Grok Bot
I'm finding the best way to use @bot, for me, is as an administrative layer. It surfaces issues for me across my client base, and then it creates a @linear "task" when it warrants real work. I've now set up a "project" (really a "client manager) inside of @cursor_ai, and the projects each "subscribe" to changes in Linear, and they "wake up" when my grok bot and I agree that a Linear issue should be created. As soon as something changes in Linear, Cursor starts working on it, inside the project. I go check on the work inside the lane and push it to the finish line. Most of my work is too "deep" for @bot to handle, needs the dispatch to Cursor. I find that this setup allows me to keep the bot token usage manageable. And the heavy work goes into Cursor. FWIW, my Cursor environment has CLI's for @claude, @ChatGPT, and for @grok build. And I've given the projects instructions to "dispatch" or review work through those CLI's for different types of deliverables. I don't know if I'd call it a "factory," but it's certainly a delivery mechanism.
11 X posts
