StealthScore workbench
Transparent, rubric-driven risk scoring from the active case-state store.
StealthScore
sigreduce-rubric-0.3-demo · authorized_local
78baseline→ 84 projected
Higher is safer. Baseline reflects current exposure; projected assumes the remediation plan completes. Calibration is case-derived.
Data mode
authorized_local
configured
Baseline
78
Projected
84
case-derived
Avg finding risk
22
0–100
Projected drivers
What moves the score after remediation
- SELF-DISCOVERY-001: Schedule recurring monitoring and add deeper source checks only after explicit approval.+6
- SELF-INTAKE-001: Keep approval gates active and require explicit consent before any external remediation.+5
Scoring rubric
Open-core, deterministic weights (SigReduce)
Severity
22%
Confidence
12%
Physical proximity
18%
Fraud utility
14%
Family exposure
12%
Social-engineering leverage
12%
Recurrence likelihood
10%
Per-finding risk scores
Sorted by weighted score
| ID | Finding | Severity | Dominant driver | Score |
|---|---|---|---|---|
| SELF-DISCOVERY-001 | Exact-name public search returned no direct match in this run | Low | Confidence | 22 |
| SELF-INTAKE-001 | Self-authorized PED scope captured | Low | Confidence | 22 |