.

Good:

Here's what I can share:
- Case study from a similar company
- 15-min intro call this week
- Quick demo if helpful

Bad:

**What I Can Offer:**
- **Case study** from a similar company
- **Intro call** this week

What NOT to Do

Generic openers:

  • "I hope this email finds you well"
  • "I'm reaching out because..."
  • "I wanted to introduce myself"

Feature dumps:

  • Long paragraphs about your product
  • Multiple value props at once
  • No clear CTA

Fake personalization:

  • "I noticed you work at [Company]" (obviously)
  • "Congrats on your role" (without context)

Markdown in emails:

  • Using bold or *italic* asterisks
  • Headers or formatted lists that won't render

Instead:

  • Lead with something specific you learned
  • One clear value prop
  • One clear ask
  • Plain text formatting only

Channel Selection

IF verified email available:
  → Email preferred (higher response rate)
  → Also provide LinkedIn backup

IF no email:
  → LinkedIn connection request
  → Follow-up message template for after connection

IF warm intro possible:
  → Suggest mutual connection outreach first

Company Configuration [CUSTOMIZE]

## Outreach Settings

- My name: [Your Name]
- My title: [Your Title]
- My company: [Company Name]
- Value prop: [One sentence - what you help with]

## Signature
[Your preferred email signature]

## Proof Points
- [Customer 1]: [Result]
- [Customer 2]: [Result]
- [Customer 3]: [Result]

## CTA Options
- Default: "Worth a 15-min call?"
- Softer: "Open to learning more?"
- Specific: "Can I send over a quick demo?"

## Tone
- [Professional / Casual / Direct]
- Industry-specific language: [Yes/No]

Example

Input: "draft outreach to the Head of Engineering at Notion"

Research finds:

  • Name: David Tibbitts
  • Recent news: Notion launched AI features
  • Hiring: 5 ML engineer roles open
  • Background: Ex-Facebook, Stanford CS

Output:

# Outreach Draft: David Tibbitts @ Notion

## Research Summary
**Target:** David Tibbitts, Head of Engineering at Notion
**Hook:** They're scaling AI features and hiring ML talent
**Goal:** Intro call about AI infrastructure

---

## Email Draft

**To:** david@notion.so
**Subject:** Notion's AI scaling + a thought

---

Hi David,

Saw Notion's AI rollout is gaining serious traction - congrats.
With 5 ML roles open, seems like you're scaling fast.

Curious how you're thinking about inference infrastructure
as usage grows. We helped [Similar Company] cut their AI
serving costs 40% while improving latency.

Worth a 15-min call to see if relevant to your roadmap?

Best,
[Name]

---

**Subject Alternatives:**
1. Notion AI + scaling question
2. Quick thought on Notion's ML hiring

---

## Email Draft Status
Draft created - check ~~email

适合场景

01

用户想查找某类 Agent Skill 时

02

需要根据任务场景推荐可安装能力包时

03

需要对比不同来源的安装命令和来源信息时

能力概览

能力 1

按任务关键词查找相关 Skills

能力 2

展示可复制的安装命令

能力 3

保留来源站点、仓库和原始说明,方便继续核验

能力 4

展示第三方安全扫描或审计结果

安装后应在对应宿主中按原始 README 的触发条件使用;具体调用方式请以来源页面和 README 为准。

平台分布

Codex

37.46%
按下载量换算3,231

Claude

30.58%
按下载量换算2,637

Cursor

18.93%
按下载量换算1,633

Gemini CLI

9.52%
按下载量换算821

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

可疑

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。来源安全扫描存在 warning/failed 结果,不能写成本站确认安全。当前只有一个来源,正式发布前建议补源仓库或其他目录站核验。

来源信息

继续浏览同类 Skills