ClawHub

ClawHub is a specialized skill registry designed for AI agents, serving as a centralized repository where developers and users can discover, share, and manage reusable capabilities that extend the functionality of agentic systems. The platform organizes skills into distinct categories including MCP Tools, Prompts, Workflows, Dev Tools, Data & APIs, Security, and Automation, enabling users to find relevant components for their specific use cases. Each skill entry is accompanied by metadata such as star count, download count, installation count, and last update timestamp, allowing the community to gauge popularity and recency. Users can sort skills by featured status, recommendation, most downloaded, most starred, most installed, recently updated, newest, or name, providing flexible discovery paths. The registry supports vector search, which means queries can be matched against skill descriptions semantically rather than relying solely on exact keyword matches, improving retrieval accuracy for agent tasks. Every skill is contributed by a publisher identified by their GitHub handle, and the platform integrates directly with GitHub for authentication, fostering an open ecosystem where skills can be versioned, forked, and improved collaboratively. For example, one published skill named “self-improving agent” captures learnings, errors, and corrections to enable continuous improvement for an AI agent; it is described for use when a command or operation fails unexpectedly or when a user corrects Claude. Another skill, “Self-Improving + Proactive Agent,” bundles self-reflection, self-criticism, self-learning, and self-organizing memory, allowing an agent to evaluate its own work, catch mistakes, and improve permanently. A third example, “Skill Vetter,” focuses on security-first