openclaw-config-center_skill

Overview

This skill builds a centralized configuration management system for OpenClaw to eliminate hard-coded values and scattered config files. It provides unified config storage, dynamic loaders, automated main-config merging, memory synchronization, multi-agent templates, and cost-optimizing router rules for a single-source-of-truth operations workflow. The goal is "change once, apply everywhere" for safer, faster maintenance.

How this skill works

The system stores modular JSON configs under ~/.openclaw/config and exposes a lightweight shell config-loader that reads values at runtime. A generate-main-config script merges module files into openclaw.json and an update-soul script synchronizes configuration changes into SOUL.md so agent memory and config stay consistent. Templates cover AGENTS.md, memoryFlush/memorySearch, multi-agent routing, and ClawRouter cost controls (model routing, token compression, response caching).

When to use it

• You need one authoritative place for OpenClaw configuration
• You want to stop editing many files when changing behavior
• You manage multiple agents, channels, or skill configs
• You need automated sync between config and agent memory
• You want to reduce API/model costs with smart routing and caching

Best practices

• Keep one JSON per responsibility: core, agents, channels, skills
• Use the provided .example templates and replace placeholders before deployment
• Run generate-main-config.sh after edits and restart gateway to apply changes
• Store secrets only in local protected files and add them to .gitignore
• Use phased rollout: add configs first, run dual-track, then switch traffic

Example use cases

• Centralize feishu bot credentials and update once to apply across agents
• Deploy a new sub-agent by dropping agent JSON into agents/ and merging main config
• Enable memoryFlush and memorySearch templates to prevent memory loss and enable semantic retrieval
• Route requests to cheaper models via ClawRouter templates and cut API spend
• Use AGENTS.md template to standardize session start, memory layers, and safety boundaries