A recent analysis of 46 popular MCP servers on Smithery revealed significant shortcomings in their tool descriptions. Approximately 69% of the 1,118 tools examined failed to specify their intended use, and 45% did not describe their return values. Furthermore, descriptions for tools within the same server were often so similar that AI agents could confuse them, with 15 servers exhibiting this issue across 90 pairs of tools. This lack of clarity can lead to incorrect tool selection by AI agents, as demonstrated by examples from PubMed and Gmail. AI
IMPACT Lack of clear tool descriptions hinders AI agent functionality and increases the risk of errors in automated tasks.
RANK_REASON Analysis of existing tools and their descriptions. [lever_c_demoted from research: ic=1 ai=0.7]
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →