A technical workflow for configuring deterministic multi-account Feishu routing and isolated agent workspaces within the OpenClaw ecosystem.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install feishu-multiagent
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-multiagent 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 OpenClaw Feishu Multi-Agent Setup?
The OpenClaw Feishu Multi-Agent Setup skill is designed for developers who need to manage multiple Feishu (Lark) bot identities simultaneously. It provides a structured approach to isolating agent personas, system prompts, and long-term memory files. By utilizing this configuration, users can ensure that messages from specific Feishu apps are routed to the correct AI agent without cross-pollination of data or identity confusion.
This setup is essential for enterprise environments where different bots serve different organizational roles. It leverages Openclaw Skills to handle complex routing logic, memory embedding provider overrides, and platform-specific error mitigation for the Feishu Open Platform.
OpenClaw Feishu Multi-Agent Setup Use Cases
- Running multiple Feishu bots with different AppIDs and AppSecrets on a single gateway.
- Resolving routing conflicts where messages are dispatched to the wrong agent or workspace.
- Creating isolated long-term memory contexts for different team assistants.
- Troubleshooting Feishu-specific errors like open_id cross-app issues or contact scope warnings.
- Hardening memory embedding configurations for agents operating in regions with restricted LLM access.
How OpenClaw Feishu Multi-Agent Setup Works
- Inspect the current state of agent bindings and Feishu channel configurations to detect duplicated directories or missing accounts.
- Create new agents with non-interactive CLI commands, ensuring each has a unique workspace and agent directory.
- Bootstrap mandatory workspace files such as SOUL.md and MEMORY.md to define the unique identity and context for the new agent.
- Modify the global configuration to include multiple Feishu accounts and bind them deterministically to specific agent IDs.
- Configure and force-index memory embeddings per agent to ensure the AI has access to isolated historical context.
- Restart the OpenClaw gateway and verify the routing logic through log inspection and explicit account pairing approvals.
OpenClaw Feishu Multi-Agent Setup Setup
First, inspect your current bindings to avoid conflicts:
openclaw agents list --bindings
openclaw config get channels.feishu
Create a new isolated agent workspace:
openclaw agents add assistant-name --non-interactive --workspace ~/.openclaw/workspace-assistant
Bind the specific Feishu account to your new agent:
openclaw agents bind --agent assistant-name --bind feishu:account-id
Finally, validate the configuration and restart the gateway to apply changes using Openclaw Skills tools:
openclaw config validate
openclaw gateway restart
OpenClaw Feishu Multi-Agent Setup Data Schema & Taxonomy
The skill organizes data across several key configuration files and directories to maintain isolation:
| Component | Path | Description |
|---|---|---|
| Global Config | ~/.openclaw/openclaw.json |
Contains the channels.feishu.accounts block and bindings array. |
| Agent Workspace | ~/.openclaw/workspace-name/ |
Root folder for agent-specific identity and prompt files. |
| Memory File | MEMORY.md |
Injected into the system prompt to provide long-term context. |
| Memory Index | memory/ |
Directory for daily memory logs and vectorized embedding chunks. |
| Auth Profiles | auth-profiles.json |
Stores provider-specific API keys for memory embedding services. |
OpenClaw Feishu Multi-Agent Setup Advanced Features
- Deterministic routing of Feishu messages based on unique account-to-agent bindings.
- Granular memory embedding configuration allowing different providers (e.g., OpenAI, OpenRouter) for each agent.
- Support for non-interactive agent creation for automated CI/CD or deployment scripts.
- Automated recovery workflows for Feishu card streaming and contact scope errors.
- Explicit pairing approval mechanisms to prevent cross-account sender allowlist leaks.
SKILL.md
Loading
Related Openclaw Skills
Skill Expert
A comprehensive 6-phase development framework for creating, optimizing, and packaging professional AI agent skills.
Bug Fixing OpenClaw Skill
A comprehensive zero-regression workflow for triaging, reproducing, and fixing bugs within OpenClaw environments.
Smart Image Loader
A versatile utility for AI agents to fetch, manage, and display local or web-based images within their workspace.
Tinmem Memory System
A persistent memory solution that allows AI agents to store and recall user preferences, project details, and interaction history across multiple sessions.
lowkey-viral
Programmatically generate AI-powered short-form vertical videos and photo carousels for social media platforms.
CIFER SDK
A developer-focused SDK for implementing ML-KEM-768 and AES-256-GCM quantum-resistant encryption on Ethereum and Ternoa networks.
Featured*
