Guides
AgentNotch guides
Short, practical answers for people running AI coding agents on macOS. The manual options first, then where AgentNotch fits.
- 01
Claude Code notifications not working
Why a Claude Code notification hook fires but no banner appears on macOS, and the checks that fix it, from Focus mode to which settings file the hook lives in.
- 02
Codex notifications on macOS
Getting alerted when Codex finishes a task, needs permission, or asks a question on macOS, from CLI notifications to an always-on menu-bar view.
- 03
Cursor agent notifications on Mac
How to catch the moment a Cursor agent finishes or stops for approval on macOS, instead of switching back to the editor to check.
- 04
How to know when Claude Code finishes
Ways to get notified when a Claude Code run finishes on macOS, from the terminal bell and Stop hooks to an always-on menu-bar view.
- 05
How to manage multiple coding agents
Running three or four AI coding agents is easy. Knowing which one needs you is the hard part. A working method, and where the attention goes.
- 06
Kimi CLI notifications on macOS
Kimi finishes quietly in a terminal tab. Here are the ways to find out it needs you, and what each one costs.
- 07
Know when Codex needs approval
Codex pauses for a yes or no before running commands. Here is how to catch those approval prompts on macOS instead of finding the agent idle.
- 08
Monitor AI coding agents on macOS
The ways to keep track of Claude Code, Cursor, and Codex on a Mac, from terminal bells and notification hooks to an always-on menu-bar dashboard that shows which session needs you.
- 09
Monitor Claude Code subagents
A Claude Code session that spawns subagents looks the same from outside as one that has stalled. Here is how to see what is running on macOS.
- 10
Monitor Claude Cowork sessions on macOS
Cowork works on its own for minutes at a time, then stops for a question you never see. Here is how to catch those pauses without keeping the Claude app in front of you.
- 11
Monitor multiple Claude Code sessions
How to run and track several Claude Code sessions at once on macOS, using tmux to organize them and a menu-bar view to catch the one that needs you.
- 12
Run Claude Code, Cursor, and Codex in parallel
A workflow for running several AI coding agents at once with git worktrees and one task per agent, and why attention becomes the bottleneck.
- 13
See how much Claude Code usage is left
Claude Code does not show your plan limits while you work. Here is how to read what is left of the 5-hour and weekly windows on macOS.
- 14
Stop babysitting your AI coding agents
Watching the terminal to catch when an agent finishes or needs a decision is a bad use of attention. Here is how to break the habit on macOS.
- 15
Track plan limits for Codex, Cursor and Kimi
Every coding agent meters you differently and shows it somewhere else. Here is how to read all of them in one place on macOS.
- 16
When Claude Code needs your input
Claude Code stops and waits when it needs permission or an answer. Here is how to catch that pause on macOS instead of finding it minutes later.
- 17
Will you run out before the limit resets?
A percent tells you where you have been. Whether the limit lasts depends on how fast you are spending it. Here is how to read pace on macOS.