Developer demand for AI tools outpaces MCP governance, creating shadow usage

The MCP ecosystem is seeing increased developer demand for tools that outpace formal governance, particularly in IDE integrations. While the catalog of risk-classified servers remains stable at 73, platforms and security teams face challenges in managing developer usage of unapproved services. Key areas of concern include GitHub Copilot MCP, OpenAI MCP (for GPT-4o, DALL-E, etc.), Figma MCP, and Anthropic Claude MCP, with a focus on visibility into token spend, code review for agent-generated outputs, and team-based allowlisting. AI

IMPACT Highlights the growing gap between developer adoption of AI tools and organizational governance, impacting IT security and cost management.

RANK_REASON The article discusses the integration and governance of AI tools within a specific ecosystem (MCP), focusing on developer adoption and platform management challenges, rather than a new release or core research.

Read on dev.to — MCP tag →

• Anthropic Claude MCP
• Claude
• Claude Code
• Cursor
• DALL-E
• Figma MCP
• GitHub Copilot
• GitHub Copilot MCP
• GitHub MCP
• GPT-4o
• MCP
• OpenAI MCP
• Windsurf

• product
• policy

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →