A centralized management system for distributing and synchronizing AI agent capabilities across different coding platforms using a 2-layer architecture.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install skill-manager-usm
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 skill-manager-usm 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 Manager?
The Skill Manager is the heart of the Universal Skill Manager (USM) architecture, designed to solve the fragmentation of capabilities across various AI environments like Cursor, Claude Code, and Openclaw. By maintaining a Single Source of Truth, it ensures that your Openclaw Skills are consistently available, versioned, and correctly scoped across all your development tools.
Built on a robust 2-layer, 2-dimension framework, this skill uses symbolic links to bridge global skill hubs with specific agent directories. This allows developers to manage a single physical file while the Skill Manager handles the complexity of visibility and platform-specific distribution through automated synchronization scripts.
Skill Manager Use Cases
- Synchronizing newly created or installed skills across all active AI agent platforms.
- Restricting specific Openclaw Skills to a single environment like Cursor or Claude Code.
- Provisioning metadata and symlinks for global or project-specific skill hubs.
- Auditing which capabilities are currently exposed to a particular AI assistant.
- Expanding the USM ecosystem by adding support for new AI agent platforms via configuration.
How Skill Manager Works
- The manager identifies the source skill in the Global Hub (~/.skills/) or Project Hub (./.skills/).
- It reads the meta.yaml file associated with the skill to determine its scope, such as universal or agent-specific.
- It references agents.yaml to locate the installation directories for registered platforms.
- The system generates or updates symlinks in the target agent directories pointing back to the hub files.
- The sync_skills.sh script is executed to finalize the distribution and ensure the environment is up to date.
Skill Manager Setup
To initialize and sync your Openclaw Skills, use the following commands:
# Run the standard synchronization
bash ~/.skills/skill-manager/scripts/sync_skills.sh
# Synchronize including a specific project directory
bash ~/.skills/skill-manager/scripts/sync_skills.sh --project-dir <path-to-your-project>
Ensure your agent platforms are defined in ~/.skills/agents.yaml before running the sync.
Skill Manager Data Schema & Taxonomy
The Skill Manager organizes data through a hierarchical structure of YAML configurations and directory mappings:
| File/Folder | Description |
|---|---|
~/.skills/ |
The Global Hub containing the physical source of all Openclaw Skills. |
meta.yaml |
Located in each skill folder; defines name, version, and agent scope. |
agents.yaml |
Defines the mapping of agent names to their respective local skill directories. |
scripts/ |
Contains the core logic for symlink generation and automated cleanup. |
references/ |
Technical schemas and provisioning instructions for the USM architecture. |
Skill Manager Advanced Features
- Support for 2-Layer Hubs allowing both global machine-wide skills and project-specific overrides.
- Granular Dimension Control to toggle skill visibility between 'universal' and specific agent lists.
- Progressive Disclosure workflow integration with specialized provisioning agents for complex setups.
- Automated handoff handling for skills generated by external installers or creators.
SKILL.md
Loading
Related Openclaw Skills
Resume Builder
An expert AI agent skill for drafting, tailoring, and rendering executive-grade white-collar resumes to high-fidelity PDF using Typst.
Muninn Memory System
Muninn Memory System is a production-grade semantic memory layer providing high-accuracy knowledge graphs and temporal reasoning for AI agents.
xz-data
A comprehensive data retrieval skill for real-time commodity market pricing, inventory, and supply chain analytics across 25 industries.
OpenClaw Upgrade Standard
A production-grade upgrade protocol for OpenClaw that prevents silent failures through automated backups, service migration, and rigorous post-upgrade testing.
MongoDB Skill
A comprehensive toolset for managing MongoDB databases through natural language, covering everything from basic CRUD to complex aggregation and geospatial analysis.
Evenium
A professional integration that enables AI agents to manage Evenium events, registrations, and attendee data through a secure CLI-driven workflow.
Featured*
