AI Environments

AIM applies a single inventory to multiple AI environments through adapters.

Supported environments

Environment	Target ID	Skills	MCP config
Claude Code	claude-code	supported	supported
Cursor	cursor	supported	supported
Codex CLI	codex	supported	supported

Gemini CLI and other environments are being considered as future adapters.

How AIM resolves paths

During aiman init and aiman doctor, AIM attempts to locate the standard directories for each AI environment.

If an environment is not found automatically, you can set the path in aim.local.yaml:

environments:
  claude-code: ~/.claude
  cursor: ~/.cursor
  codex: ~/.codex

aim.local.yaml is local to the machine and must not be committed to Git.

Skills

A skill from skills/<name>.md is installed in the format expected by the specific AI environment. AIM does this during apply (local iteration) and sync (applying the published state):

aiman apply   # apply local changes
aiman sync    # apply the published state from Git

MCP servers

An MCP Item from mcp/<name>.yaml is applied by the adapter to the AI environment's config file.

For Claude Code and Cursor, the mcpServers JSON structure is used. For Codex CLI, the mcp_servers TOML structure is used.

Secret values and env variables are substituted from the local aim.local.yaml.

Diagnostics

aiman doctor

doctor shows:

• found and missing AI environments;
• path issues;
• the state of the inventory repository;
• missing required MCP env variables.

---

What's next

• Concepts — what an adapter, Skill Item, and MCP Item are.
• Security and Limits — what AIM writes to AI environments.
• Troubleshooting — if an environment is not found or the path is non-standard.