A production-grade blueprint for orchestrating a team of AI agents through Feishu using a foreman-worker model and file-driven task queues.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install feishu-multi-agent
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 feishu-multi-agent 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 Feishu Multi-Agent Orchestration?
This skill provides a comprehensive framework for managing a decentralized AI workforce within the Feishu (Lark) ecosystem. By implementing a foreman-worker architecture, it allows one master agent to delegate tasks to multiple specialized subordinate agents. This approach leverages Openclaw Skills to transform simple chat interfaces into a powerful automation hub. The system utilizes file-driven task queues and persistent work cycles to ensure that agents remain productive without manual intervention, effectively bypassing typical workspace sandbox limitations through clever bash-based orchestration.
Feishu Multi-Agent Orchestration Use Cases
- Managing an automated content factory where multiple agents handle research, writing, and publishing.
- Coordinating a software development team where different agents handle frontend, backend, and testing tasks.
- Implementing persistent AI assistants that monitor markets, social media, or internal data and report back to a main channel.
- Setting up a supervisor agent that monitors the health and task status of other agents to ensure 24/7 reliability.
How Feishu Multi-Agent Orchestration Works
- Multiple Feishu bot applications are created and bound to unique agent IDs within the OpenClaw configuration.
- A foreman agent is designated to monitor the task directories and session activities of all subordinate agents.
- Task queues are managed through physical directory structures (pending and done) within each agent's workspace.
- The foreman agent uses bash commands to write task files into the worker workspaces, bypassing standard tool restrictions.
- Subordinate agents follow a heartbeat logic, continuously checking their pending task folders for new instructions.
- Cron jobs are configured to trigger the foreman's patrol logic and the workers' execution cycles at regular intervals.
Feishu Multi-Agent Orchestration Setup
- Create Feishu custom apps for each agent at open.feishu.cn and enable Bot and WebSocket capabilities.
- Configure the
channelsandbindingsin your OpenClaw configuration to map Feishu account IDs to specific agent IDs. - Initialize the task directory structure for each worker agent:
mkdir -p ~/.openclaw/workspace-worker/tasks/{pending,done}
- Define the SOUL.md for the foreman and HEARTBEAT.md for workers to establish the operational logic.
- Register the automated patrol cron job:
openclaw cron add --name "foreman-patrol" --every 30m --agent main --session main --system-event "Execute patrol. Check worker queues and assign tasks."
Feishu Multi-Agent Orchestration Data Schema & Taxonomy
The orchestration relies on a standardized file-based metadata taxonomy and directory structure:
| Path | Purpose | Format |
|---|---|---|
tasks/pending/ |
Queue for upcoming tasks | Markdown files with metadata headers |
tasks/done/ |
Archive of completed tasks | Markdown files moved after execution |
SOUL.md |
Core management instructions | Markdown instructions for the foreman |
HEARTBEAT.md |
Operational loop logic | Markdown instructions for workers |
logs/gateway.log |
Communication logs | Standard OpenClaw gateway logs |
Feishu Multi-Agent Orchestration Advanced Features
- Sandbox Workarounds: Utilizes bash execution to perform cross-workspace file writing where standard write tools fail.
- Isolated Session Cycles: Leverages the isolated session flag in cron jobs to prevent long-term context window degradation.
- Status Reporting: Automatically generates periodic productivity reports for human supervisors every 4 hours.
- Automated Error Recovery: Includes health check routines to identify and restart disconnected Feishu WebSocket clients.
- Custom Heartbeats: Supports varied work frequencies per agent, allowing high-intensity workers to run more often than low-priority observers within Openclaw Skills.
SKILL.md
Loading
Related Openclaw Skills
Feishu App Setup Automation
An automated browser-based skill designed to handle the entire lifecycle of Feishu open platform application configuration for AI agents.
AI Companion Agent for Feishu
A comprehensive framework for deploying an autonomous AI companion on Feishu featuring long-term memory, voice messaging, and image generation.
AI Companion Agent Setup
A comprehensive framework for deploying lifelike AI agents with personality, persistent memory, and multi-modal communication capabilities.
Agent Batch Guard
A comprehensive framework and configuration guide designed to prevent AI agents from freezing during large-scale batch operations by managing session transcript bloat.
Crypto Auto Progression
An automated execution engine that drives the crypto-hedge-backtest project through cron-based task management and reporting.
Guardrails Safe Tools
A security-first execution layer that enforces governed permissions for shell commands, outbound messaging, and external API actions.
Featured*
