Broad dumb questions #393
-
|
Beta Was this translation helpful? Give feedback.
Replies: 3 comments 1 reply
-
|
I can't speak to the team flag, but offhand I don't see any reason this wouldn't work with any tool/model as long as it can be fed the instructions. |
Beta Was this translation helpful? Give feedback.
-
|
Great questions! Let me address each: 1. Multi-agent compatibility (Claude Code + Jules/Gemini) Yes, beads is agent-agnostic. It's a CLI tool that any agent can use as long as it can run shell commands. The symlink approach ( 2. Team mode ( The
If you're using branches sometimes but not always, 3. Mixed agent teams (Claude Code, Codex, etc.) Should work fine! Beads uses git for sync, and the merge driver handles concurrent edits. The ID system is adaptive - it automatically uses longer hash IDs as your issue count grows to minimize collision risk. With the default settings:
The main consideration is making sure each agent's instruction file includes the beads workflow. Most agents that can run shell commands should work. |
Beta Was this translation helpful? Give feedback.
-
|
Thanks! |
Beta Was this translation helpful? Give feedback.
Great questions! Let me address each:
1. Multi-agent compatibility (Claude Code + Jules/Gemini)
Yes, beads is agent-agnostic. It's a CLI tool that any agent can use as long as it can run shell commands. The symlink approach (
GEMINI.md→AGENTS.md) works well - just make sure yourAGENTS.mdincludes the beads workflow instructions. The key is that agents learn to callbd create,bd list,bd close, etc.2. Team mode (
bd init --team)The
--teamflag runs a setup wizard that configures:beads-metadatabranch for issue commits