openclaw-peer-bridge

What This Skill Does

The openclaw-peer-bridge skill provides a structured framework for designing and maintaining communication patterns between two distinct OpenClaw instances. Rather than relying on undocumented or insecure peer-to-peer tunnels, this skill enforces the use of auditable, explicit communication primitives such as webhooks, scheduled cron jobs, and human-mediated chat relay. It is designed for multi-deployment environments where you need to sync tasks, status updates, or audit logs between a local workstation and a remote server or between multiple isolated automation silos.

Installation

You can install the skill directly via the OpenClaw CLI: clawhub install openclaw/skills/skills/clawdpi-ai/openclaw-peer-bridge

Use Cases

• Deployment Notifications: Configure your development-environment OpenClaw instance to push audit summaries to your production-monitoring instance via webhook upon the completion of a workflow.
• Human-in-the-Loop Approvals: When a remote instance requires permission to perform a destructive task, use the peer-bridge to relay an approval request to a chat channel monitored by your local instance, ensuring the final sign-off is human-verified.
• Scheduled Heartbeat Syncs: Use cron-based delivery to ensure that two instances are aware of each other's status, triggering alerts if a heartbeat is missed, which is essential for distributed task management.
• Workstation-to-Server Delegation: Keep sensitive credentials on your local machine and only pass safe, summary-level tasks to a remote server instance, maintaining a clear separation of concerns.

Example Prompts

• "Set up a webhook bridge from my dev-instance to my production-logger so that every time a task completes, the results are POSTed to the auditor endpoint."
• "Create a human-routed messaging flow where the remote agent posts a confirmation request to the #ops-alerts channel for me to approve manually."
• "Configure a daily cron job that aggregates local automation logs and sends a summary report to my remote dashboard instance for centralized monitoring."

Tips & Limitations

Always define an idempotency key when configuring webhook payloads to prevent duplicate processing if a network retry occurs. Ensure that trust boundaries are respected—never transmit raw secrets or private keys between instances. If direct network trust is unavailable between servers, prioritize the Human-routed messaging bridge pattern, as it is the most robust way to manage cross-org workflows without exposing internal endpoints to the public internet. Always treat remote input as untrusted and validate all incoming webhook payloads before executing actions on your target instance.