Agent skill

The skills/ix/ directory is an agent skill. It teaches an agent which ix command answers which question, how to keep answers bounded, and how to recover from common errors.

It follows the Claude Code skill format and the agents.md standard, so Claude Code, Codex, Cursor and other agents load the same skill from their own skills directory.

Install

From a checkout of the Ix repository:

bash scripts/install-skill.sh             # deploy to every agent found on this machine
bash scripts/install-skill.sh --dry-run   # preview where it would go

Start a new agent session so the skill is loaded, then ask:

Set up Ix and map this repo.

The skill’s bootstrap script checks for Node.js 22+, git, Docker and ripgrep, installs the CLI if it’s missing, starts the backend and maps the repository.

What’s in it

File	Purpose
SKILL.md	The entry point: when to use Ix and the core workflow
references/commands.md	Routes a goal to a command. Rendered here as the command reference
references/flags.md	Every flag of every command. Rendered here as the flag reference
references/output-formats.md	Choosing text, json or llm
references/troubleshooting.md	Recovering from common failures
scripts/bootstrap.sh / .ps1	First-run setup

Skill or MCP?

They work together. The MCP server gives an agent native tools. The skill teaches an agent that runs shell commands how to use the CLI well. If your client supports MCP, register the server with ix mcp install too.