AgentSync CLI
A CLI that keeps instructions, commands, skills and MCP in .agents and links them into the configs of different agents
- 76 ★ on the repository
- Last commit
Extended access
Why this level
- `apply` creates and replaces symbolic links to agent settings files in the project
- Distributing MCP configuration changes which servers all your agents see
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add cli/agentsyncDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- `apply` creates and replaces symbolic links to agent settings files in the project
- Distributing MCP configuration changes which servers all your agents see
This is third-party code. Review the repository files before installing.
What it does
AgentSync is a fast portable CLI in Rust with an npm wrapper. Different agents expect settings in different places: `CLAUDE.md` for Claude Code, `GEMINI.md` for Gemini CLI, `.cursor/rules/` for Cursor, `AGENTS.md` for OpenCode, plus their own folders for commands and skills. AgentSync keeps one source of truth in `.agents/` and creates symbolic links to the right places, and lays out MCP server configuration for each agent. The loop is short: `agentsync init` creates or migrates configuration, `agentsync init --wizard` asks questions, `agentsync apply` creates or refreshes links, `agentsync status` shows sync state and `agentsync skill` manages installed skills. An edit in one file reaches every agent at once, with no copies and no drift. The command can run through `npx @dallay/agentsync apply` without installing, be installed globally through npm, pnpm, yarn or bun, added to `package.json` as `prepare` and attached to Lefthook or Husky Git hooks so links refresh after a branch switch.
Who it is for. Teams that work with several coding agents at once and are tired of maintaining the same instructions and skills in different files.
Good fit when
- A project uses Claude Code, Cursor and Gemini CLI and instructions drift between files
- A new team member needs settings for all agents with one command
- MCP server configuration should be distributed to all agents from one place
- After a branch switch or merge, symbolic links should be refreshed by a Git hook
Not a fit when
- You work only on Windows without symlink support: extra setup or WSL is needed
- An agent expects a format AgentSync does not know: support follows a table of known agents
- You need content that differs per agent: there is one source and the links point at the same files
Example request
Move my instructions and skills into .agents and distribute them to Claude Code, Cursor and Gemini CLI through agentsyncLimitations
On Windows you need the right to create symbolic links or WSL, with details in a separate guide. Building from source needs Node.js 22.22 and Rust 1.89. Agent support follows a table in the README, and for GitHub Copilot and VS Code some command and skill paths are absent. Links mean an edit in `.agents/` is visible to every agent at once, so a mistake in a shared file spreads everywhere as well. Installing from releases requires checking the checksum, which the README describes.
How to disable. Delete the created links by hand or with the revert command from the guide and run `npm uninstall -g @dallay/agentsync`. The `.agents/` folder stays as ordinary project files.
FAQ
How can I try it without installing?
With `npx @dallay/agentsync apply`. pnpm has `pnpm dlx`, bun has `bunx`, and for a persistent command use `npm install -g @dallay/agentsync`.
Which agents are supported?
Claude Code, Gemini CLI, Cursor, GitHub Copilot, VS Code, OpenCode, Z-Code, MiniMax Code and others. The README holds a table of paths for instructions, commands and skills.
What does the .agents folder hold?
The single source: instructions in `AGENTS.md`, commands, skills and the `agentsync.toml` settings file. AgentSync links to them from the places agents expect.
How do I automate applying the links?
Add `agentsync apply` to the `prepare` script in `package.json` or to a Git hook via Lefthook, Husky or simple-git-hooks. Then links refresh after install and branch switches.
Is there a Rust install?
Yes, `cargo install agentsync` installs the CLI from crates.io. Prebuilt binaries are in GitHub releases with checksums.
Discussion
No questions or reviews yet.
Related
CLI toolsdotagents is a Sentry CLI that wires skills, MCP servers, hooks and plugins from one agents.toml into each agent's folders.
One config for Claude Code and Codex in every repo: shared rules, skill packs, second-model review and a destructive-command guard
A portable .agent folder with memory, skills and protocols that plugs into many coding agents and keeps its knowledge when you switch tools
A bridge between local CLI agents: Codex, Cursor or Claude Code delegate work to Antigravity, Grok, Kimi and seven more.
The squads CLI runs teams of agents defined in markdown files in git: a lead briefs, workers act, a verifier checks.
A skill for messaging between Claude Code, Codex, Gemini CLI and other CLI agents through a shared local SQLite database, written in bash
Comparison with similar tools
Cards that share the main search query, and how they differ.
README in short
The README explains the problem of scattered settings and shows a source-of-truth diagram for `.agents/`. A table of agents with paths, install routes through package managers, cargo and binaries, a quick start and notes on Windows and Git hooks follow.