Share your thoughts, 1 month free Claude Pro on us
See more
Feedback
Search any
task
Search any
task
SOTA Malicious Skill Detection benchmarks and papers with code | Wizwand
Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Tasks
Malicious Skill Detection
Benchmarks
Dataset Name
SOTA Method
Dataset Name
SOTA Method
Metric
Trend
Results
Last Updated
Adversarial Set 400 skills
HSS
Attack Success Rate (ASR)
0
28
1mo ago
MaliciousAgentSkills
SafeClaw-R
B
9
13
3mo ago
SKILLVETBENCH labeled set 1.0 (78 malicious + 22 benign skills)
SKILLVETBENCH
Balance
0.95
9
1mo ago
ClawHub
SkillVetBench
Overall Detection Rate
95
9
1mo ago
ClawHub Unsafe File Ops 1.0 (n=10)
SkillVetBench
Catch Rate
100
9
1mo ago
ClawHub Prompt Injection 1.0 (n=19)
SkillVetBench
Catch Rate
100
9
1mo ago
ClawHub Command Injection 1.0 (n=27)
SkillVetBench
Catch Rate
100
9
1mo ago
ClawHub Overall 1.0
SkillVetBench
Overall Balance
95
9
1mo ago
MaliciousAgentSkillsBench (404 malicious, 502 benign)
BIV
Recall
100
9
2mo ago
Clean Base Set 100 skills
CISCO
Not Flagged Rate
99
7
1mo ago
100-skill dataset (OpenClaw)
SkillScanner (LLM)
TPR
98
7
1mo ago
MalSkillsBench
Skill Scanner
True Positives (TP)
93
6
3mo ago
906-skill benchmark
Behavioral integrity (BIV)
F1 Score
94.6
4
24d ago
Wild Skills
Cisco (static)
Ori Score
98.6
3
22d ago
SkillJect (Malicious)
SKILLDETONATE
Originality
98
3
22d ago
Reviewed Sample (65 Malicious, 13 Benign)
Locate-and-Judge
True Positives (TP)
62
3
1mo ago
SkillFortifyBench (total)
SkillFortify
Precision
100
1
4mo ago
NPA Skills
-
-
0
1mo ago
Showing 18 of 18 rows
25 / page
50 / page
100 / page
1
Search any
task
Search any
task
Privacy Policy
Terms of Service
FAQs
Swarm Docs