← sources

How do you manage a shared setup across multiple coding agents?

source: reddit (reddit:AI_Agents)published 2026-10-21 00:18ingested 2026-10-22 02:36

retained text

I'm looking for a way to manage multiple coding agents without having to maintain completely separate setups for each one.

Ideally, I'd like to switch easily between Claude Code, Codex, OpenCode, GitHub Copilot, Antigravity, etc., depending on which models are best, pricing, usage limits, and so on.

I'm particularly interested in having a single source of truth for things like AGENTS.md, skills, MCP servers, hooks, and other shared configurations, both globally and per project. Ideally, it should also sync across multiple machines and environments (Windows, WSL (ubuntu)).

I realize some agent-specific settings and permissions probably need to be managed separately, potentially through something like chezmoi, since their setup are very different from coding agents to coding agents.

Has anyone built a setup like this? Are there any good open-source tools for managing it all centrally, or is a custom dotfiles repo with symlinks/scripts still the way to go?

  
**Edit:** Thanks for the replies! A dotfiles repo with some simple scripts seems to be the common approach.

I've also come across tools like Rulesync, Microsoft APM, Skillshare, and others in my research.

Has anyone actually tried these? Do they make maintenance easier when agents update, or do they just add another layer to maintain?

I'd also like to keep both my own and downloaded skills in one version-controlled library, synced across machines.

Not looking to overengineer this, just wondering if any of these tools are worth using instead of maintaining everything manually.

document fetched 2026-10-22 00:19 · canonical https://reddit.com/r/AI_Agents/comments/1xb8lj6/how_do_you_manage_a_shared_setup_across_multiple