Content Adaptation for Xiaohongshu, Douyin, Bilibili, WeChat

v3.0.3

写完一篇内容,想要发到小红书、抖音、B站、公众号还要分别适配?通过渐进式确认流程,帮助你快速生成多平台适配文案,确保每个平台的内容都符合字数和风格规范。触发词:多平台分发、发到XX平台。

1· 184·0 current·0 all-time

Install

OpenClaw Prompt Flow

Install with OpenClaw

Best for remote or guided setup. Copy the exact prompt, then paste it into OpenClaw for jrnonr/viral-note-distributor.

Previewing Install & Setup.
Prompt PreviewInstall & Setup
Install the skill "Content Adaptation for Xiaohongshu, Douyin, Bilibili, WeChat" (jrnonr/viral-note-distributor) from ClawHub.
Skill page: https://clawhub.ai/jrnonr/viral-note-distributor
Keep the work scoped to this skill only.
After install, inspect the skill metadata and help me finish setup.
Use only the metadata you can verify from ClawHub; do not invent missing requirements.
Ask before making any broader environment changes.

Command Line

CLI Commands

Use the direct CLI path if you want to install manually and keep every step visible.

OpenClaw CLI

Bare skill slug

openclaw skills install viral-note-distributor

ClawHub CLI

Package manager switcher

npx clawhub@latest install viral-note-distributor
Security Scan
VirusTotalVirusTotal
Benign
View report →
OpenClawOpenClaw
Benign
high confidence
Purpose & Capability
Name/description (multi‑platform copy adaptation) match the included SKILL.md, agent templates, style/hook references, and the small formatting script. There are no unexpected required binaries, env vars, or config paths that would be unrelated to content generation.
Instruction Scope
Runtime instructions describe a progressive two‑round confirmation flow, per‑platform generation, and validation steps. The agents read local reference files (platform styles, hooks, cover guide) and user input; they do not instruct reading system files, environment variables, or transmitting data to unknown endpoints. The only external interaction noted is an example suggestion to pass cover_image_prompt to an image generation tool (DALL‑E/Stable Diffusion) — this is optional and described as a separate tool.
Install Mechanism
No install spec; this is instruction‑first with one small helper script (format_output.py) used for formatting JSON results. There are no downloads, package installs, or archive extracts in the skill bundle.
Credentials
The skill requires no environment variables or credentials. References to external image generators are examples only and would require separate credentials if the user chooses to use them; the skill itself does not request those.
Persistence & Privilege
Skill flags are default (always: false, user-invocable true). There is no code that writes to other skills' configs or requests permanent presence or elevated privileges.
Assessment
This skill appears internally consistent and low‑risk: it only generates platform‑specific copy, reads the included style/hook reference files, and formats outputs. Before using: (1) do not paste secrets or private credentials into the content you want adapted — generated prompts (especially cover image prompts) could be sent to third‑party image services if you choose to use them; (2) note the skill does not auto‑post to platforms (no API credentials requested) — if you want posting automation, expect additional env vars and re‑review those for proportionality; (3) review the provided references and sample outputs to ensure style/tone and hashtag behaviour meet your expectations; (4) if you enable cover image generation, confirm what external image service you use and its privacy/credential implications.

Like a lobster shell, security has layers — review code before you run it.

latestvk975mt8k0gwkqh0b77cah2k2m584hrzc
184downloads
1stars
6versions
Updated 2w ago
v3.0.3
MIT-0

爆款内容多平台分发 Skill

核心设计原则

渐进式披露(Progressive Disclosure):绝不一次性暴露所有上下文。每一轮只释放用户当前需要确认的信息,等用户确认后再推进。

用户输入(可能碎片化)
     ↓
【Round 1】暴露:话题 + 基调 + 受众 + 角度
           用户确认 / 修改方向
           ↓
【Round 2】暴露:关键信息点(key_points)供用户增删
           用户确认 / 补充
           ↓
【生成阶段】各平台并行生成 → 汇总输出

Round 1 — 方向确认

判断输入类型

输入字数状态操作
< 20字信息不足追问补充
20-200字,碎片化需要重写先重写整理再继续
> 200字,结构化直接进入 Round 1提炼摘要

暴露字段

字段来源是否暴露
topic提炼
angle提炼
target_audience推断
mood推断
focus提炼/推断(踩坑/推荐/深度分析/情绪宣泄)
platforms关键词识别
key_points提炼❌ 隐藏
original_content原始输入❌ 隐藏(已蒸馏)

Round 1 Prompt 模板

## Round 1 · 方向确认

我理解你想说:

📌 话题:{topic}
🎯 角度:{angle}
👥 受众:{target_audience}
💡 基调:{mood}
🎯 侧重:{focus}(踩坑/推荐/深度分析/情绪宣泄)
📺 平台:{platforms}

请确认这个方向对不对,或者告诉我需要怎么调整。

状态转移

  • 用户确认("对的/可以/继续")→ 进入 Round 2
  • 用户修改 → 更新字段,重新暴露 Round 1
  • 信息不足 → 追问补充

Round 2 — 关键点确认

触发条件

Round 1 已确认

生成 Key Points

基于用户原始输入 + Round 1 确认的方向,提取 3-5 个关键信息点

暴露字段

在 Round 1 基础上,追加暴露 key_pointskey_points[].source 隐藏)

Round 2 Prompt 模板

## Round 2 · 关键点确认

话题:{topic}(已确认)
角度:{angle}(已确认)

📋 我提炼的关键信息点:

  1. {key_point_1}
  2. {key_point_2}
  3. {key_point_3}

请确认这些点是否准确,你可以:
- 删除不需要的点
- 添加遗漏的信息
- 修改某条的表达

确认后我开始生成各平台内容。

状态转移

  • 用户确认("对的/可以/生成吧")→ 进入生成阶段
  • 用户增删修改 → 更新 key_points,重新暴露 Round 2
  • 用户要求改方向 → 退回 Round 1

生成阶段

并行派生平台 Agent

platforms = [用户指定的平台列表]
for platform in platforms:
    spawn platform-agent(topic, key_points, angle, focus, mood, platform, with_cover)

子 Agent 任务输入

{
  "task": "platform_adaptation",
  "platform": "平台名",
  "topic": "话题",
  "key_points": ["关键点1", "关键点2"],
  "angle": "角度",
  "target_audience": "受众",
  "mood": "基调",
  "focus": "侧重",
  "with_cover": true/false,
  "platform_style": "(从 references/platform-styles.md 读取)",
  "hooks_library": "(从 references/hooks-library.md 读取)"
}

校验与输出

主 Agent 校验各平台输出的字段完整性和字数合规性,通过后删除所有字段名,转为可复制纯文本输出给用户。

用户面向输出模板

📰 公众号

【标题】
AI时代,程序员真的要早做打算了

【正文】
(完整正文段落,可直接复制粘贴到公众号后台)

#程序员 #AI时代

---
发布建议:建议周二或周四晚间发布;阅读时长约5分钟

JSON 仅在 Agent 内部流转,不展示给用户。


意图识别与平台匹配

平台关键词

单平台:小红书/小红书帖子/xhs、抖音/抖音文案/短视频、B站/bilibili、公众号/微信公众号

全平台:多平台/全平台/4个平台

判断逻辑

扫描平台关键词
     ↓
找到"全平台"关键词? → 平台列表 = 全部4个
找到具体平台词? → 只加入对应平台
什么都没找到? → 询问用户

特殊处理

用户跳过确认直接要求生成

用户输入很详细但没确认就要求生成,仍走 Round 1 → Round 2 流程。

快速路径(用户已明确所有信息)

满足以下全部条件时,Round 1 + Round 2 合并展示,用户一次确认后直接生成:

  • 输入字数 > 200字
  • 包含明确话题 + 受众 + 至少3个关键点
  • 平台明确

用户说"帮我发到全平台,内容是OpenClaw体验"

信息不足(<20字),触发 Round 1 追问:"先补充一下想说的核心观点是什么?"


文件结构

viral-note-distributor/
├── SKILL.md                    # 本文件:技能入口 + 完整 workflow
├── agents/
│   ├── master.md               # 主控Agent:Round推进、状态转移、子Agent调度
│   └── platform-agent.md       # 平台生成Agent模板
└── references/
    ├── platform-styles.md      # 各平台图文风格定义
    ├── hooks-library.md        # 爆款钩子库
    └── cover-image-guide.md    # 封面图生成指南

Comments

Loading comments...