An analysis of over 12,500 MCP and agent skill repositories on GitHub reveals a highly polarized ecosystem. The top 1% of repositories command nearly half of all stars, indicating extreme concentration at the top. A significant portion, 62.8% of MCP servers, have fewer than 100 stars, suggesting a lack of a robust middle class in this development space. Agent skills, however, show more recent activity and higher median star counts compared to servers, though the overall distribution remains skewed. AI
IMPACT Reveals a highly concentrated development landscape, suggesting potential challenges for new projects and developers seeking broader visibility.
RANK_REASON Analysis of public GitHub data on MCP and agent skill repositories. [lever_c_demoted from research: ic=1 ai=0.7]
- anthropics/skills
- GitHub
- Hacker News
- headroomlabs-ai/headroom
- MCP
- MCPMarkets
- obra/superpowers
- October 17, 2026
- openclaw/openclaw
- punkpeye/awesome-mcp-servers
- sansan0/trendradar
- vinta/awesome-python
- X
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →