SleeplessDroid

@lawrencecchen sadly it breaks when spawning agent teams...somehow it chooses the model id not compatible with 3rd party one (kimi25turbo from firepass). Can we config/finetune the initial agent spawn terminal command somehow?

@AnPod you can set env variables they should be passed in. so maybe in your zshrc you could do: fireclaudeteams() { export CLAUDE_CODE_THING=... cmux claude-teams --dangerously-skip-permissions "$@" }

Introducing cmux Claude Code Agent Teams: `cmux claude-teams --dangerously-skip-permissions` Teammates/subagents spawn as native cmux pane splits. They stack vertically in a right column and auto-equalize as agents spawn and exit. `cmux claude-teams` automatically sets CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS=1 and shims tmux on PATH with cmux's tmux-compat layer, so you don't need to update your Claude config. All arguments forward to Claude Code, and it works over cmux SSH too. Out in the latest version of cmux (0.63.x).

@lawrencecchen hey, since last update if I select in opencode within a cmux a text (auto copy to clipboard) and try to cmd+v it, I just get the first letter of the initially selected text. Also on cmd+q and reopen of cmux all my opened tabs/workspaces get lost. Any plans on fix?

@lawrencecchen idk the best ux, i think desktop ADEs usually open the file explorer then to the right

@lawrencecchen any chance to get file browser natively in cmux like here?

@lawrencecchen can you make a btn inside each pane to open the current dir in Zed, VSCode,... ? This would be great as i cant find file explorer in cmux, which apart of browser would be great. See other ADEs

@theo Synnthetic.new thank me l8r

So, uh, what subscription should I be using for my OpenClaw now? 🙃

@MilksandMatcha To build the best infra, tools and apps the internet has ever seen!

Giving away 5 Codex Pro plans Each person will get 3 months of free Codex Pro (highest tier). Winners will be selected from comments in 48 hours, comment below why you want it.

@NikoStarus @0xSero Bruh 12 hours…touch grass, have some kids or smthng

@0xSero terminal-native agent orchestration is the way. we run a similar setup -- multiple coding agents coordinated through tmux with dedicated roles. the polish matters because you end up living in these tools 12+ hours a day

I found the winner the level of polish makes me happy. Support Droid, Hermes, Codex, Pi, and much much more. Clean design, uses terminals as it should. github.com/generalaction/…

@0xSero What about running in sandboxes. Lots of supply chain attacks these days and will probably get worse. Any recommendations?

@rwrrll @dgrcode @dhh @FireworksAI_HQ U do plan / task breakdown with glm, codex or opus. And burn the tokens with kimi. Poetry

@dgrcode @dhh @FireworksAI_HQ IME it’s very very good at reasonably contained tasks with an obvious outcome, not very good on system design or deep understanding of complexity Much stronger bias to action than the big boys, so less pondering and flambéing, much more doing 100% a useful tool in the belt!

Kimi K2.5 Turbo is hilariously fast on @FireworksAI_HQ.

@le0maiajr @theo I guess ill stay on open weights models some more then..

@AnPod @theo max $200

People keep asking me about Conductor vs T3 Code, here is my thoughts (open source ftw)

@le0maiajr @theo On which plan

the quota is draining abnormally fast during active sessions this week (since Monday) i burned through the 5hr window on a single PR. some users are saying disabling the 1M context window helps. all of this during the 2x promo they launched as a thank you lol. and not a word from Anthropic about it

@q_yeon_gyu_kim @justsisyphus @OpenAI So whats bow the best model setup for omo?

wow gpt 5.4 soooo good it feels a bit different from opus obviously but it could be @justsisyphus on oh-my-opencode first time ever gpt sisyphus experimenting now very good model @OpenAI

@ryancarson Pattern ident.: The agent keeps "abandoning" the step. It does the work, commits changes, updates progress.txt, but doesn't properly signal step completion to antfarm. The issue: Antfarm steps require explicit completion via CLI (step complete <step-id>), not just text output

@openclaw any chance to adding Viber as a channell to interact with?

@opencode is there a documented way to disable auto-update?

@justsisyphus hey man, happy bday. But something wierd happening with my OmO setup since Opencode 1.1.1 released (i instruct Sysiphus, but Build executes it)

[Q:] 6.7K ! We Just Beat Apple! And.. Orchestrator?! I was born exactly 23 years and one day ago, haha. Our project has just surpassed 6.7k stars; we overtook google/mcp a while ago, and today we even surpassed one of Apple's repositories. Everyone. Thanks to you, I've had the most memorable day of my life. Honestly, all I did today was have lunch with my family and code with my friends. But I think this day will just quietly remain in my memory. I simply wanted to say thank you. I want to live up to your expectations! I am building the best agent harness. You have just joined this journey. And we are quite late, right? True Orchestrator. With Decent Multi model, Multi Agent. Soon.

@thdxr @toeknee___ would this not be the same as :brainstorm? github.com/obra/superpowe…

@toeknee___ coming next week

why do you use both cursor and opencode together when cursor has an entire agent view?

github.com/AnPod/Switch-O… Interactive CLI (bash) to switch between oh-my-opencode configuration profiles (global or project-local .opencode/) with an arrow-key menu—quickly swap providers/models and restart OpenCode to apply.