A sophisticated management tool for routing Telegram messages between various Claude model backends and provider types.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install message-routing
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 message-routing 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 Claude Code Switchboard?
The Claude Code Switchboard is a specialized component within the ecosystem of Openclaw Skills designed to provide granular control over AI message processing. It acts as a traffic controller, allowing developers to define exactly how incoming Telegram messages are handled—whether they are dispatched to a local Claude Code CLI instance for deep reasoning or routed to a high-speed API for immediate responses. This flexibility ensures that your AI agents remain responsive and cost-effective.
By centralizing configuration management, this skill eliminates the need for manual file editing. It provides a standardized interface to toggle between model versions like Opus, Sonnet, and Haiku, while also managing the lifecycle of the gateway services required to process these requests. It is an indispensable utility for developers looking to build resilient and adaptable communication workflows using Openclaw Skills.
Claude Code Switchboard Use Cases
- Dynamically switching between high-capability models and low-latency models based on current task requirements.
- Implementing automated fallback logic to maintain service uptime when a primary model provider is unresponsive.
- Debugging message flow issues by accessing real-time gateway logs directly from the command interface.
- Reducing operational overhead by utilizing API-only modes for lightweight tasks within the framework of Openclaw Skills.
- Restoring stable baseline configurations quickly after testing experimental model routing setups.
How Claude Code Switchboard Works
- The user initiates a routing change using the /telegram-routing command to specify a new primary model or fallback state.
- The skill parses the request and accesses the central configuration file located at ~/.openclaw/openclaw.json.
- It updates the agents.defaults.model.primary and fallback fields with the appropriate model references.
- The skill triggers a restart of the OpenClaw gateway service to ensure the new routing logic is active.
- Incoming Telegram messages are intercepted by the gateway, which identifies the configured backend and dispatches the payload to either the local CLI or the Anthropic API.
- Processing results and errors are captured in the gateway logs for verification and troubleshooting within the environment of Openclaw Skills.
Claude Code Switchboard Setup
To integrate this functionality into your collection of Openclaw Skills, follow these configuration steps:
# Check your current routing status and active models
/telegram-routing status
# Switch your primary model to Claude Sonnet for a balance of speed and intelligence
/telegram-routing model sonnet
# If you experience latency, quickly switch to API fallback mode
/telegram-routing fallback
# Validate your configuration if you encounter errors
openclaw doctor
Claude Code Switchboard Data Schema & Taxonomy
The skill organizes routing metadata within the OpenClaw configuration file using a structured hierarchy to ensure compatibility with other Openclaw Skills.
| Key Field | Type | Description |
|---|---|---|
model.primary |
String | The active model reference (e.g., claude-cli/opus or anthropic/haiku). |
model.fallbacks |
Array | A list of secondary models used in sequence if the primary provider fails. |
agents.defaults |
Object | The core configuration block containing routing and provider preferences. |
gateway.log |
File | A local log file located at ~/.openclaw/logs/gateway.log tracking all message traffic. |
Claude Code Switchboard Advanced Features
- Support for shorthand model references that automatically resolve to full provider paths (e.g., opus to claude-cli/opus).
- Automated gateway service orchestration that handles service restarts and configuration validation.
- Lightweight API mode for high-performance scenarios where local CLI overhead is not desired.
- Robust fallback chains that ensure continuity of service across different Openclaw Skills during provider outages.
- Integrated log tailing capabilities to monitor live message processing and routing decisions.
SKILL.md
Loading
Related Openclaw Skills
Intern Report Generator
An intelligent internship documentation suite for generating professional daily logs, weekly summaries, and career reflections.
Predict Skill
A sophisticated local forecasting and prediction system designed to analyze trends, generate scenarios, and assess risks using structured data analysis.
Konektor Agent API
A machine-readable API designed for AI agents and automation tools to manage leads, track conversions, and monitor workspace analytics.
Baby & Maternity Compliance and Data Privacy
A comprehensive framework for baby and maternity brands to automate compliance management and transparent data privacy practices.
Quick Converter
A lightweight, zero-dependency utility for instant offline unit, currency, crypto, and time conversions within AI agent workflows.
Skill Evolver
A comprehensive skill lifecycle manager designed to discover, orchestrate, and evolve reusable AI agent capabilities.
Featured*
