求职：摘要 + 清单机器人 + 投递队列

提示词

Adapt this outline. Do not treat it as a finished prompt. Fill every [BRACKET]. Do not apply, email, or message a recruiter unless I say send.

Three bots, one pipeline:

1. Digest — each morning, scan [BOARDS / ALERTS / INBOX] for roles matching [TITLE, LOCATION, MUST-HAVES, HARD NOS]. Return a short digest: company, role, why it fits, link, date seen. Skip anything already in the list file.
2. List bot — maintain /workspace/job-search/list.md. Columns: company, role, link, status (new / researching / ready / queued / sent / dead), notes. You may add and recategorize. You may not mark sent unless I confirm.
3. Apply queue — given a "ready" row, draft the tailored resume bullets and a short message. Show me the packet. Stop. After I send (or tell you to send), move the row to sent and log the date.

Rules: never invent a job that is not on a live listing. Never fabricate employment dates. Never auto-submit on LinkedIn/Indeed/etc.
Share templates from the source post (install if you use Grok Bot):
- digest: https://x.ai/bot/Nk-vzuWqTvqSed-G8-Za5
- list bot: https://x.ai/bot/uVNOsoe-iWf4ZOUdfgo5R
- apply queue: https://x.ai/bot/5RXN9P3CxnIIwgcmvVWEp
Start with the digest. Do not run the apply queue until the list file exists.

它是怎么搭起来的

Haseeb Mir（@HaseebMir91）发了一套三个模板，“来自我实际运行求职的方式”，附三个 x.ai/bot 分享链接：摘要、清单机器人、投递队列。GrokHub 后来镜像了相关页面；本站仍以他的帖子为准。

安装这三个分享，或者按下面的大纲自己重建。
按顺序串起来：摘要机器人写入候选职位，清单机器人是唯一事实来源，投递队列只负责起草。
提交仍由你来点。自动投递只会让你变成垃圾邮件发送者，然后被封号。

为什么它好用

求职这件事会死在一个塞满东西的聊天窗口里。三个座位正好对应真实的状态——看到了、在跟进、在起草——而且有了分享链接，你可以先试他的分工，再写自己的版本。

本案例收集自公开的 Grok Bot 社区，版权归原作者所有。提示词属于参考资料：运行前请先审阅内容，并且不要让 bot 超出你实际授予的权限行事。