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

Agent Safety Evaluation on Agent-SafetyBench aggregated clean and five attack types

26.31UBR

SAFEHARNESS

Updated 3d ago

Evaluation Results

Method	Links
SAFEHARNESS 2026.04		26.31	20.2	77.92	59.4	3.8	709
SAFEHARNESS 2026.04		28.74	26.4	94.25	67.5	3.48	1,152
System-Prompt 2026.04		31.88	29.2	91.75	62.3	3.14	0
SAFEHARNESS 2026.04		33.76	28.7	83.67	54.4	2.96	678
SAFEHARNESS 2026.04		34.59	32.3	95.42	62.9	2.89	1,137
SAFEHARNESS 2026.04		38.14	36.9	94.83	57.5	2.62	1,181
SAFEHARNESS 2026.04		39.18	38	95.08	56.5	2.55	1,477
System-Prompt 2026.04		39.25	35.5	91.5	55.2	2.55	0
LlamaFirewall 2026.04		39.73	33.9	86.83	54	2.52	148
System-Prompt 2026.04		40.67	38.7	94.67	55.7	2.46	0
Guardrail 2026.04		40.79	34.8	86.83	52.6	2.45	82
System-Prompt 2026.04		43.08	42.6	96.92	54.3	2.32	0
Guardrail 2026.04		44.8	41.8	90.58	48.5	2.23	54
Unprotected 2026.04		44.99	39.4	86.5	48.1	2.22	0
LlamaFirewall 2026.04		45.99	42.8	91.33	48.6	2.17	96
LlamaFirewall 2026.04		46.83	44.8	94.67	49.5	2.14	168
Guardrail 2026.04		47.19	45	95	49.9	2.12	88
Unprotected 2026.04		48.22	45.2	91.25	46	2.07	0
System-Prompt 2026.04		49.52	48.4	96.42	47.7	2.02	0
Unprotected 2026.04		51.18	49.2	95.25	45.6	1.95	0
LlamaFirewall 2026.04		51.26	49.5	95.75	46.4	1.95	184
Guardrail 2026.04		52.81	51.3	96.42	44.7	1.89	67
System-Prompt 2026.04		54.2	53	95.33	41.9	1.85	0
LlamaFirewall 2026.04		56.51	54.3	95.42	40.6	1.77	182
Guardrail 2026.04		56.54	55.7	96.25	40.5	1.77	75
LlamaFirewall 2026.04		56.72	55.3	94.92	39.3	1.76	215
Unprotected 2026.04		57.32	56.2	95.67	39.1	1.74	0
Guardrail 2026.04		59.26	56.9	94.92	37.4	1.69	99
Unprotected 2026.04		61.83	61.6	95.42	33.6	1.62	0
Unprotected 2026.04		61.98	60.9	95.33	33.9	1.61	0