Skip to content

chenjin-cmd/agent-skills-launch-pack_

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

1 Commit
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

起号专家 Agent Skills / Account Launch Agent Skills

一组面向中文内容运营场景的 agent skills,覆盖微信公众号、小红书、抖音和 X/Twitter 的合规起号、定位、选题、内容计划和复盘。

A bilingual agent skills pack for Chinese content operators, covering compliant account launch planning, positioning, topic systems, content calendars, and review loops for WeChat Official Accounts, Xiaohongshu, Douyin, and X/Twitter.

语言 / Language


中文说明

Skills

Skill 用途
wechat-account-launch-expert 微信公众号起号、定位、选题库、文章简报、发布节奏和周复盘
xiaohongshu-account-launch-expert 小红书账号定位、笔记简报、内容日历、转化路径和复盘
douyin-account-launch-expert 抖音新号冷启动、观看理由、9 条视频实验、互动和复盘
x-twitter-cold-start-expert 中文 X/Twitter 冷启动、定位、回复区曝光、主贴转化和 7 天计划

安装

默认安装到 Codex skills 目录:

./install.sh --all

安装指定 skill:

./install.sh wechat-account-launch-expert

安装到自定义 agent skills 目录:

./install.sh --all --target ~/.codex/skills

也可以使用环境变量:

AGENT_SKILLS_DIR=~/.claude/skills ./install.sh --all

查看可安装的 skill:

./install.sh --list

仓库结构

.
├── skills/
│   ├── wechat-account-launch-expert/
│   ├── xiaohongshu-account-launch-expert/
│   ├── douyin-account-launch-expert/
│   └── x-twitter-cold-start-expert/
├── tools/
│   └── validate_skills.py
├── tests/
├── manifest.json
└── install.sh

每个 skill 文件夹都包含 SKILL.md,并可附带:

  • references/: 方法论、流程和检查清单。
  • agents/: 面向不同 agent UI 的展示信息。

验证

运行单元测试:

python3 -m unittest discover -s tests -v

运行发布包校验:

python3 tools/validate_skills.py .

校验器会检查:

  • skills/ 是否存在。
  • 每个 skill 是否有 SKILL.md
  • SKILL.md 是否包含 namedescription frontmatter。
  • skill 内容里是否出现常见本机路径、API key、token、cookie 或授权头。

兼容性

这个仓库使用通用的 skills/<skill-name>/SKILL.md 结构。默认安装路径面向 Codex,也可以通过 --targetAGENT_SKILLS_DIR 安装到其他支持 skills 的 agent 环境。

安全说明

这些 skills 只提供内容运营规划、诊断和复盘框架,不承诺涨粉、收益、播放量或爆款结果。涉及平台规则、广告能力、处罚边界、导流方式和收益门槛时,应以平台当前官方规则为准。

许可证

MIT License. See LICENSE.


English README

Skills

Skill Purpose
wechat-account-launch-expert WeChat Official Account launch planning, positioning, topic library, article briefs, publishing cadence, and weekly review
xiaohongshu-account-launch-expert Xiaohongshu account positioning, note briefs, content calendar, conversion path, and review loop
douyin-account-launch-expert Douyin cold start planning, viewing reasons, 9-video experiments, interaction design, and performance review
x-twitter-cold-start-expert Chinese X/Twitter cold start planning, positioning, reply-based discovery, post/thread conversion, and a 7-day execution plan

Install

Install all skills into the default Codex skills directory:

./install.sh --all

Install one skill:

./install.sh wechat-account-launch-expert

Install into a custom agent skills directory:

./install.sh --all --target ~/.codex/skills

You can also use an environment variable:

AGENT_SKILLS_DIR=~/.claude/skills ./install.sh --all

List available skills:

./install.sh --list

Repository Structure

.
├── skills/
│   ├── wechat-account-launch-expert/
│   ├── xiaohongshu-account-launch-expert/
│   ├── douyin-account-launch-expert/
│   └── x-twitter-cold-start-expert/
├── tools/
│   └── validate_skills.py
├── tests/
├── manifest.json
└── install.sh

Each skill folder contains SKILL.md and may also include:

  • references/: playbooks, workflows, and checklists.
  • agents/: display metadata for different agent UIs.

Validate

Run the unit tests:

python3 -m unittest discover -s tests -v

Run the package validator:

python3 tools/validate_skills.py .

The validator checks that:

  • skills/ exists.
  • Each skill has a SKILL.md.
  • Each SKILL.md includes name and description frontmatter.
  • Skill content does not contain common local paths, API keys, tokens, cookies, or authorization headers.

Compatibility

This repository uses the generic skills/<skill-name>/SKILL.md layout. The default install target is Codex, but --target and AGENT_SKILLS_DIR let you install the skills into other agent environments that support skill folders.

Safety Notes

These skills provide planning, diagnosis, and review frameworks for content operations. They do not promise followers, revenue, views, or viral outcomes. For platform rules, ad features, enforcement boundaries, external-link behavior, and monetization thresholds, use the current official platform rules as the source of truth.

License

MIT License. See LICENSE.

About

No description, website, or topics provided.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors