Agentrig

Welcome

Reuse the AI workflows that work best for you and your team.

Agentrig documentation

Agentrig helps devs and teams package the AI workflows that already work, then reuse them across projects, tools, and teams. Instead of rebuilding the same setup for every repo or assistant, you can author it once as plugins, compose it into rigs, and install it natively where people already work.

Agentrig supports portable AI workflows and native integrations

What is a plugin?

A plugin is a folder with root plugin.json, portable skills/ and mcp.json components, and optional AgentRig-specific files under ai.agentrig/. AgentRig plugin manifests are Agent Plugins 1.0.0 conformant, with AgentRig-specific fields under extensions["ai.agentrig"]. The CLI's public install contract is the AgentRig InstallBundle resolved from approved registry/catalog listings.

How AgentRig helps

How Agentrig standardizes and reuses AI workflows across projects and tools

Agentrig is not limited to one assistant or editor. You can export and install the same plugin as a native integration for the tools many developers already work in every day.

  • Reusable workflow plugins: package skills, MCP config, hooks, agents, commands, and rules into registry-installable plugins that can be shared across projects.
  • Composable rigs for teams: combine plugins into shared setups for a stack, repo, or org so teams can standardize how they work with AI.
  • Native delivery where people work: export the same workflows as integrations for Claude Code, Codex, and Cursor, with tracked installs and safer cleanup.

See Marketplace exports for provider-native output layouts and Rigs for team composition.

Quick paths

On this page