PurpleForge Coverage Engine
Gradient Boosting ranker for prioritizing MITRE ATT&CK detection gaps in Purple Team exercises.
Model Description
| Property | Value |
|---|---|
| Type | GradientBoostingRegressor |
| Version | 1.0.0 |
| Techniques catalog | 39 |
| Detection rules | 30 |
| Training samples | 39 |
Intended Use
Purple Team exercises — prioritize which detection gaps to remediate after attack simulation.
Training Metrics
See train_metrics.json — R² typically > 0.99 on technique feature vectors.
Benchmark (linked dataset)
| Metric | Value |
|---|---|
| Detection recall | 80.4% |
| Gap recall | 100% |
| Avg coverage | 75.2% |
Limitations
- Synthetic exercise data only
- Requires
gap_ranker.joblibartifact at inference - Does not replace commercial BAS/SOAR platforms
Built by Aria AI Security Engineering.