Independent Security Ratings
for AI Tools

SpiderRating scores MCP servers, Claude skills, and AI tools across three dimensions: description quality, security analysis, and metadata health. Deterministic. Transparent. Open source.

Browse DirectoryHow It Works
235
Items Rated
46+
Security Rules
5.0/ 10
Avg Score
100%
Open Source

Top Secure

View all →
#NameScore
1
Afrexai Accessibility EngineSkill1kalin
C6.9
2
Afrexai Agent EngineeringSkill1kalin
C6.5
3
Byr CliSkill1morebuild
C6.4
4
Kimi Agent PolicySkill2233admin
C6.3
5
Mapping SkillSkill16miku
C6.3

High Risk

View all →
#NameScore
1
Pydantic AiMCPpydantic
F1.7
2
AiMCPvercel
F1.8
3
Agent ToolkitMCPstripe
F1.9
4
Doppel Erc 8004Skill0xm1kr
F2.0
5
Universal TradingSkill0xmomo-ngclubs
F2.0

Grade Distribution

A
0 (0%)
B
0 (0%)
C
168 (71%)
D
34 (14%)
F
33 (14%)

Most Common Security Issues

Sandbox Not Configured208 servers (88.5%)
Dangerous Eval158 servers (2.1%)
SSRF130 servers (2.1%)
Prototype Pollution38 servers (0.9%)
Child Process Injection25 servers (0.9%)