STATE OF MCP SECURITY

Every public MCP server, graded

Automated, reproducible security grades for MCP servers pulled from the public registry. Each report shows exactly what failed and how to reproduce it. We read the server's own advertisement only, and never call a tool.

18
servers scanned
72%
graded D or worse
2
clean (grade A)
9
grade F
A ai.advisorsai/store-readiness http:https://advisorsai.ai/store-readiness-mcp clean A ai.agenticterminal/directory http:https://mcp.agenticterminal.ai/mcp clean B agency.kesey/pretrip npm:pretrip-mcp 1 med B ai.afmr/discovery http:https://afmr.ai/api/rpc 1 med B ai.afmr/discovery http:https://afmr.ai/mcp 1 med D ai.adeu/adeu npm:@adeu/mcp-server 1 high 2 med E ai.aetherwealth/mcp npm:@aetherwealth/mcp 1 high 12 med E ai.agenticaffiliate/affiliate-networks-mcp npm:affiliate-networks-mcp 3 high 13 med E npm:@modelcontextprotocol/server-filesystem npm:@modelcontextprotocol/server-filesystem 9 med F ac.inference.sh/mcp http:https://api.inference.sh/mcp 1 crit 3 med F ac.tandem/docs-mcp http:https://tandem.ac/mcp 1 crit 2 med F ad.inside/inside-ads http:https://app.inside.ad/api/mcp 1 crit 2 med F agency.goji/goji http:https://mcp.goji.agency/mcp 1 crit F ai.aarna/atars-mcp http:https://mcp.aarna.ai/mcp 1 crit 1 med F ai.adoraads/beauty http:https://mcp.adoraads.ai/beauty 1 crit 1 high 1 med F ai.agentberg/agentberg http:https://agentberg.ai/mcp 1 crit 1 med F ai.agenticfabricationnetwork/ufp http:https://agenticfabricationnetwork.ai/mcp 1 crit 4 med F ai.agentlookups/counterscript http:https://rx.agentlookups.ai/mcp 1 crit 1 med

Point-in-time automated analysis of each server's public metadata (initialize and tools/list). Grades reflect deterministic checks only. Generated 2026-08-29. A server owner who wants a listing corrected or removed can get in touch.