A meta-system for Openclaw Skills that analyzes natural language to auto-select, chain, and execute complex multi-skill workflows.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install mupeng-skill-router
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 mupeng-skill-router 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 Skill Router?
Skill Router acts as the intelligent gateway for the entire agent ecosystem, specializing in intent classification and federated skill composition. By analyzing user input in real-time, it determines whether a request requires a single action or a multi-stage pipeline of Openclaw Skills. The system utilizes a high-performance v2 architecture that scans only the frontmatter of skill files, resulting in an 83% reduction in token consumption compared to traditional methods.
This skill is designed to handle the complexity of modern AI workflows by mapping triggers to specific execution paths. Whether it is managing a simple git commit or orchestrating a five-tier production pipeline involving research, analysis, and deployment, Skill Router ensures that every task is handled by the most qualified sub-skills in the optimal order.
Skill Router Use Cases
- Orchestrating full content pipelines from SEO planning and copywriting to social media posting.
- Managing automated research workflows that monitor competitors, scrape data, and deliver daily reports via email.
- Implementing safe software release cycles by chaining code reviews, git automation, and deployment disciplines.
- Creating comprehensive morning routines that aggregate health monitoring, token usage tracking, and system notifications.
- Facilitating content recycling by analyzing performance metrics and auto-triggering repackaging skills.
How Skill Router Works
- The router scans the frontmatter of all available Openclaw Skills to match triggers using description and trigger fields.
- It identifies the execution script (Shell or JavaScript) defined in the skill metadata.
- Scripts are executed directly using an exec protocol with pre-defined environment variables for workspace, events, and memory.
- Output is processed as JSON or plain text, and event files are generated in the events directory to facilitate data sharing.
- Subsequent skills are triggered automatically based on post-execution hooks or error-handling rules defined in the chain.
Skill Router Setup
To configure the environment for Openclaw Skills routing, ensure your workspace is correctly mapped and the yq utility is available for metadata parsing:
# Define the Openclaw environment variables
export WORKSPACE=$HOME/.openclaw/workspace
export EVENTS_DIR=$WORKSPACE/events
export MEMORY_DIR=$WORKSPACE/memory
# Verify skill triggers by scanning frontmatter
for skill in skills/*/SKILL.md; do
yq eval '.name, .description, .trigger, .run' "$skill"
done
Skill Router Data Schema & Taxonomy
Skill Router organizes data through a structured hierarchy to ensure seamless handoffs between Openclaw Skills:
| Directory | Purpose |
|---|---|
| /events | Stores {type}-{date}.json files used for inter-skill communication and data passing. |
| /memory | Houses persistent data and context logs for long-term learning and evaluation. |
| /workspace | The root execution context where individual skill scripts (run.sh or run.js) reside. |
| Frontmatter | YAML metadata in SKILL.md defining triggers, dependencies, and execution hooks. |
Skill Router Advanced Features
- Complex intent detection that maps a single natural language prompt to a multi-skill chain (1:N mapping).
- Tiered execution graph that organizes skills into Collection, Analysis, Production, Deployment, and Tracking layers.
- Automated hook registration for pre-hook validation, post-hook event generation, and on-error notifications.
- Built-in safety mechanisms including manual approval gates for external actions and infinite loop prevention.
- High-efficiency metadata scanning that scales to dozens of Openclaw Skills while maintaining low token overhead.
SKILL.md
Loading
Related Openclaw Skills
SEO Audit Agent
A specialized diagnostic agent that identifies, explains, and prioritizes SEO issues to improve organic search visibility.
saas-decomposer
A SaaS-to-AIaaS conversion engine that decomposes software functions into autonomous AI-driven tasks.
Release Discipline
A strategic quality gate for AI agents and developers that prevents version spam by enforcing a 24-hour cooldown and mandatory quality checklists.
RAG Engineer
An expert RAG systems architect skill focused on bridging the gap between raw data and LLM understanding through optimized retrieval pipelines.
trend-radar
An automated trend intelligence skill that monitors global tech markets and generates importance-weighted weekly reports.
Web-Claude Unified Search
A high-reliability search orchestration skill that implements a 3-tier fallback strategy using Brave, DuckDuckGo, and browser-based Claude.ai.
Featured*
