MCP Security Bench (MSB)

Evals

Benchmark / dataset

Attacks against Model Context Protocol in agents.

Category
Prompt injection & agent security
Insurance exposure
Prompt injection
Publisher
Academic
Publisher type
Academic
License
MIT
Status
Maintained (<12mo)
Last commit
2026-03-24
Stars
51
Notes
ICLR 2026.

View source on GitHub →

Wanich Risk Lab, 2026.