- It'd be great to change the default branch used for creating new workspaces.
- I'd like the ability to add custom tools to the "Open in..." menu.
> It'd be great to change the default branch used for creating new workspaces. Yeah you can actually change this now! If you click the repo name you can make changes to the "setup script". If you added `git checkout -b "branch name"` it would run that on every new workspace instance.
orig_branch=$(git branch --show-current) && git checkout dev && git pull && git branch -D "$orig_branch" && git checkout -b "$orig_branch"
But no way to find out if there’s any data sent to your servers etc, unless I’m missing some links?
Right now the app uses GitHub's OAuth sign in (https://docs.github.com/en/apps/oauth-apps/building-oauth-ap...) which unfortunately doesn't allow for fine-grained permissions (it will only have access to organization code if you explicitly grant it) . We're switching our sign-in to a GitHub App so we can make the permissions fine-grained.
Or, skip the integration and use your local GitHub CLI auth.
More info: https://docs.github.com/en/apps/oauth-apps/building-oauth-ap...
More-over, can you document the GitHub permissions needed and which GitHub App(s) this tool uses? Are you using device-flow, online oauth-flow, etc? And where are the Oauth tokens stored if so? Is there any server-side component where you might be storing tokens?
Thanks!
It mostly works, except we don't have a clean flow for docker: shared system daemon & repository means need to manually tag & run by branch/project (docker compose -p ...), which is friction for the LLM and even more setup than we want
As a heavy multi session Claude Code user, this may be what finally converts me to cloud IDEs...
But I'm an old school vim/emacs keyboard only hacker, so terminal all the way is super efficient for me
Anyways, excellent work!
What I do want is simple git worktree management for an already-checked-out repo on my machine, no Github permissions or dependency re-installation (copying node_modules, etc.).
So starting a new worktree requires additional setup and isn’t as simple as just checking out a new branch
Internally our workflow looks like this: - we have a script that sets up a repo — copies env variables, runs pnpm i, inits a db, etc - we have a field in the repo settings called “setup script.” every time you make a new workspace, that script runs
Hopefully will be much improved over the next week or two!
Babysitting one process, but giving it all of my attention[1] can lead to rapid progress.
Pushing forward multiple efforts simultaneously can lead to rapid progress if code is largely unrelated and you keep a sharp eye on branching and commits.
However, it’s remarkably easy to make a mistake and be enamored by how well the AI can use git to fix your mistake.
But the cost ends up being adherence to it, specification or architecture, or being distracted into scope creep.
The new scope might be useful for factoring, but that is also the kind of thing that interfere with your Main processor trying to handle.
And it’s remarkably easy to get lost in the state when you have to manage validation well.
The more mature the project and more intense the expectation of scm hygiene, the more punitive small mistakes can become.
It is not really a problem in my process prior to Claude code, which was almost entirely turned based chat.
I still have configured fairly complex and detailed session workflows to preserve state amidst the regular threat of auto compact.
Even I even have documentation, describing a plan to migrate into git worktrees myself, but haven’t been able to justify the move, as I continue to learn to be productive with a single main session focus.
[1] I don’t know if there is such a thing as constant focus on a complex genetic programming effort because of the delays involved. You can stare at the funny verbs anthropic picks while the tokens go brrrr for only so long.
Yes, the bottleneck is going to still be human review. But with only a single Claude Code (or Cursor), I spent a lot of time (a couple minutes at a time) waiting for the agent to run through its loop completing what I asked it to do. In this time I found I was naturally switching context, but not always productively (checking slack, email, whatever). I called it "agent lag" time.
Now with these worktrees tools, I can easily bounce between tasks I have the agents working on during these natural lulls. Overall I'm able to get more done and spend my time more efficiently.
Also, Cursor has had the option to run multiple agents for a while (in tabs), but I always found it cumbersome because all the agents are operating on the same git checkout, so you have to take special care for them not to conflict with each other. Worktrees completely solve that problem.
Would love to discuss more what kind of testing setup you’re looking for, want to shoot me an email at jackson@melty.sh?
works like you'd expect
I’m sure the authors would appreciate well-thought alternative suggestions and assistance.
I'm the only user at the moment, and I really enjoy the workflow. I run about four claude-codes at once this way. It's a little underbaked but I think this is the way a lot of people are going to go. Seems like the 'par' tool in a sibling comment is a similar approach.
Containers do make things easier, especially since agents can see the log output of any service super easily. To do the same thing outside a container you need to send logs somewhere the agent can see.
Agreed that worktrees seem clearly better.
You are INSANE to authorize this app on anything other than throwaway code.
@charlieholtz care to comment?
It will only have access to organization code if you explicitly grant it. We're working on switching our sign-in to a GitHub App so we can make the permissions fine-grained.
In case people aren't reading, here's the c+v of what Conductor gets access to currently:
--
This application will be able to read and write all public and private repository data. This includes the following:
Code
Issues
Pull requests
Wikis
*Settings*
*Webhooks and services*
*Deploy keys*
*Collaboration invites*
Or, skip the integration and use your local GitHub CLI auth.
- I switch in between planning and execution in the middle of the conversation using Terminal a lot, it would be nice to have it here as well rather than defining how I want Claude to think in advance.
- Entering messages when the agents can result in task lists to save time.
I’m also experimenting different UX with CC, here is Claude Code running in Slack if anybody is interested. https://peerbot.ai
> - Entering messages when the agents can result in task lists to save time.
Ah yes, queuing coming soon too! Do you prefer if the queued messages send when Claude is finished or interrupt it?
Sadly, this is lost with conductor.
I just don’t feel as joyful using it.
Being able to quickly hit escape to interject. Escape again to see the conversation history.
The key bindings should be exactly the same.
I don't think I'm looking for an interface to replace CC - it's a great interactive terminal app.
I just want a better way to manage the sessions.
The horror!
On macOS my environment is Sublime Text as editor, iTerm as terminal. I open two full screen iTerm windows of four panes each, start Claude Code Opus 4 and type "just task" in each pane. CLAUDE.md tells the session to run this justfile command and follow the directions, which generally include instructions to continue by doing "just task" again. Each session may edit its <file>.c and <file>.h but not shared headers; if it has a problem with this I touch STOP in the task directory, the other sessions eventually are told to wait, and then the trouble child session can edit what it wants.
I'm sure in a few years AI will read and fix my 57 C file code base all at once, but for now this workflow has proved invaluable.
The irony? This is the dumbest parallel task manager imaginable yet it works. And our code base has a similar feel: The storage management is a set of fixed sized stacks that outperformed malloc() by 4x back in the day. Etc. Etc.
I learned this gift from my Dad. He devised the Bayer filter for digital photography back in 1974, it looks like ten minutes work (it wasn't) but we're still using it.
Reason I ask is I have gotten outstanding results from gemini pro 2.5, which has a very large context.
It already accepts very large individual messages, however if the code base is still too big for a single message, it is possible to send it all in multiple messages and still have the model keep it all in context.
For large, system wide refactors, this can provide planning that claude code can't or won't do efficiently.
Sometimes, I'll have claude code come up with a plan, feed all the related code to gemini, and then provide claude code gemeni's improved plan.
Then after the work is done, have gemeni validate the implementation.
I've looked a bit into automation around this but haven't picked something yet.
If anyone has a solid workflow for this already, I'm all ears.
1. Place all your repos inside a parent directory and name them accordingly (frontend, backend, devops, etc.)
2. Each repo has its own claude file.
3. The parent repo has a claude file that links these claude files.
I’ve had claude successfully navigate across these and uses services with back-end, setup terraform infra and then complete the implementation for the front-end.
I'm surprised people are running multiple agents, and are able to check their outputs diligently.
If you don’t like code so much you can just do something else with your life. Have you tried sales or finance?
Let it running in parallel "unwatched" will end up with nothing but a pile of code that will have to be re-written.
It was not that way at the beginning, when I started in May 24th (2 days post Sonnet/Opus 4 announcement), however, last 2 - 3 weeks, they've tweaked they system substantially (I mean I have no other explanation) and its behavior is despicable.
I have see a few on X complaining the same (along with far more about GPT o3 and 4o) but the culture of "opaque transparency" in the AI industry left us with nothing but guesses.
You can see in various subreddits the spikes in complaints--and there is definitely one in week(s) on Claude Code.
I generally take this as a sign of needing to get more skilled at context control and validation. And that by doing so, you'll be in only a better place once the product you're relying returns to baseline or leaps ahead again.
I also suspect it is tuned for ultimate experience upon release to grade well on benchmarks and online reports, then tuned back to a level they can bear as they burn investors money.
the need is clear, I am just expecting transparency.
The tools like Cursor have privacy mode that ensure that.
Lots of enterprises use Azure, GCP, AWS for the models for these reasons as well.
Regarding the code, I don't believe it is a concern for most of the people, because they use 3rd party git repo anyway, which has exactly the same level of risk
> the support for submodules is incomplete
This has been a stumbling block for me. In C firmware, there’s no package management so submodules tend to be how we bring in libraries.
My experience has been that worktrees simply do not work with submodules.
Made it to solve for myself. Published because my friends loved it and wanted to use it.
This is a simple problem and users should be able to open existing repositories, view multiple projects, get notified when Claude needs attention (done or needs user input)
jamil7•6mo ago
Charlieholtz•6mo ago