Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Malicious-skill detection on 906-skill benchmark
Loading...
94.6
F1 Score
Behavioral integrity (BIV)
41.976
55.638
69.3
82.962
Jun 30, 2026
F1 Score
Recall
False Positive Rate (FPR)
Updated 24d ago
Evaluation Results
Method
Method
Links
F1 Score
Recall
False Positive Rate (FPR)
Behavioral integrity (BIV)
axis=behavior
2026.06
94.6
97.8
7.2
LLM-only audit
axis=behavior
2026.06
93
-
-
Structural fingerprint
axis=identity
2026.06
73.5
81
32
Rule-based scanner
axis=n/a
2026.06
44
-
-
Feedback
Search any
task
Search any
task