Hey everyone, Said here, the maker of CrewTower.
On October 1, Claude Code shipped mods. Hooks could already tell an outside app what Claude was doing and answer a permission prompt. They could never talk back. A mod can.
A mod is a small JavaScript or TypeScript file that Claude Code loads into its own process. It gets called on every event (a tool call, a prompt, a turn ending, part of the screen being drawn) and can watch it, rewrite it, or take it over. It can draw panes and buttons, add slash commands, read context numbers directly, and reach outside Claude Code.
That last part is what I needed. CrewTower lives in the MacBook notch and shows every agent session you run. With a small mod, each Claude card in the notch can now talk to its session:
@said_altan sessions sitting there waiting for me is very real, so this makes a lot of sense. what I'd want from the notch is picking a feature spec file and sending it as the next prompt, instead of retyping the context every time I start a new task.
@leoheo yeah retyping the same context every time gets old fast. right now it's just text you type, but picking a spec file and sending it is a nice idea, gonna think about how that'd work
@said_altan nice one. my sessions often run long tasks like a video render, and on the card I only see working. would be cool to see what it is doing right now, like last tool call or a progress line. is this possible with the mods?
@joertemple Of course. Actually there is an activity button on the card showing you what your agent is doing at the moment
Being able to send Claude the next message without switching back to the right terminal tab looks useful. Can I edit or cancel that message while Claude is still working? Sometimes its answer changes what I want to ask next.
@den_is1 not yet. once you send it, it just waits and goes in when the turn ends. you can hit Stop on the card but you can't edit it from there. had the same thing happen to me actually, so I'll probably add that
I love this this is so creative, I think a lot of vibecoders would need this at a point. If you don't mind me asking how long did it take to build this all?
@gagik_hakobyan1 around 2 weeks
Congrats Said on the CrewTower launch and the new Claude Code mod. The context breakdown with a hint to compact at a task break is the part I would lean on most. How does the card decide it is at a task break: just the turn ending, or something like tests passing or a commit? Having it offer the compact right at that moment would be my answer to your question.
@genedai thanks Gene! right now it's pretty simple, it's just when the turn ends and Claude's waiting on you. and it only suggests compacting once the context gets big, around 300k on 1M. tests passing or a commit would be a much better moment to suggest it tbh, hadn't thought of that