← 返回 Skills 市场
103
总下载
0
收藏
0
当前安装
1
版本数
在 OpenClaw 中安装
/install cell-x-content-mentor
功能描述
X/Twitter content strategy skill for Chinese creators, indie builders, and expert accounts. Use when Codex needs to decide whether an idea belongs on X, turn...
使用说明 (SKILL.md)
X Content Mentor
Overview
把 X 上的内容工作,当成一个系统,不当成“随手写条推文”。
这个 Skill 主要解决五类问题:
- 不知道发什么
- 知道要发什么,但不知道怎么写成 X 原生内容
- 已经写了,想判断问题在哪里
- 账号有内容但增长停滞
- 想做一次结构化账号诊断并沉淀历史策略
它不是通用 tweet 生成器,而是一个路由式内容策略 Skill。
Quick Start
- 先判断当前属于哪个场景:选题、写作、审阅、增长,还是账号诊断。
- 锁定受众、语言、目标动作,不要跳过。
- 只读取当前场景需要的 reference,不一次性全读。
- 如果真正的问题不在 X 写法,明确指出并切给更合适的 Skill。
- 结束时给一个可直接执行的下一步。
Default Contract
默认采用以下约定,除非用户另有说明:
- 输出解释用中文
- 如果目标平台是英文 X,正文内容用英文,解释与策略仍可用中文
- 优先判断“值不值得发”和“该怎么发”,再讨论具体 wording
- 不承诺爆款,不把历史算法经验当作绝对规律
- 如做过账号诊断,优先利用已有沉淀
Workflow
Step 1: Scope the Input
先判断用户手上有什么:
- 一个项目、经验、观察,想发出去
- 一条已经写好的短帖或 Thread
- 一个增长困惑
- 一个要分析的账号
同时锁定四个变量:
- 目标受众是谁
- 想要什么结果:增长、建认知、招募、导流、产品验证
- 发中文、英文,还是双语
- 当前账号大致处于什么阶段
如果这些信息缺得太厉害,不要假装确定。
Step 2: Route to the Right Scene
按下面路由处理:
| 场景 | 什么时候进入 | 按需读取 |
|---|---|---|
| A. 选题与方向 | 用户不知道发什么,或不知道某个项目怎么转成 X 内容 | references/topic-engine.md |
| B. 写短帖 / Thread | 用户已经有方向,想起草内容 | references/writing-workshop.md |
| C. 审阅现有内容 | 用户给了现成帖子,需要诊断或改写 | references/review-scorecard.md + references/writing-workshop.md |
| D. 增长与策略 | 用户问涨粉、定位、分发、导流 | references/growth-engine.md |
| E. 账号诊断 | 用户要系统看账号数据与内容结构 | references/account-diagnosis.md + references/review-scorecard.md |
Step 3: Apply Scene Rules
Scene A: Topic and Direction
- 把用户素材先转成“可在 X 上成立的角度”,再写帖
- 优先判断它适合:
- short post
- thread
- series
- or not X at all
- 如果主题 promise 太弱、材料太薄、目标不清,先指出,不硬写
Scene B: Writing
- 先决定短帖还是 Thread
- 先给 2-4 个 hook 方向,再展开成正文
- 对英文 X 内容:
- 正文用英文写
- 不做中文句子直译式英文
- Thread 默认要有:
- opening promise
- structured middle
- 收束或 CTA
Scene C: Review
- 先判断这是题的问题、hook 的问题、结构的问题,还是账号语境的问题
- 默认先给诊断,再决定是否给改写版
- 区分:
- 同一路径优化
- 换一种路径重写
Scene D: Growth
- 先判断阶段:冷启动、验证期、放大期
- 不要一上来讲宏大增长理论
- 优先指出当前最限制增长的一件事:
- 定位太散
- 题材不对
- 写法抓力不足
- 互动动作缺失
- 没有把高表现内容扩成可复利资产
Scene E: Account Diagnosis
- 先判断是否已有历史数据
- 如无数据,按可获得性分级采样
- 输出不只给结论,还要沉淀到用户数据目录
- 如样本不足,明确标注
Step 4: Hand Off When Needed
以下情况不要硬留在本 Skill 内:
- 题目本身不成立,先交给
$content-strategy-diagnosis - 只是前两句抓力弱,先交给
$hook-optimizer - 结构对了但风格太平,交给
$celf-style-writer - 想找对标账号或案例,交给
$benchmark-filter
User Data Persistence
账号诊断默认把用户历史信息保存在:
user-data/x-content-mentor/\x3Cusername>/
├── profile.md
├── posts_\x3Cdate>.json
├── posts_\x3Cdate>.md
├── report_\x3Cdate>.md
└── strategy.md
规则:
- 如果已有
strategy.md,默认静默读取,用来提高后续建议的针对性 - 如果上次诊断已超过 30 天,提醒重新诊断
- 如果只有部分样本,照做,但把样本量不足写进结论
Output Format
标准输出默认参考 assets/x-strategy-pack-template.md。
账号诊断默认参考 assets/account-diagnosis-template.md。
Hard Rules
Do not:
- 把所有请求都简化成“我给你写三条 tweet”
- 用过度确定的算法数字伪装成实时事实
- 忽略受众、语言、账号阶段
- 只给漂亮 wording,不给结构性判断
- 为了看起来专业,把简单动作说得很复杂
Always:
- 先判断场景
- 先说最关键的问题或机会
- 明确区分高置信度共识与阶段性经验
- 在账号诊断里留下可复用的策略沉淀
- 让建议能跟仓库内其他 Skill 接力
Resource Map
- references/topic-engine.md
- 读这个文件,把项目、经验、观察转成适合 X 的题材与角度。
- references/writing-workshop.md
- 读这个文件,处理短帖、Thread、hook 与语言适配。
- references/growth-engine.md
- 读这个文件,判断账号阶段、分发动作与增长策略。
- references/review-scorecard.md
- 读这个文件,诊断现有内容、做复盘与反模式排查。
- references/account-diagnosis.md
- 读这个文件,做数据采样、诊断结构与用户数据留存。
- assets/x-strategy-pack-template.md
- 标准场景的输出模板。
- assets/account-diagnosis-template.md
- 账号诊断输出模板。
安全使用建议
This skill appears to do what it says: provide structured topic, writing, review, growth, and account-diagnosis guidance using the included reference files and templates. Before installing, consider: 1) Data persistence — it will write reports and posts to user-data/x-content-mentor/<username>/ and will, by default, silently read an existing strategy.md to inform suggestions; if you want explicit consent each time, ask for that behavior to be changed. 2) Privacy of uploaded analytics — don't provide any sensitive tokens, full analytics exports, or private credentials unless you're comfortable with them being stored. 3) Browser scraping suggestions in the docs imply it might fetch public profile data; confirm what data sources the agent will actually access in your environment. If those points are acceptable, the skill is coherent with its stated purpose.
功能分析
Type: OpenClaw Skill
Name: cell-x-content-mentor
Version: 0.1.0
The skill bundle is a comprehensive framework for X/Twitter content strategy, focusing on topic selection, writing, and account diagnosis. It operates by processing user-provided data and storing strategy reports in a local directory (`user-data/x-content-mentor/`) for persistence, which is consistent with its stated purpose. No evidence of malicious code, data exfiltration to external endpoints, or harmful prompt injection was found across the SKILL.md or reference files.
能力评估
Purpose & Capability
Name, description, and included reference files (topic-engine, writing-workshop, growth-engine, review-scorecard, account-diagnosis, templates) align with a content-strategy/mentoring skill. There are no unrelated environment variables, binaries, or install steps requested.
Instruction Scope
Runtime instructions are limited to routing content tasks, applying writing/review/growth heuristics, and reading the bundled reference files. However, the SKILL.md explicitly instructs the skill to persist and read user-specific data under user-data/x-content-mentor/<username>/ and states it will '默认静默读取' (silently read) an existing strategy.md. Reading/writing the skill's own data is coherent with the purpose but the silent-read behavior is a privacy/consent concern.
Install Mechanism
No install spec or code is present; this is an instruction-only skill so nothing is downloaded or executed on install. This reduces install-time risk.
Credentials
The skill requests no environment variables, no credentials, and no config paths beyond its own user-data directory. That is proportionate for a content-strategy/documentation-style skill.
Persistence & Privilege
The skill asks to persist account diagnostics and strategy files in user-data/x-content-mentor/<username>/. Persisting its own data is reasonable for the stated functionality, but the default behavior to silently read existing strategy.md (without explicit per-run confirmation) is noteworthy. always:false and no elevated privileges are set.
如何使用
- 确保已安装 OpenClaw(本地或 Docker 部署)
- 在对话框中输入安装命令:
/install cell-x-content-mentor - 安装完成后,直接呼叫该 Skill 的名称或使用
/cell-x-content-mentor触发 - 根据 Skill 的参数说明提供必要输入,即可获得结构化输出
版本历史
v0.1.0
Initial public release
元数据
常见问题
X 内容策略 是什么?
X/Twitter content strategy skill for Chinese creators, indie builders, and expert accounts. Use when Codex needs to decide whether an idea belongs on X, turn... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 103 次。
如何安装 X 内容策略?
在 OpenClaw 或 Claude Code 对话框中运行命令「/install cell-x-content-mentor」即可一键安装,无需额外配置。
X 内容策略 是免费的吗?
是的,X 内容策略 完全免费,采用 MIT-0 许可证,可自由下载、安装和使用。
X 内容策略 支持哪些平台?
X 内容策略 跨平台运行,可在任意部署了 OpenClaw / Claude Code 的环境中使用(cross-platform)。
谁开发了 X 内容策略?
由 cellinlab(@cellinlab)开发并维护,当前版本 v0.1.0。
推荐 Skills