Grok Build just landed in your terminal. It pulls live X data directly into your workflow, something no other AI CLI tool does right now.

You need to try this if you're running the Hermes agent. Stop using it from your CLI. This open-sourced GitHub repo turns Hermes into a fully interactive app. That's the difference between a tool you tolerate and one you actually use. https://github.com/fathah/hermes-desktop

How to NEVER hit usage limits with the new Opus 4.8 model (rapid-fire style): • Use plan mode MORE than actual coding (reduces errors and debugging later) • Insert this into your project instructions: "I am using Opus 4.8 for all tasks inside this project - while we build, proactively give me tips for avoiding hitting my rate limits • start NEW convos regularly - don't let context bloat get in the way of Opus 4.8's intelligence • Use dynamic workflows (just include "workflow" in your prompt) to build repeatable systems • Stick to Low -> Medium effort for tasks that don't involve complex math/coding problems Combining these things should ensure you don't hit rate limits in the new model.

/react doctor is a new way to fix every bug inside your React app. One scan, and every issue is caught & resolved. Fully open-sourced and ready to install now as an agent skill. Terminal command: " npx react-doctor@latest "

It transforms Claude Code into a structured development platform with powerful tools and intelligent agents. It also injects behavioral changes into Claude. Check out the full framework: https://github.com/SuperClaude-Org/SuperClaude_Framework