List your company

Red-teaming tool-calling agents on MCP Filesystem 1.0 (test)

68.8Success Rate (RR)Zero-Shot (ZS)
-1.08817.05635.253.344Mar 21, 2026

Evaluation results

MethodLinks
Zero-Shot (ZS)Mode=Zero-Shot2026.0368.87.8
Multi-Turn (MT)Mode=Multi-Turn2026.0337.531.2
Iterative Refinement (IR)Mode=Iterative Refinement2026.032540.6
Standard Evolution (SE)Mode=Evolutionary Search2026.0310.971.9
T-MAPBackbone=DeepSeek-V3.2, Target Model=GPT-5-mini2026.031.684.4