Repository navigation
Trying Concord with Gemini CLI and another coding agent #30248
alexechoi
started this conversation in
Show and tell
Replies: 0 comments
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
I'm building Concord, an open-source MCP server that lets coding agents share work-state across harnesses.
There is a Gemini CLI integration, and I'd like feedback from people who use Gemini alongside Claude Code, Cursor, or Codex. The useful test is simple: have one agent claim a module, then ask Gemini who owns it before starting related work. Task decisions and handoffs stay available to both sessions.
Separate Gemini CLI sessions can coordinate through Concord too. One can work on implementation while another handles tests, with their own identities, file claims, and shared task context.
npm install -g @concord-ai/concord-mcp cd your-repo concord setupRestart Gemini after setup. For the documented monitor startup flow, launch it with
concord adapters gemini. Idle message delivery depends on its session-owned monitor running; plain Gemini sessions need an initial turn to start that monitor.Gemini CLI guide.
Concord Cloud connects local, remote, and async agents in one shared workspace. Work can continue across machines and at different times, with task context and handoffs available for the next agent. Try Concord Cloud.
If you try it, could you share your Gemini CLI version, whether two agents register distinctly, and whether a peer message reaches Gemini during a turn and while idle? A short failure report would be useful too.
All reactions