OpenClaw Peer Bridge

---

name: openclaw-peer-bridge

description: Design and operate communication patterns between two OpenClaw instances. Use when a user wants one OpenClaw deployment to send status, tasks, reminders, audit results, or webhook callbacks to another OpenClaw deployment, or when defining a bridge between local and remote OpenClaw agents, dashboards, or automation flows.

---

# OpenClaw Peer Bridge

Design communication between two OpenClaw instances using supported primitives only.

Prefer explicit, auditable flows: webhook delivery, cron delivery, channel routing, or user-mediated relay.

Do not invent peer-to-peer commands or claim native cross-instance session sharing unless it exists in the current deployment.

Core patterns

1. Event push via webhook

Use when one OpenClaw instance needs to notify another system or service immediately.

Pattern:

1. Source OpenClaw runs an isolated cron or task.

2. Delivery mode is `webhook`.

3. Target side exposes an HTTP endpoint that accepts the event and decides what to do next.

Use for:

2. Human-routed messaging bridge

Use when the only guaranteed common surface between two instances is a chat channel.

Pattern:

1. Instance A sends a structured message to a human-controlled channel.

2. Human or downstream automation forwards or triggers Instance B.

3. Keep payloads short, explicit, and idempotent.

Use for:

3. Scheduled relay

Use when exact timing matters more than low latency.

Pattern:

1. Source instance schedules a cron job.

2. Cron produces either an announcement or webhook.

3. Target instance or receiving system processes the scheduled event.

Use for:

4. Local-to-remote control pattern

Use when a workstation OpenClaw needs to coordinate with a server-side OpenClaw.

Recommended split:

Design rules:

Workflow

Step 1: define the bridge objective

State the business outcome first:

Step 2: choose the transport

Choose the smallest sufficient transport:

Step 3: define message contract

Specify:

Prefer JSON for machine processing and short bullet summaries for human review.

Step 4: define trust boundary

Document:

Never assume both OpenClaw instances share memory, sessions, or auth state.

Step 5: implement with reversible changes

Prefer:

Step 6: verify end to end

Check:

Output expectations

When using this skill, produce one of these deliverables:

Guardrails

Quick examples

Example: daily remote server audit to local operator

Design:

Example: deployment completion handoff

Design: