Overview

AIM Loadout is a local-first CLI tool for collecting, storing, and syncing your AI inventory — skills, MCP servers, and configuration — across Claude Code, Cursor, Codex CLI, and future AI environments.

The name "Loadout" is a metaphor for a developer's toolkit when working with AI tools: a complete set of skills and instruments you can quickly reload on a new machine with aiman sync.

You collect inventory from files and AI environments, store it in a Git repository, manage it from a single local CLI, and sync it to any supported AI environment without copying files by hand.

What AIM solves

An active AI-tools user quickly ends up with fragmented settings:

• skills in Claude Code;
• local instructions and skills in Cursor;
• MCP servers with different configuration formats;
• secrets and environment variables that can't be stored in Git;
• multiple laptops or development environments.

AIM Loadout turns all of this into one managed inventory repository and one repeatable workflow.

Core cycle

The full inventory lifecycle: collect → publish → sync.

Collect — populate your inventory from a file or an AI environment:

aiman add skill path/to/skill.md             # skill from a local file
aiman add mcp path/to/server.yaml            # MCP server from a local file
aiman import skill review-code --from claude-code  # skill from an installed AI environment
aiman import mcp context7 --from claude-code       # MCP server from an installed AI environment

Publish and deliver — apply, publish, and sync:

aiman apply          # apply local changes without touching Git
aiman push           # publish validated inventory
aiman sync           # pull the published state on another machine

apply is for local iteration: edit a skill, apply it, check it in your AI tool, repeat. push publishes the ready state. sync fetches the published inventory and applies it to AI environments on the current machine.

The core MVP focus: one user, one inventory repository, a reliable publish → sync cycle. If that cycle does not work reliably, everything else is secondary.

What the current product includes

• Git-based inventory repository.
• Skill Item in Markdown with YAML frontmatter.
• MCP Item in YAML.
• Adapters for Claude Code, Cursor, and Codex CLI.
• Local config aim.local.yaml for paths, hash markers, and MCP env values.
• Commands: init, switch, add skill, add mcp, import skill, import mcp, apply, push, sync, status, doctor.

Adding a new AI environment requires writing an adapter. This is an architectural property of the product, not a marketing promise.

What's next

• Getting Started — walk through the first working scenario.
• Concepts — inventory, Library Items, adapters, and aim.local.yaml.
• Workflows — collecting, publishing, and syncing inventory.
• CLI Reference — the exact command contract.