Malicious Package Detection on Evaluation Dataset
99.5AccuracyPYGUARD
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| PYGUARDTechnology=Knowledge-driven2026.01 | 99.5 | 99.9 | 99.08 | 99.49 | 2 | 19 | |
| GuardDogTechnology=Rule2026.01 | 94.12 | 92.38 | 94.42 | 93.39 | 141 | 101 | |
| SocketAITechnology=LLM2026.01 | 90.95 | 92.01 | 88.12 | 90.03 | 161 | 250 | |
| CerebroTechnology=ML2026.01 | 89.12 | 90.21 | 85.88 | 88 | 196 | 297 | |
| HerculeTechnology=Dynamic2026.01 | 88.76 | 90.67 | 84.47 | 87.46 | 183 | 327 | |
| SAP-RFClassifier=Random Forest, Technology=ML2026.01 | 86.74 | 92.2 | 78.05 | 84.54 | 139 | 462 | |
| SAP-XGBClassifier=XGBoost, Technology=ML2026.01 | 65.55 | 80 | 34.39 | 48.11 | 181 | 1,381 | |
| SAP-DTClassifier=Decision Tree, Technology=ML2026.01 | 59.99 | 68.68 | 25.42 | 37.1 | 244 | 1,570 | |
| PyPI WareCheckTechnology=Rule2026.01 | 52.99 | 49.66 | 99.19 | 66.18 | 2,117 | 17 | |
| OSSGadgetTechnology=Rule2026.01 | 50.01 | 48.13 | 88.38 | 62.32 | 2,000 | 244 | |
| Bandit4MalTechnology=Rule2026.01 | 33.37 | 34.72 | 48.69 | 40.54 | 1,927 | 1,080 |