Plugin

ClawRecipes

The installation layer for OpenClaw teams. Scaffold roles, workflows, memory structure, and repeatable operating patterns so your agent stack feels like a real system, not a pile of prompts.

Browse Marketplace Explore ClawKitchen GitHub

Scaffold full teams

Generate production-shaped teams with roles, shared workspace conventions, ticket flow, and instructions already wired together.

Scaffold single agents

Spin up one focused specialist fast when you want a single operator instead of an entire staffed team.

Managed memory

Keep durable team context in files and structured memory instead of relying on one long fragile transcript.

Installable operating systems

Use marketplace recipes, skills, and proven setups to get to a working system quickly instead of rebuilding process from scratch.

Quick Install

OpenClaw plugin

openclaw plugins install @jiggai/recipes

Restart gateway

openclaw gateway restart

Feature 1

Orchestrators that turn prompts into operating systems

ClawRecipes is not just a scaffold command. It is the install layer for real operating models. Teams can install recipes that already define roles, file structure, workflow expectations, escalation paths, and ticket discipline. Instead of inventing a team from scratch every time, you start from a known-good operating shape and adapt from there.

●Install ready-made multi-agent team blueprints
●Keep role boundaries, authority, and handoffs explicit
●Start from repeatable structure instead of vibes

Feature 2

Workflow triggers for real execution, not demos

Recipes pair naturally with orchestrated workflows. Trigger work from cron schedules, events, or manual launches, then move it through research, LLM steps, approvals, handoffs, media generation, and publishing. The result is not just agent output, it is a repeatable delivery system.

●Manual, scheduled, and event-driven workflow entry points
●Structured movement across node types and roles
●Artifacts and outputs preserved in the workspace

Feature 3

Managed memory that survives the next run

Most agent setups lose the plot as soon as context shifts. ClawRecipes keeps team memory, instructions, tickets, and artifacts grounded in files the whole system can read. That makes continuity inspectable, reviewable, and reusable across sessions, runs, and team members.

●Durable workspace context instead of hidden state
●Memory-aware execution patterns for long-lived work
●Clean history in git for review and trust

Feature 4

Skills for a team or for one specialist

Some capabilities belong across the whole team. Others belong only to one role. ClawRecipes gives you that flexibility. Install skills at the right scope so teams share the right operating rules while individual agents can still specialize deeply where needed.

●Attach skills to a team-wide operating model
●Give one role deeper tooling or domain behavior
●Keep capability packaging installable and maintainable

Command Reference

List + inspect recipes

openclaw recipes list
openclaw recipes show <recipeId>

Scaffold team + agent

openclaw recipes scaffold-team <recipeId> -t <teamId> --apply-config
openclaw recipes scaffold <recipeId> --agent-id <agentId> --apply-config

Tickets workflow

openclaw recipes tickets --team-id <teamId>
openclaw recipes move-ticket --team-id <teamId> --ticket <ticket> --to testing
openclaw recipes complete --team-id <teamId> --ticket <ticket>

Maintenance

openclaw recipes remove-team --team-id <teamId>
openclaw recipes cleanup-workspaces --yes

Coming soon

Managed AI teams

Want a team designed, installed, and tuned for your company? Join the list for the managed AI team product and service.

Tell us what kind of managed AI team help you want and we'll follow up when early spots open.