Unity

Unity is an open-source CLI tool built in TypeScript by nicolasakf that keeps Agent Skills and agent rule files synchronized across multiple AI coding agents from a single canonical source. It mirrors skills and rules into the native folders expected by tools like Codex, Orion, Claude Code, Cursor, and OpenCode, eliminating drift when you switch between agents. The project is available on GitHub under the MIT License.

What It Is

Unity is a synchronization layer for AI coding agent configuration — specifically the "skills" (reusable agent capabilities from the Agent Skills ecosystem) and "rules" (agent instruction files) that different coding agents each store in their own proprietary locations. Instead of manually maintaining separate copies for each tool, Unity designates a canonical source tree (~/.agents/ for user-level, <repo>/.agents/ for project-level) and propagates changes bidirectionally to every registered agent mirror.

How the Sync Model Works

Unity operates on two scopes:

• User scope — skills at ~/.agents/skills, rules at ~/.agents/rules, synced globally across all registered projects
• Project scope — skills at <repo>/.agents/skills, rules at <repo>/.agents/rules, scoped to a single repository

The unity watch --pull command starts a file watcher that monitors both the Unity source directories and the agent mirror locations. When a change is detected in any mirror (e.g., an agent edits its own skill file), Unity pulls that change into the canonical source and then pushes it back out to all other mirrors. Without --pull, only the source directories are watched and outbound propagation is one-directional until a manual unity pull or unity sync is run.

Setup Path

Getting started requires three steps:

1. Install globally via npm: npm install -g @nicolasakf/unity
2. Run unity init to select target agents (e.g., codex, orion) and optionally register project repository paths
3. Start the watcher with unity watch --pull

A --non-interactive flag is available for use by AI coding agents themselves, allowing automated initialization without interactive prompts. Additional CLI workflows for pull, push, sync, env push, project registration, diagnostics, and import repair are documented in the project's full setup guide.

Target Audience and Use Case

Unity is aimed at developers who work across multiple AI coding agents and want consistent skill and rule definitions without manual duplication. It is particularly relevant for teams or individuals using Agent Skills — a registry of reusable agent capabilities — who need those skills available regardless of which coding agent they are currently using.

Current Status

The repository was created in April 2026 and last updated in August 2026, with the most recent push in June 2026. It is an early-stage open-source project with 7 stars and no forks at the time of indexing. The core sync, watch, pull, push, and init commands are documented and functional, with configuration and full setup guides available in the repository docs.