CVSS-depth calculator. 13 metrics. Live scoring. Shareable vector string.
Think CVSS, but for AI operational risk. Score a pattern across impact, exploitability, autonomy level, sector criticality, and observability. Draft methodology, iteration welcome.
Loading scorer…
Intrinsic properties of the risk, independent of deployment context.
Significant harm; widespread or hard to reverse.
Easily exploitable with basic skill or common tools.
Humans monitor and can intervene; AI acts within bounds.
Exploitable remotely over a network without adjacency.
No privileges or authentication required.
Impact propagates beyond the vulnerable component.
Deployment-specific factors that amplify or reduce risk.
Regulated industry: finance, healthcare, legal, etc.
Sensitive business or personal data; restricted access.
Serves multiple organizations or customer tenants.
Random or statistical sampling of outputs.
How current, proven, and fixable is this risk right now?
Actively exploited in the wild.
An unofficial workaround exists but no official fix.
Confirmed by vendor or authoritative third party.
AIOR Score
Score Breakdown
Final = (Base + Env) × Temporal
Vector String
AIOR:1.0/I:H/E:H/A:O/AV:N/PR:N/SC:C/SE:R/DS:C/SCL:M/HO:S/EM:W/RL:W/RC:CShareable URL
Draft methodology, v1.0. Scoring uses a weighted multiplicative model inspired by CVSS v3. Weights and formula subject to iteration. CVSS reference