docs

OpenHop docs

Reference material for the people who already know what OpenHop is. If you're new, start with the project README and npx openhop demo.

Contents

Doc	What's in it
install.md	Every install path — npx openhop init, OpenSkills, the Claude Code /plugin command, and the manual file-drop layout for each supported client.
yaml.md	YAML schema reference. Every step kind, every data shape, every keyword the skill is allowed to emit.
cli.md	Full CLI reference — every command, every flag, every exit code.
architecture.md	What runs where, why it's local-first, and the wire shape between the agent, the CLI, the API, and the web renderer.

The animated playground lives at https://naorsabag.github.io/openhop/. It loads the same renderer used locally — same sprites, same animation, same inspector — but stores nothing server-side (the flow is encoded into the URL hash). Good for sharing a snapshot without anyone installing anything.

Conventions

• Commands prefixed with $ are run by you. Output below is what you'd see.
• npx openhop … and openhop … are interchangeable once the package is installed globally or as a dev dependency. The docs use npx because it works without any prior install.
• Anything labeled agent-facing is what your AI coding agent emits or consumes — you generally don't write it by hand. The source of truth for the agent's contract is skills/openhop/SKILL.md.