Risk Graphprototype

Gap board

Absence as a first-class object: 31findings the graph can see about what it can't see. Every gap gets an owner and becomes an Action.

Broken lineage · 8

“ATO attempts / 10k logins” on RSK-014

critical

Bound to auth_events (warehouse), but the source or observations are stale. Reports using this metric are blocked until resolved or annotated.

→ Fix the source refresh or re-bind the metric.

“Successful ATO rate” on RSK-014

critical

Bound to auth_events (warehouse), but the source or observations are stale. Reports using this metric are blocked until resolved or annotated.

→ Fix the source refresh or re-bind the metric.

“MFA adoption” on RSK-014

critical

Bound to auth_events (warehouse), but the source or observations are stale. Reports using this metric are blocked until resolved or annotated.

→ Fix the source refresh or re-bind the metric.

“90-day churn of ATO'd accounts” on RSK-014

critical

Bound to Fraud ops board (Looker), but the source or observations are stale. Reports using this metric are blocked until resolved or annotated.

→ Fix the source refresh or re-bind the metric.

“Mean time to lockout” on RSK-014

critical

Bound to Fraud ops board (Looker), but the source or observations are stale. Reports using this metric are blocked until resolved or annotated.

→ Fix the source refresh or re-bind the metric.

“Privileged exports / week” on RSK-033

critical

Bound to auth_events (warehouse), but the source or observations are stale. Reports using this metric are blocked until resolved or annotated.

→ Fix the source refresh or re-bind the metric.

“DSAR responses on time” on RSK-041

critical

Bound to Fraud ops board (Looker), but the source or observations are stale. Reports using this metric are blocked until resolved or annotated.

→ Fix the source refresh or re-bind the metric.

“Chargeback rate” on RSK-052

critical

Bound to Chargeback spreadsheet, but the source or observations are stale. Reports using this metric are blocked until resolved or annotated.

→ Fix the source refresh or re-bind the metric.

Unmeasured risk · 1

RSK-021 · SMS pumping / artificially inflated traffic

critical

No KRIs are defined for this risk — exposure is asserted, not observed.

→ Add starter KRIs from the library and bind them to a registered source.

Unwatched vector · 2

SMS pumping / AIT

critical

No detections cover this attack vector (monetization stage). Linked: RSK-021.

→ Add a detection and register its TPR/FPR metrics.

Promo abuse

critical

No detections cover this attack vector (monetization stage). Linked: RSK-052.

→ Add a detection and register its TPR/FPR metrics.

Paper policy · 1

POL-05 · Acceptable Messaging Use Policy

critical

No control enforces this policy. It states a rule nothing checks.

→ Link the controls that enforce it, or open an Action to build one.

Vendor review lapsed · 2

VND-01 · Primary cloud provider (tier 1)

critical

Last due-diligence review 2025-07-24, and it holds customer PII. $4.04M of residual exposure runs through it (RSK-008, RSK-041).

→ Run due diligence: security questionnaire, SOC 2 report, and continuity terms.

VND-04 · Offshore support BPO (tier 2)

warning

Never reviewed, and it holds customer PII. $2.54M of residual exposure runs through it (RSK-033, RSK-041).

→ Run due diligence: security questionnaire, SOC 2 report, and continuity terms.

Hand-fed metric · 1

“Vendor SLA breaches / quarter” on RSK-008

warning

Manual entry, last observed 2026-09-24 (58 days ago).

→ Bind to a registered data source so observations flow automatically.

Stale assessment · 1

RSK-008 · Vendor concentration — cloud region outage

warning

Last assessed 2025-09-18 (429 days ago).

→ Reopen the assessment cycle.

Unmapped requirement · 1

ISO 27001 A.8.16 · Monitoring activities

warning

No control is mapped to this requirement.

→ Map an existing control or open an Action to build one.

Untested control · 1

CTL-112 · Credential-stuffing WAF rules

warning

No test evidence in 229 days — effectiveness decayed to 15% in every linked residual score.

→ Run a control test and attach evidence.

Stale source · 4

auth_events (warehouse)

warning

Expected refresh every 1d; last refreshed 2026-11-10 (11 days ago).

→ Fix the pipeline or reassign the owner.

Fraud ops board (Looker)

warning

Expected refresh every 1d; last refreshed 2026-11-09 (12 days ago).

→ Fix the pipeline or reassign the owner.

SDAT domain-risk API

warning

Expected refresh every 1d; last refreshed 2026-11-10 (11 days ago).

→ Fix the pipeline or reassign the owner.

messaging_billing (warehouse)

warning

Expected refresh every 1d; last refreshed 2026-11-10 (11 days ago).

→ Fix the pipeline or reassign the owner.

Unmodeled technique · 8

Fraud kill chain FKC-1 · Synthetic identity creation

warning

No attack vector in the register maps to this account creation technique, so nobody has decided whether it applies here.

→ Model it as a vector and link the risks it would realize, or record why it does not apply.

Fraud kill chain FKC-6 · Cash-out through resold balances

warning

No attack vector in the register maps to this cash-out technique, so nobody has decided whether it applies here.

→ Model it as a vector and link the risks it would realize, or record why it does not apply.

MITRE ATT&CK T1078 · Valid accounts

warning

No attack vector in the register maps to this initial access technique, so nobody has decided whether it applies here.

→ Model it as a vector and link the risks it would realize, or record why it does not apply.

MITRE ATT&CK T1566.002 · Phishing: spearphishing link

warning

No attack vector in the register maps to this initial access technique, so nobody has decided whether it applies here.

→ Model it as a vector and link the risks it would realize, or record why it does not apply.

MITRE ATT&CK T1530 · Data from cloud storage

warning

No attack vector in the register maps to this collection technique, so nobody has decided whether it applies here.

→ Model it as a vector and link the risks it would realize, or record why it does not apply.

Messaging abuse MA-02 · Smishing from compromised accounts

warning

No attack vector in the register maps to this phishing technique, so nobody has decided whether it applies here.

→ Model it as a vector and link the risks it would realize, or record why it does not apply.

Messaging abuse MA-03 · Snowshoe sending across number pools

warning

No attack vector in the register maps to this filter evasion technique, so nobody has decided whether it applies here.

→ Model it as a vector and link the risks it would realize, or record why it does not apply.

Messaging abuse MA-04 · Brand-impersonating sender IDs

warning

No attack vector in the register maps to this impersonation technique, so nobody has decided whether it applies here.

→ Model it as a vector and link the risks it would realize, or record why it does not apply.

Policy review overdue · 1

POL-02 · Security Monitoring & Detection Standard v1.4

warning

Review was due 2026-10-06. 2 control(s) enforce a policy nobody has re-approved.

→ Re-approve or revise, and re-baseline the controls it governs.