Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Agent defense evaluation on AgentDojo

69.15Utility under Attack

Sanitizer

19.219632.182345.14558.1077Oct 6, 2025
Updated 26d ago

Evaluation Results

MethodLinks
2025.10
69.1574.090
2025.10
68.68-0.47
2025.10
67.2585.5327.82
2025.10
58.5958.410
2025.10
56.2873.136.84
2025.10
55.6471.6641.65
2025.10
54.553.60
2025.10
52.4676.291.27
2025.10
50.0370.0118.15
2025.10
50.0183.0257.69
2025.10
32.9168.040.95
2025.10
21.1441.497.95