A powerful orchestration skill for managing flexible, multi-role AI agent development teams ranging from 2 to 10 members.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install multi-agent-team-by-halfmoon82
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install multi-agent-team-by-halfmoon82 using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is Multi-Agent Dev Team v2.2?
The Multi-Agent Dev Team v2.2 skill is a comprehensive framework designed to scale OpenClaw capabilities from single-agent tasks to full-scale team orchestration. By utilizing Openclaw Skills, users can deploy between 2 and 10 specialized agents to work collaboratively on complex software projects. The skill supports multiple named teams coexisting within a single instance, each with its own independent configuration and role assignments.
This version introduces advanced logic for model auto-detection, allowing the system to match specific roles—like Architects or Frontend developers—with the most appropriate LLM based on reasoning strength or coding specialization. It transforms how developers interact with AI by providing structured workflow templates and a robust protocol for agent spawning and lifecycle management.
Multi-Agent Dev Team v2.2 Use Cases
- Establishing a full-stack development team with dedicated roles for PM, Architect, and QA.
- Managing parallel development squads (e.g., a dedicated frontend team and a backend team) with independent task tracking.
- Implementing strict project control through a standard 9-step collaboration workflow.
- Rapidly deploying lean 3-person teams for hotfixes and agile feature iterations.
How Multi-Agent Dev Team v2.2 Works
- Run the interactive setup wizard to define team names and select between 2 to 10 agent roles.
- Assign specific model types to each role using the auto-detection feature which scans your openclaw.json config.
- Select a collaboration template (standard, quick, solo, or custom) to define how agents pass tasks between one another.
- The wizard automatically updates the OpenClaw configuration and initializes the required workspace directory structure.
- Activate the team using the wake-up protocol commands in the chat interface.
- Orchestrate subagent tasks via the sessions_spawn method, ensuring agents operate within their defined responsibilities.
Multi-Agent Dev Team v2.2 Setup
To integrate this capability into your Openclaw Skills environment, execute the setup wizard provided in the skill directory.
# Navigate to your skill directory and run the wizard
node wizard/setup.js
# For multi-team support, specify a team name
node wizard/setup.js --team alpha
After configuration, restart the gateway to apply changes:
openclaw gateway restart
Multi-Agent Dev Team v2.2 Data Schema & Taxonomy
The skill organizes its operations using a hierarchical file structure within the workspace to maintain team manifests and project history.
| Location | Data Type | Purpose |
|---|---|---|
teamtask/teams/*.json |
JSON Manifests | Definitions for team roles, models, and workflows. |
teamtask/tasks/ |
Task Logs | Storage for project-specific execution data. |
agents/ |
Agent Config | Subagent SOUL files and team-prefixed ID configurations. |
memory/ |
Telemetry | Weekly optimization summaries and skill usage analytics. |
Multi-Agent Dev Team v2.2 Advanced Features
- Multi-team co-existence allowing unique prefixes like alpha-frontend and beta-backend to avoid ID collisions.
- Weekly optimization tasks via OpenClaw Cron that analyze telemetry to automatically update skill mappings.
- Mandatory timeout governance providing graded retries, circuit breakers, and explicit failure classification.
- Model fallback chains that automatically switch providers if the primary model hits rate limits or is unavailable.
- Skill learning telemetry that tracks agent_id, skill_name, and context for continuous performance analysis.
SKILL.md
Loading
Related Openclaw Skills
Model Failover Doctor
A diagnostic and repair utility designed to resolve persistent All models failed errors and routing mismatches within the OpenClaw ecosystem.
Config Preflight Validator
A specialized local validation utility designed to verify configuration files and patches against the official schema before deployment.
Complex Task Three-Step Methodology
A sophisticated three-step framework designed to guide AI agents through zero-cost screening, lightweight assessment, and deep DAG-based planning for complex task execution.
timely
A command-line interface for Apple Reminders that supports natural language dates and location-based geofencing triggers.
Skill Priority Setup
Automated skill tiering and injection policy configuration tool for optimizing Openclaw Skills performance.
Skill Quick Index
A high-speed discovery and indexing tool that maps natural language intents to specific local AI agent functions.
Featured*
