Replying to @LukeParkerDev
also experienced this even just yesterday. if your ChatGPT conversation is compacting and you attmept to steer at the same time, the icons just flicker as though state is overwriting each other.
Claude desktop seems super polished though, even on Linux
Replying to @nexxeln
home/end key doesnt work in subagent view. only in the main agent view.
Replying to @imnotchalk
Replying to @thdxr
considering it is faster to open the panes in a new tab (yoyo-code.com/why-is-github-…), one cannot blame the user
Replying to @kaushikgopal @pidotdev
Lots of themes between this and OpenCode DCP, which similarly has the model pick messages to remove!
One of my favorite features of Herdr is that it has great ways for agents to introspect it's state and update it. tmux also has this capability, but it feels so much more natural using Herdr.
Some example prompts and outputs, this is really powerful to get Herdr spawning agents.
Been using Herdr lately to manage my agent sessions and really enjoying it. There's still more work to be done to have better integration with jj, however it has excellent hooks with all agent harnesses.
Highly recommend! Next on the list is to try out Collie to prompt remotely.
Replying to @RhysSullivan
Replying to @jxnlco
Adding a new MCP server that supports OAuth but not DCR shows no error in ChatGPT app when you add it, the "authenticate" button basically does nothing. You have to figure out to manually edit your toml config to add the oauth mode.
Replying to @badlogicgames
5.6 in particular loves to do an end-of-turn subagent to review its work
Replying to @RhysSullivan
if you're not careful someone will mitm steal your slop bowl
Replying to @jakubkrehel
1. accessibility-contrast.md says positive means light text on dark, but positive means dark text on light (see APCA formula)
3. APCA recommends 75, not 60, for body text, and ideally 90.
Replying to @RhysSullivan
There's experiments in the MCP to add this as well -- github.com/modelcontextproto…
Replying to @xubinrencs @kimmonismus
OpenAI servers have been broken since Friday for Asia users. See github.com/openai/codex/issu… and the linked OpenCode thread where lots mention their experience. It's an underlying fault of the HTTP API, not the app itself.



