How to get notified when Claude Code or Codex needs input
Get a desktop notification when Claude Code or Codex stops at a permission prompt or finishes: Claude Code hooks, Codex settings, and one setup for all agents.
NOTES FROM THE HIVE
Get a desktop notification when Claude Code or Codex stops at a permission prompt or finishes: Claude Code hooks, Codex settings, and one setup for all agents.
Why closing the terminal or SSH kills Claude Code and Codex, and four fixes: tmux, headless mode, Hive, and an always-on machine. Plus how to resume.
A five-step review routine for code written by Claude Code, Codex and other agents: isolate, overview, read by hunk, what to watch for, and run it yourself.
Zellij and Hive compared for running Claude Code, Codex and other agents: what Zellij does well, what agents need, and which to use when.
What changed in Hive 0.4: macOS and Windows, spaces that follow your cd, worktree terminals, right-click menus, layouts, and optional integrations.
Four rules for running Claude Code, Codex and other agents in parallel: separate checkouts, visible status, sessions that survive, and one place per project.
Give every coding agent its own checkout with git worktree: the commands, the pitfalls (dependencies, .env files, branches) and how to clean up safely.
An honest comparison of tmux and Hive for running AI coding agents: what tmux does well, what Hive adds, and when to use which (or both).