Setup & Installation
Or with the ClawHub CLI, for registry-managed skill folders outside a full OpenClaw workspace:
Version History
Add send_prompt, land_agent actions and workflow step type docs
What This Skill Does
Tutti orchestrates teams of AI coding agents, Claude Code, Codex, and Aider, from a single declarative config file. Agents run in isolated git worktrees, and the CLI handles launching, dispatching prompts, verifying output, and landing work back to the main branch.
Managing Claude Code, Codex, and Aider from one config file removes the need to manually coordinate separate terminals, branches, and context windows for each agent.
When to use it
- Running parallel feature work across multiple AI agents on separate branches
- Sending a prompt to a specific agent and capturing its output for review
- Running a verification workflow before merging agent-produced code
- Landing an agent's branch or opening a PR when work is complete
- Generating a context handoff packet when an agent's session is near its limit
Example Workflow
Here's how your AI assistant might use this skill in practice.
User asks: orchestrate the team, run the verify workflow, and land the frontend agent's work as a PR
- 1Run doctor_check to verify tt, tmux, and tutti.toml are in place
- 2Run launch_team to start all agents defined in tutti.toml
- 3Poll team_status until agents are idle
- 4Run verify_team --strict to gate on passing verification
- 5Run land_agent frontend --pr to open a pull request from the agent's branch
All agents launched, verification passed, and a pull request opened for the frontend agent's branch
Requirements
Accounts, API keys, or tools you or your AI assistant may need to set up while using this skill.