fix(lightweightRiskModel): fix duplicate quality check and add null g…

[Div1912]

Summary

Two bugs fixed in src/lib/lightweightRiskModel.js:

Bug 1 - Duplicate totalPayments check inflates quality score

getDataQualityScore() awarded 25 points for totalPayments > 10 and another 25 points for totalPayments > 0. The second condition is always satisfied when the first is, so a wallet with >=10 payments can never score below 50/100 regardless of other data quality indicators (counterparties, asset diversity). The intent was to reward having any transactional activity alongside counterparty diversity. The last check has been changed to totalPayments > 0 && uniqueCounterparties > 0 to reflect that intent.

Bug 2 - No null guard in calculateRiskScore()

If metrics was null or undefined, the call would propagate into normalizeFeatures() and silently catch an error, falling through to fallbackRiskCalculation({}) with no logged context. An explicit null check is now added at the top of calculateRiskScore() so callers get a clear error message.

Files Changed

• src/lib/lightweightRiskModel.js

Testing

• getDataQualityScore({ totalPayments: 15, uniqueCounterparties: 0, assetDiversity: 0 }) should now return score: 25 (not 50).
• • calculateRiskScore(null) should log a clear error and return the fallback result.

[vercel]

Someone is attempting to deploy a commit to the mericcintosun Team on Vercel.

A member of the Team first needs to authorize it.