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

Malicious behavior measurement on AgentHarm Harmful

0Harm Rate

Rule Traverse

-1.247.1315.523.87Mar 3, 2026Mar 13, 2026Mar 24, 2026Apr 4, 2026Apr 15, 2026Apr 26, 2026May 7, 2026
Updated 2mo ago

Evaluation Results

MethodLinks
2026.05
010000
2026.05
090.314.61.4
2026.05
0.69646.11.8
2026.05
0.689.235.83.9
2026.05
1.196.6702.4
2026.05
1.782.453.79.5
2026.05
2.395.568.83.1
2026.05
2.894.975.311
2026.05
2.886.950.46.6
2026.05
3.481.337.916
2026.05
5.193.286.86.3
2026.05
5.175.650.812.4
2026.05
5.151.734.621.4
2026.03
69157-
2026.03
69468-
2026.05
6.893.275.614.1
2026.03
79267-
2026.05
7.464.842.723.4
2026.03
88962-
2026.05
889.885.69.1
2026.05
886.984.911.1
2026.03
98752-
2026.03
98659-
2026.03
98863-
2026.03
11011-
2026.05
12.556.366.629.2
2026.05
14.221.652.441.9
2026.03
187458-
2026.05
23.91.159.859.6
2026.05
255888.938.1
2026.05
27.8067.467.4
2026.05
29.11.263.663.3
2026.03
31031-