What changed
OpenAgentSkill just added 500 new high-star skills from GitHub. The batch focuses on agent frameworks, browser automation, RAG, data workflows, developer tooling, and practical automation projects with at least 500 stars.
This matters because agent builders usually do not need another generic directory. They need a fast way to spot tools that already have community signal, understand where each tool fits, and decide whether it belongs in a real workflow.
Notable additions
Agent and orchestration workflows
Start with Hermes Agent, AI Agents for Beginners, Deer Flow, and Agno if you are comparing frameworks, examples, and multi-step agent patterns.
Browser and web automation
Firecrawl, Browser Use, Crawlee Python, and SeleniumBase are useful when agents need to collect, test, or transform web data instead of only chatting about it.RAG, memory, and knowledge workflows
Chroma, GraphRAG, LightRAG, Onyx, and Haystack give the batch a strong retrieval and knowledge-management spine.Developer productivity
Daytona, Vibe Kanban, Claude Code Tools, and RepoMaster are worth checking if your agent workflow touches code review, environments, project planning, or repeated repository tasks.How to use the update
Use New Arrivals to scan the freshest imports. Switch to Most Starred when you want stronger community proof. For production evaluation, open the skill page, inspect the GitHub repository, then test the install command in a small agent workflow before promoting it into your daily setup.
What is next
The importer will keep expanding the marketplace while filtering out protocol-server-only projects. The blog will now track these launches as OpenAgentSkill Update: shorter launch notes for discovery, plus deeper scenario guides for skills that deserve a closer look.