ZeeLin Twitter/X AutoPost

ZeeLin Twitter/X 自动发推 + 回关 + 涨粉运营 — 通过 openclaw browser / Browser Relay 操作网页版 Twitter/X，无需 API Key。用户先在自己的浏览器登录并挂上 Relay，Agent 负责撰写推文并发布、一键回关粉丝、蓝V互关（认证关注者回关）...

Security Scan

The skill largely does what it says (automates X via the OpenClaw Browser Relay), but the runtime instructions rely on absolute local paths and reference external scripts that are not included, and they instruct the agent to exec shell scripts — this mismatch and the exec behavior warrant caution.

Purpose & Capability

The skill's stated purpose (automating Twitter/X via the browser relay without API keys) matches the included tweet.sh and follow_back.sh which use openclaw browser. However SKILL.md examples use hard-coded absolute paths under /Users/youke/.openclaw/workspace/skills/... and also reference scripts in a different skill folder (twitter-x-operations/scripts/*.sh) that are not present in this package. Those external references are incoherent and could cause the agent to try to exec arbitrary, absent, or user-local scripts.

Instruction Scope

Instructions explicitly tell the agent to run exec commands that call local bash scripts. The scripts capture and parse full browser snapshots and may echo fragments of page content for diagnostics (which could include private information). The instruction set also directs the agent to search for and comment on other users' posts (automated interactions), which is within the described purpose but has behavioral and policy risks. The use of exec to run absolute path scripts (some not bundled) broadens runtime scope beyond the skill bundle.

Install Mechanism

No install spec — instruction-only plus two included shell scripts. Nothing is downloaded or written during install. This is low-risk from an installation vector perspective.

Credentials

The skill does not declare any required environment variables or credentials, which is consistent with a browser-relay approach. The tweet.sh optionally uses OPENCLAW_CLI if set (not declared). The more important concern is that browser snapshots (page content) are read and parsed; while necessary for operation, snapshots could contain sensitive data.

Persistence & Privilege

always is false and the skill is user-invocable. It does not request persistent elevated privileges or modify other skills' configs in the package. Autonomous invocation (default allowed) is normal and not by itself flagged here.

What to consider before installing

This skill appears to implement browser-based posting and follow-back via included shell scripts, but exercise caution before installing/using it. Key points to check before proceeding: 1) Inspect the included scripts (tweet.sh, follow_back.sh) yourself — they run commands via 'openclaw browser' and may print browser snapshots; verify you are comfortable with that. 2) SKILL.md examples call absolute paths (/Users/youke/...) and reference scripts under twitter-x-operations (follow_back_verified.sh, comment.sh) which are not bundled — ensure those external scripts are legitimate and present on your machine before executing. 3) Running exec to invoke shell scripts executes local code — do not run if you don't trust the source. 4) Automated commenting/following can risk account suspension (rate limits / ToS violations) and may post or reveal private page content; limit volumes and test with a throwaway account first. 5) Prefer running in a controlled environment and confirm the OpenClaw Relay is the browser session you intend to control. If you are unsure about the missing referenced scripts or absolute-path usage, treat this package as untrusted until you can verify or sanitize the commands.

ZeeLin Twitter/X 自动发推 + 回关 + 涨粉运营 🐦

通过 openclaw browser / Browser Relay 操作网页版 Twitter/X：支持发推（撰写 + 发布）、回关（粉丝列表一键回关）、蓝V互关（认证关注者回关）、深度评论、以及在求关注/互关类帖子下主动打招呼。用户先在自己的浏览器登录并挂上 Relay，Agent 用脚本完成操作，无需 API Key。

飞书下：发推/评论时优先直接发一个 exec；回关/蓝V互关默认带较长超时，减少 request timed out。

概述

• 发推：Agent 撰写推文 → 使用已登录且挂上 Relay 的 X 标签页 → Agent 输入并发布
• 回关：在关注者列表中自动点击回关
• 蓝V互关：在认证关注者列表中自动回关
• 深度评论：对指定帖子写评论并发布
• 涨粉互动：主动寻找 follow for follow / f4f / 互关 / 求关注 类帖子，在下面自然打招呼，增加曝光与涨粉

---

何时触发

发推

• 「帮我发一条推特/推文」
• 「自动在 X 上发帖」
• 「围绕某个热点写一条推特并发布」
• 「每天 XX 点自动发推」「设置定时推特」

回关 / 蓝V互关

• 「回关」「帮我回关」「回关推特」
• 「有人关注我了」「关注者列表回关」
• 「蓝V互关」「认证关注者回关」「蓝V回关」

互动 / 涨粉

• 「帮我评论这条推文」
• 「在涨粉推文下打招呼」
• 「帮我找求关注的帖子互动」
• 「今天做下推特运营」

---

回关与蓝V互关（必须用 exec，不要用 browser 逐步点）

用户说「回关 / 蓝V互关 / 认证关注者回关」时，第一反应：用 exec 执行脚本，不要自己用 browser 打开页面、snapshot、click。

普通回关

{"tool": "exec", "args": {"command": "bash /Users/youke/.openclaw/workspace/skills/zeelin-twitter-web-autopost/scripts/follow_back.sh Gsdata5566 https://x.com 5", "timeout": 90000}}

蓝V互关 / 认证关注者回关

优先调用已合并进本 skill 的运营脚本：

{"tool": "exec", "args": {"command": "bash /Users/youke/.openclaw/workspace/skills/twitter-x-operations/scripts/follow_back_verified.sh Gsdata5566 https://x.com 5", "timeout": 90000}}

• 飞书下默认建议 5 人，更稳
• 执行完后根据输出回报：「已回关 X 人」/「已蓝V互关 X 人」

---

总体流程（发推）

Step 1：确认用户的 X 网址

首次使用时，询问用户：

「请提供你访问 X/Twitter 的网址（例如 https://x.com 或 https://twitter.com）」

记住用户提供的 BASE_URL，后续所有操作基于它。不要自行假设网址。

Step 2：先准备已登录的 Relay 标签页

1. 让用户在自己的 Chrome 中打开用户提供的 X 网址并登录
2. 让用户在该标签页挂上 OpenClaw Browser Relay，确认 Badge 为 ON
3. 后续一律通过 openclaw browser / Relay 操作该标签页
4. 不要默认改用 agent-browser，因为它是独立浏览器，不共享用户现有登录 session
5. 只有在用户明确要走独立浏览器，且已经保存过登录态时，才考虑 agent-browser state load

Step 3：撰写推文内容

• 用户给了完整文案 → 直接使用
• 用户给了主题/方向 → 用模型生成（≤240 字符）
• 用户要求全自动 → 自行选热点并撰写

Step 4：发布推文

优先使用现成脚本：

bash /Users/youke/.openclaw/workspace/skills/zeelin-twitter-web-autopost/scripts/tweet.sh "推文内容" https://x.com

或在需要时用浏览器流程补救。

Step 5：回报结果

告诉用户：

• 发布成功/失败
• 推文全文
• 推文 URL（如果能拿到）

---

深度评论（用户给帖子链接）

1. 用户给出一条 X 帖子链接
2. 先写一条自然、有信息量、有趣的评论
3. 确认后执行：

{"tool": "exec", "args": {"command": "bash /Users/youke/.openclaw/workspace/skills/twitter-x-operations/scripts/comment.sh \"评论内容\" \"帖子URL\" https://x.com", "timeout": 60000}}

---

涨粉帖打招呼（主动互动）

目标：在 follow for follow / f4f / 互关 / 求关注 / follow back 类帖子下友好评论，提升曝光和回关率。

推荐流程

1. 用 X 搜索页搜索相关关键词
2. 找 3～5 条帖子即可，不要一次太多
3. 每条写略有变化的友好评论，例如：
   • 「刚看到，已 fo，欢迎回关～」
   • 「有同感，先关注啦，常互动」
   • 「已支持，互相关注一起涨」
4. 逐条执行评论脚本：

{"tool": "exec", "args": {"command": "bash /Users/youke/.openclaw/workspace/skills/twitter-x-operations/scripts/comment.sh \"评论内容\" \"https://x.com/xxx/status/123\" https://x.com", "timeout": 60000}}

5. 最后汇总告诉用户已互动多少条

注意： 单次建议 3～5 条，避免太像机器刷评。

---

定时发布

当用户要求定时发推时，使用 openclaw cron。

询问参数

• 频率：每天 / 每周 / 一次性
• 时间：几点
• 时区：默认 Asia/Shanghai
• 内容策略：固定文案 / 每次自动写新的
• 语言：中文 / 英文

创建示例

openclaw cron add \
  --name "daily-tweet" \
  --description "每天自动撰写并发布推文" \
  --cron "0 10 * * *" \
  --tz "Asia/Shanghai" \
  --message "请执行 zeelin-twitter-web-autopost skill：用用户的X网址打开推特，撰写一条英文AI热点推文并发布，不要与之前重复"

---

exec 命令速查

操作	命令
发推	bash /Users/youke/.openclaw/workspace/skills/zeelin-twitter-web-autopost/scripts/tweet.sh "推文内容" https://x.com
回关	bash /Users/youke/.openclaw/workspace/skills/zeelin-twitter-web-autopost/scripts/follow_back.sh Gsdata5566 https://x.com 5
蓝V互关	bash /Users/youke/.openclaw/workspace/skills/twitter-x-operations/scripts/follow_back_verified.sh Gsdata5566 https://x.com 5
评论	bash /Users/youke/.openclaw/workspace/skills/twitter-x-operations/scripts/comment.sh "评论内容" "帖子URL" https://x.com

以上均通过 exec 执行；回关/蓝V互关建议 timeout: 90000，评论建议 timeout: 60000。

---

安全与风控

• 不要自动输入密码，登录由用户自己完成
• 这个 skill 默认依赖 openclaw browser / Browser Relay，不要切到 agent-browser 去要求用户重新登录
• 不发违法、仇恨、违规内容
• 发帖频率建议每天不超过 3–5 条
• 主动互动单次建议 3–5 条，避免刷屏
• 失败最多重试 1–2 次

---

TL;DR

• 用户说「发推」→ 发推脚本
• 用户说「回关」→ follow_back.sh
• 用户说「蓝V互关」→ follow_back_verified.sh
• 用户说「评论这条」→ comment.sh
• 用户说「找涨粉帖互动」→ 搜 3～5 条 + 逐条 comment.sh