PulseAugur
EN
LIVE 06:58:13

Analysis finds 99.9% of MCP servers untestable in production

A recent analysis of 22,561 Multi-Agent Conversation Protocol (MCP) servers revealed a significant trust gap, with only 18 servers being independently verifiable at runtime. The dominant method of vetting these servers relies on static code scans, which do not assess the actual behavior or reliability of the live service. This leaves a vast majority of MCP servers as black boxes, making it difficult for agents to ensure the tools they call are dependable and secure against runtime attacks like tool poisoning. AI

IMPACT Highlights critical security and reliability gaps in the agent tool ecosystem, pressuring developers to implement runtime verification.

RANK_REASON Analysis of a technical protocol's ecosystem security and verifiability. [lever_c_demoted from research: ic=1 ai=0.7]

Read on dev.to — MCP tag →

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

COVERAGE [1]

  1. dev.to — MCP tag TIER_1 English(EN) · Dinesh Kumar ·

    We tried to independently verify all 22,561 MCP servers. Only 18 could be checked.

    <p>I maintain a deduplicated index of 22,561 MCP servers. I tried to independently verify all of them at runtime. Not by scanning the source in the repo, but by actually reaching the running server to check it responds and behaves. Only 18 could be checked.</p> <p>Here is what th…