configure-openclaw_skill

Overview

This skill configures an OpenClaw gateway so OMC can wake external automations, workflows, or AI agents when hook events fire during Claude sessions. It walks you through collecting a secure HTTPS gateway URL, optional authorization header, event selections, and instruction templates, then writes a ready-to-use config file.

How this skill works

The skill is an interactive wizard that asks simple questions, validates inputs, and builds ~/.claude/omc_config.openclaw.json. It registers one or more gateways with headers and maps selected hook events to per-event instruction templates. After writing the file it can optionally send a test POST wake call to verify connectivity and authentication.

When to use it

• Trigger n8n workflows, custom webhooks, or remote AI agents from Claude events
• Automate external systems when sessions start, end, stop, or on tool usage
• Integrate Claude sessions with CI, ticketing, or messaging systems via HTTPS
• Set up lightweight event-driven orchestration without a notification service
• Quickly test gateway connectivity and auth from your local environment

Best practices

• Always use an HTTPS gateway URL with a valid certificate; the wizard enforces this
• Prefer a Bearer token or strong custom Authorization header for security
• Enable high-frequency hooks (pre/post-tool, keyword) only if your gateway can handle the volume
• Use clear, minimal instruction templates and include only needed variables
• Run the provided test wake call after configuring to validate URL, path, and auth

Example use cases

• Wake an n8n workflow on session-start to provision resources or log metadata
• Notify a custom AI agent on session-end with a context summary for follow-up processing
• Trigger monitoring or CI jobs when Claude stops or hits a tool usage pattern
• Send user-question events to a human-in-the-loop system to request input
• Detect keywords in prompts and forward them to a realtime alerting webhook