Our new X account is live! Follow @wizwand_team for updates

Tool-use agent security evaluation on SIREN

23.56Explicit Directive (UA)

CaMeL

Updated 4d ago

Evaluation Results

Method	Links
CaMeL 2026.01		23.56	44.83	3.8	20.11	17.88	25.14	10.66	30.51	2	0	11.68	25.34	24.87	0	46.79
VIGIL 2026.01		17.24	16.09	52.17	1.09	21.79	24.02	28.31	0	14.67	3.33	27.53	8.13	40.57	0.32	74.49
CaMeL 2026.01		17.24	45.98	4.35	16.3	13.97	37.43	13.97	33.09	1.33	0	10.74	27.84	26.55	0	30.84
Vanilla ReAct 2026.01		15.52	64.94	11.96	69.57	15.08	54.19	12.87	56.62	8.67	48.67	12.93	58.92	30.56	16.65	65.31
VIGIL 2026.01		14.37	22.99	25.54	1.63	17.88	37.99	19.85	0	12.67	2.67	18.46	11.99	39.3	0.21	40.82
DRIFT 2026.01		10.92	49.43	13.59	59.78	13.41	55.87	23.16	5.51	14	4	15.85	33.06	47.63	10.22	55.1
Spotlighting 2026.01		10.34	62.07	10.87	69.02	15.64	48.04	11.4	32.72	9.33	52	11.57	50.89	21.29	9.8	73.47
MELON 2026.01		8.62	63.22	21.74	54.35	12.29	47.49	15.44	3.31	9.33	6	13.87	32.64	24.66	0.42	43.88
DRIFT 2026.01		8.05	62.07	25	28.26	10.06	64.8	17.28	6.25	10	13.33	14.6	32.64	59.75	14.12	76.53
DeBERTa 2026.01		5.75	58.62	15.76	40.76	7.26	40.22	14.71	9.93	4.67	28.67	10.32	33.26	8.85	1.48	34.69
Tool-Filter 2026.01		4.02	40.23	5.43	1.09	5.59	22.91	11.4	13.6	7.33	18.67	7.19	18.56	6.53	2.32	48.98
Vanilla ReAct 2026.01		3.45	88.51	26.09	71.2	25.7	65.36	12.13	75.37	13.33	67.33	15.95	73.83	39.52	38.88	79.59
Tool-Filter 2026.01		2.87	49.43	3.26	0.54	2.79	22.35	9.56	0.37	4.67	43.33	5.11	20.13	7.48	0.11	45.92
Spotlighting 2026.01		2.3	87.93	52.72	40.22	29.05	74.3	11.03	58.82	8	60	20.33	63.61	43.94	39.83	77.55
DeBERTa 2026.01		2.3	90.23	16.3	58.15	10.61	66.48	15.81	7.72	4	32.67	10.64	47.24	21.29	8.11	43.88
MELON 2026.01		1.72	87.93	46.74	37.5	24.58	61.45	18.38	0.74	2.67	12	19.5	36.7	35.63	0.21	71.43