creator-outreach

Compare original and translation side by side

🇺🇸

Original

English

🇨🇳

Translation

Chinese

Creator Outreach

创作者外联

Follow shared release-shell rules in:

```
postplus-shared
```
release-shell rules

Use this skill when the user wants to:

find creator or influencer leads across TikTok, Instagram, and X
combine multiple discovery routes into one candidate pool
extract public contact signals such as email or link-in-bio URLs
convert normalized platform datasets into a single outreach-ready lead table
prepare outreach drafts before Gmail or CRM delivery

Read these references before implementation:

skills/50-publishing/creator-outreach/references/lead-schema.md

skills/50-publishing/creator-outreach/references/routing.md

遵循以下共享的release-shell规则：

```
postplus-shared
```
release-shell规则

当用户有以下需求时，可使用此技能：

在TikTok、Instagram和X平台上寻找创作者或网红线索
将多种挖掘渠道整合为一个候选池
提取公开联系信号，如邮箱或主页链接（link-in-bio）
将标准化平台数据集转化为统一的可用于外联的线索表格
在通过Gmail或CRM发送前准备外联草稿

实施前请阅读以下参考文档：

skills/50-publishing/creator-outreach/references/lead-schema.md

skills/50-publishing/creator-outreach/references/routing.md

Core Method

核心方法

Do not lock this skill to one niche, one follower band, or one creator archetype.

Use a two-stage model:

build a candidate pool from one or more discovery routes
run one shared shortlist pass across those candidates

This skill is the shortlist and outreach layer above the platform skills. It should not assume:

only mid-tier creators matter
only one niche matters
keyword account search is the only discovery path
content-first is always the best path

Choose the route that fits the job.

Default posture:

keep the default flow light
add deeper evidence only when the user request clearly needs it

Do not turn every creator discovery task into a full competitor-plus-comments-plus-market-fit investigation by default.

不要将此技能限定于单一细分领域、单一粉丝量级或单一创作者类型。

采用两阶段模型：

通过一种或多种挖掘渠道构建候选池
对这些候选者执行统一的筛选流程

此技能是平台技能之上的筛选和外联层，不应预设：

只有中量级创作者才重要
只有单一细分领域才重要
关键词账号搜索是唯一的挖掘路径
内容优先始终是最佳路径

选择符合任务需求的渠道。

默认原则：

保持默认流程轻量化
仅当用户需求明确需要时，才添加更深入的验证信息

不要默认将每一项创作者挖掘任务都变成全面的竞品分析+评论分析+市场适配调研。

Discovery Routes

挖掘渠道

Use one or more of these routes before lead building:

在构建线索前，可使用以下一种或多种渠道：

1. Official pool route

1. 官方池渠道

Examples:

TikTok Shop Seller Center creator marketplace
platform-native creator marketplace exports
manually assembled seller-center handle lists

Use when:

the user already has an official marketplace export
the platform exposes filtering for category, market, follower band, or engagement
the goal is to build a first-pass candidate pool quickly

Current capability:

this skill can ingest exported or manually copied handle lists after they are normalized into local datasets
this skill does not currently automate clicking through seller-center UI
use this route only when the user already has official-pool data or explicitly wants that route

示例：

TikTok Shop商家中心创作者市场
平台原生创作者市场导出数据
手动整理的商家中心账号列表

适用场景：

用户已有官方市场导出数据
平台支持按品类、市场、粉丝量级或互动率筛选
目标是快速构建首轮候选池

当前能力：

此技能可接收已标准化为本地数据集的导出或手动复制的账号列表
此技能目前无法自动操作商家中心UI
仅当用户已有官方池数据或明确要求使用该渠道时，才采用此路径

2. Keyword search route

2. 关键词搜索渠道

Do not search only category nouns.

Prefer combinations like:

core product term +
```
review
```
core product term +
```
unboxing
```
core product term +
```
haul
```
problem/use-case term +
```
routine
```
commerce tags such as
```
TikTokMadeMeBuyIt
```
,
```
TikTokShop
```
, or platform-native shopping tags when relevant

Use when:

the user needs broad market discovery
the user wants creators actively posting around a product or use case
there is no strong seed list yet

This is a common default discovery route when no stronger seed source exists.

不要仅搜索品类名词。

优先使用组合关键词，例如：

核心产品词 +
```
review
```
核心产品词 +
```
unboxing
```
核心产品词 +
```
haul
```
问题/使用场景词 +
```
routine
```
电商标签，如
```
TikTokMadeMeBuyIt
```
、
```
TikTokShop
```
，或相关平台原生购物标签

适用场景：

用户需要广泛的市场挖掘
用户希望找到围绕产品或使用场景积极发布内容的创作者
目前没有明确的种子列表

当没有更优质的种子来源时，这是常见的默认挖掘渠道。

3. Competitor backtracking route

3. 竞品回溯渠道

Start from:

competitor brand accounts
competitor product videos
competitor collab posts
competitor comment sections

Inspect:

creators already collaborating with the competitor
posts with shopping or partner signals
active commenters asking for link, shade, size, price, or buying help
likely KOC accounts that can be expanded later

Use when:

the user wants faster conversion-oriented discovery
the user has direct competitors with visible creator activity
the goal is to shorten the path from discovery to outreach

This is an optional enhancement layer, not a default requirement.

从以下内容入手：

竞品品牌账号
竞品产品视频
竞品合作帖
竞品评论区

重点关注：

已与竞品合作的创作者
带有购物或合作标识的帖子
主动询问链接、款式、尺寸、价格或购买渠道的活跃评论者
可后续拓展的潜在KOC账号

适用场景：

用户希望更快找到转化导向的创作者
用户有明确的竞品且其创作者活动可见
目标是缩短从挖掘到外联的路径

这是可选的增强层，而非默认要求。

4. Content-first route

4. 内容优先渠道

Find relevant posts first, then pull the authors, then enrich the profiles.

Use when:

the user wants people who are clearly publishing the right content
account names and bios are weak signals
the job is format-aware or content-angle-aware discovery

This is often the best default when follower band and content fit both matter.

先找到相关帖子，再提取作者信息，然后完善其资料。

适用场景：

用户希望找到明确发布匹配内容的创作者
账号名称和简介的参考价值较低
任务需要关注内容格式或角度的挖掘

当粉丝量级和内容适配性都重要时，这通常是最佳默认选择。

5. Known-handle enrichment route

5. 已知账号完善渠道

Use when:

the user already has handles
the next step is profile enrichment, contact extraction, or outreach prep

适用场景：

用户已有账号列表
下一步需要完善资料、提取联系方式或准备外联

Screening Dimensions

筛选维度

After discovery, shortlist candidates with a shared framework.

Always consider:

account type:
- individual creator
- KOC
- educator / consultant
- brand / product account
- aggregator / media page
content fit:
- does the account repeatedly post around the product, problem, use case, or audience
audience fit:
- target market
- likely audience segment
- language
- market relevance
style fit:
- visual style
- creator tone
- trustworthiness
- cultural fit
engagement quality:
- likes
- comments
- shares
- comment intent quality
outreach readiness:
- public email
- link-in-bio
- past collab signals
- stable posting behavior

Do not treat follower count as the only priority.

Follower band should be chosen per campaign:

head creators for reach
mid-tier creators for repeatable paid or affiliate work
small creators or KOCs for low-cost seeding or testing

挖掘完成后，使用统一框架筛选候选者。

需始终考虑以下维度：

账号类型：
- 个人创作者
- KOC
- 教育者/顾问
- 品牌/产品账号
- 聚合/媒体账号
内容适配性：
- 账号是否持续发布与产品、问题、使用场景或受众相关的内容
受众适配性：
- 目标市场
- 潜在受众群体
- 使用语言
- 市场相关性
风格适配性：
- 视觉风格
- 创作者调性
- 可信度
- 文化适配性
互动质量：
- 点赞数
- 评论数
- 分享数
- 评论意图质量
外联就绪度：
- 公开邮箱
- 主页链接（link-in-bio）
- 过往合作标识
- 稳定的发布频率

不要将粉丝数量作为唯一优先级。

粉丝量级应根据活动需求选择：

头部创作者用于扩大曝光
中量级创作者用于可重复的付费或联盟合作
小量级创作者或KOC用于低成本推广或测试

Default vs Optional Depth

默认流程 vs 可选增强

Default flow:

pick one sensible discovery route
collect platform data
build leads
extract public contact signals
score and shortlist

Optional enhancement layers:

official-pool import
competitor backtracking
comment-intent evidence
TikTok Shop commerce signals
deeper market / language / style review

Only add these when:

the user explicitly asks for them
the core shortlist is weak and the next-best move is to deepen evidence
the campaign type clearly depends on that signal

Avoid stacking all optional layers into the first pass.

默认流程：

选择一个合理的挖掘渠道
收集平台数据
构建线索
提取公开联系信号
评分并筛选

可选增强层：

官方池导入
竞品回溯
评论意图验证
TikTok Shop电商信号
更深入的市场/语言/风格审核

仅在以下情况添加增强层：

用户明确要求
核心筛选名单质量不佳，需要补充验证信息
活动类型明确依赖该信号

避免在首轮流程中叠加所有可选增强层。

Current Automation vs Manual Judgment

当前自动化能力 vs 人工判断

This skill can score automatically from local normalized data:

follower scale
visible contact signals
niche overlap from bio and profile text
platform preference
coarse account type inference

This skill cannot reliably infer from current data unless the platform dataset already includes it:

follower geography
precise audience market distribution
deep style quality judgment
language nuance and cultural fit
competitor collaboration history from off-platform evidence
comment-level purchase intent unless comments were explicitly collected upstream

When these factors matter, treat them as required manual or upstream-research checks, not hidden assumptions.

此技能可通过本地标准化数据自动评分：

粉丝规模
可见的联系信号
简介和资料文本中的领域匹配度
平台偏好
粗略的账号类型推断

除非平台数据集已包含相关信息，否则此技能无法可靠推断：

粉丝地域分布
精准的受众市场分布
深度风格质量判断
语言差异和文化适配性
来自平台外的竞品合作历史
评论层面的购买意图（除非上游已明确收集评论数据）

当这些因素重要时，应将其视为必要的人工或上游调研环节，而非隐含假设。

Workflow

工作流程

1. Choose the discovery route first

1. 首先选择挖掘渠道

Pick the route based on the user request:

official pool export -> normalize or manually structure it, then enrich
broad discovery -> platform search route
competitor scouting -> collect creator candidates from competitor content first
content-angle discovery -> content-first route
known handles -> direct enrichment route

If the user gives no special constraints, prefer the lightest route that is likely to produce a usable shortlist.

根据用户需求选择渠道：

官方池导出数据 -> 标准化或手动整理，然后完善资料
广泛挖掘 -> 平台搜索渠道
竞品调研 -> 先从竞品内容中收集创作者候选者
内容角度挖掘 -> 内容优先渠道
已知账号 -> 直接完善渠道

如果用户没有特殊限制，优先选择最轻量化且能生成可用筛选名单的渠道。

2. Collect platform data

2. 收集平台数据

Do not scrape raw platform data directly from this skill unless the platform skill is missing something critical.

Use:

```
skills/20-research/tiktok-research
```

skills/20-research/instagram-account-research

```
skills/20-research/x-research
```

The expected input to this skill is normalized profile data or normalized profile-like exports from those platform skills.

The lead builder can also accept richer candidate-like datasets when they already include fields such as:

```
route
```
```
topicFit
```
```
audienceFit
```
```
sourceEvidence
```
engagement proxies
comment or commerce evidence

除非平台技能缺少关键信息，否则不要通过此技能直接抓取原始平台数据。

使用以下技能：

```
skills/20-research/tiktok-research
```

skills/20-research/instagram-account-research

```
skills/20-research/x-research
```

此技能的预期输入是来自这些平台技能的标准化资料数据或类资料导出数据。

线索构建工具也可接收更丰富的候选数据集，只要其包含以下字段：

```
route
```
```
topicFit
```
```
audienceFit
```
```
sourceEvidence
```
互动指标
评论或电商验证信息

3. Build a unified lead table

3. 构建统一线索表格

Convert normalized profile datasets into one shared lead schema:

bash

node ${CLAUDE_SKILL_DIR}/scripts/build_creator_leads.mjs \
  --inputs <work-folder>/.postplus/tiktok-profiles.json,<work-folder>/.postplus/instagram-profiles.json,<work-folder>/.postplus/x-profiles.json \
  --output <work-folder>/.postplus/creator-leads.json

将标准化资料数据集转化为统一的线索 schema：

bash

node ${CLAUDE_SKILL_DIR}/scripts/build_creator_leads.mjs \
  --inputs <work-folder>/.postplus/tiktok-profiles.json,<work-folder>/.postplus/instagram-profiles.json,<work-folder>/.postplus/x-profiles.json \
  --output <work-folder>/.postplus/creator-leads.json

4. Extract contact signals

4. 提取联系信号

Only use public signals already visible in platform data:

email in bio / signature / description
website or link-in-bio

Do not guess email addresses.

bash

node ${CLAUDE_SKILL_DIR}/scripts/extract_contact_signals.mjs \
  --input <work-folder>/.postplus/creator-leads.json \
  --output <work-folder>/.postplus/creator-leads-enriched.json

仅使用平台数据中已有的公开信号：

简介/签名/描述中的邮箱
网站或主页链接（link-in-bio）

不要猜测邮箱地址。

bash

node ${CLAUDE_SKILL_DIR}/scripts/extract_contact_signals.mjs \
  --input <work-folder>/.postplus/creator-leads.json \
  --output <work-folder>/.postplus/creator-leads-enriched.json

5. Score and shortlist

5. 评分并筛选

Before drafting, score and filter the lead list when the user wants a shortlist rather than the full dataset.

Important:

treat score as a first-pass ranking, not final truth
apply campaign-specific overrides for market fit, style fit, and creator type
if the dataset does not support those checks, call them out explicitly
do not add deeper evidence collection unless the user request or result quality justifies it

bash

node ${CLAUDE_SKILL_DIR}/scripts/score_creator_leads.mjs \
  --input <work-folder>/.postplus/creator-leads-enriched.json \
  --brief <work-folder>/.postplus/brand-brief.json \
  --platforms tiktok,instagram,x \
  --output <work-folder>/.postplus/creator-leads-scored.json

node ${CLAUDE_SKILL_DIR}/scripts/shortlist_creator_leads.mjs \
  --input <work-folder>/.postplus/creator-leads-scored.json \
  --min-score 45 \
  --top 20 \
  --output <work-folder>/.postplus/creator-leads-shortlist.json

在起草外联内容前，如果用户需要筛选名单而非完整数据集，需对线索列表进行评分和筛选。

重要提示：

将评分视为首轮排名，而非最终结论
根据活动需求对市场适配性、风格适配性和创作者类型进行自定义调整
如果数据集不支持这些检查，需明确说明
除非用户需求或结果质量要求，否则不要添加更深入的验证信息收集

bash

node ${CLAUDE_SKILL_DIR}/scripts/score_creator_leads.mjs \
  --input <work-folder>/.postplus/creator-leads-enriched.json \
  --brief <work-folder>/.postplus/brand-brief.json \
  --platforms tiktok,instagram,x \
  --output <work-folder>/.postplus/creator-leads-scored.json

node ${CLAUDE_SKILL_DIR}/scripts/shortlist_creator_leads.mjs \
  --input <work-folder>/.postplus/creator-leads-scored.json \
  --min-score 45 \
  --top 20 \
  --output <work-folder>/.postplus/creator-leads-shortlist.json

6. Generate outreach drafts

6. 生成外联草稿

Use a structured brief to generate a draft set.

Minimal brief:

json

{
  "brandName": "Example Brand",
  "productName": "Example Product",
  "niche": "creators relevant to the product, audience, or use case",
  "offer": "a creator partnership or product collaboration",
  "whyYou": "your content feels relevant to this product and audience",
  "cta": "Would you be open to hearing more?",
  "signature": "Name | Brand"
}

Then generate drafts:

bash

node ${CLAUDE_SKILL_DIR}/scripts/generate_outreach_drafts.mjs \
  --leads <work-folder>/.postplus/creator-leads-shortlist.json \
  --brief <work-folder>/.postplus/brand-brief.json \
  --output <work-folder>/.postplus/creator-outreach-drafts.json

使用结构化简报生成草稿集。

极简简报：

json

{
  "brandName": "Example Brand",
  "productName": "Example Product",
  "niche": "creators relevant to the product, audience, or use case",
  "offer": "a creator partnership or product collaboration",
  "whyYou": "your content feels relevant to this product and audience",
  "cta": "Would you be open to hearing more?",
  "signature": "Name | Brand"
}

然后生成草稿：

bash

node ${CLAUDE_SKILL_DIR}/scripts/generate_outreach_drafts.mjs \
  --leads <work-folder>/.postplus/creator-leads-shortlist.json \
  --brief <work-folder>/.postplus/brand-brief.json \
  --output <work-folder>/.postplus/creator-outreach-drafts.json

Output Expectations

输出预期

Return:

a shortlist of creators worth contacting
fit scores and fit reasons that explain ordering
contact signals with explicit provenance
explicit notes on missing evidence when market fit or style fit could not be verified from the dataset
risks:
- no email found
- only link-in-bio found
- niche mismatch
- weak audience scale
- uncertain target-market fit
- uncertain style or cultural fit
outreach drafts ready for review or Gmail draft creation

需返回：

值得联系的创作者筛选名单
适配分数和排序理由
带有明确来源的联系信号
当无法从数据集中验证市场适配性或风格适配性时，需明确说明缺失的验证信息
风险提示：
- 未找到邮箱
- 仅找到主页链接（link-in-bio）
- 领域不匹配
- 受众规模较小
- 目标市场适配性不确定
- 风格或文化适配性不确定
可直接用于审核或创建Gmail草稿的外联草稿

Decision Rules

决策规则

If the user asks for generic creator discovery, choose the discovery route first instead of defaulting to one search pattern.
If the user already has normalized profile datasets, jump straight to lead building.
If the user has official marketplace exports or a seller-center shortlist, treat that as a first-class input.
If the user wants competitor-based discovery, collect creator candidates from competitor assets before running enrichment.
If the user does not explicitly ask for competitor, comment, official-pool, or commerce-layer evidence, do not force those layers into the default pass.
If the user wants cold email or partnership prep, always extract contact signals before drafting.
Default to drafting, not sending.

如果用户要求通用创作者挖掘，应先选择挖掘渠道，而非默认采用单一搜索模式。
如果用户已有标准化资料数据集，直接进入线索构建环节。
如果用户有官方市场导出数据或商家中心筛选名单，将其作为优先输入。
如果用户需要基于竞品的挖掘，先从竞品资产中收集创作者候选者，再进行资料完善。
如果用户未明确要求竞品、评论、官方池或电商层验证信息，不要将这些层强制加入默认流程。
如果用户需要冷邮件或合作准备，在起草前务必提取联系信号。
默认仅生成草稿，不发送。

Notes

注意事项

This skill is the cross-platform layer above individual platform research skills.
Gmail sending should stay in a separate delivery step.

此技能是独立平台调研技能之上的跨平台层。
Gmail发送应保留在单独的交付环节。