seeing if/how graphs come back in

Is this graph renderer Collaborator's own? Which one is using...?

fork of

Thank you very much! I've been using @warpdotdev Split Panes, but Collaborator has a very attractive expandable view, is there any chance I'll use Warp inside it?

@warpdotdev I also quite like warp. Which exactly about warp would you like in collaborator? The spit panes?

@warpdotdev I like the options to customize the look, the tab between answers it creates, the context information of each folder, the Autocomplete and history, integrated MCPs, list of commands saved with alias... There are many things.

What is bro cooking 👀

Seems like i really need it

Bro you strike me like an evil genius

😅 evil, maybe

evil maybe, genius right on cause your canvas approach is kinda sick tbh confusing but sick fr, hence the evil prefix

looking great will try it later today

👍🏼 would love to hear your thoughts

Hey, I built something similar Skill management is mostly really messy and full of BS, and a lot of it is just people obsessing over nice-looking graphs. So we built better-skills[.]dev, a better way to manage, create, share, and visualize your skills. Check it out!

Love seeing dev tools experiment with visuals

What do you mean by “if” Say more please

Don’t wanna push features just to push features

I totally get it bro What’s the mechanism for if graphs emerging. Anything I can read. What’s the research direction or architecture here Seems very interesting.

Hint: I’m watching MiroFish very closely :)

Okayyy I’m yet to check it out. See this. Basically graph emerges unless explicitly asked not to. And it’s not really a graph . More like a hyperlink. A pointer with a descriptor. I already have clusters now. And waiting to see how reasoning chains emerge

node connections random or thematic?

Would love to see source control panel in there tbh would be so useful

Say more?

Usecase is my workflow, I generally stage/unstage changes as I’m working on multiple tasks also manage worktrees Would love to see vscode’s source control panel in this

nice, been looking for something like this. how's the latency on larger graphs?

Completely depends on your single-threaded CPU perf, so pretty good on Apple Silicon for graphs <1k nodes. What are you looking to do with a graph?

knowledge graph visualization for an AI agent system. ~260 markdown docs connected by cross-references, wanted to see the topology

For global search and tracing tool use over a CC session Or for interpretability - if you tally the tool call or retrieval counts and grow the node size based on how much the file is surfaced. That would be helpful for understanding the topology of the graph

Go to byte level