What it does

This skill transforms your AI agent into a Senior Commercial Credit Underwriter. It synthesizes complex alternative data into structured underwriting briefings for small-and-medium businesses (SMBs). By correlating five distinct data streams—bank cash flow, accounting records, credit bureaus, utility/payroll data, and public records—it provides a 360-degree view of borrower health.

Why use this skill

Standard prompting often leads to surface-level financial analysis or "hallucinated" credit worthiness. This skill implements a strict weighting protocol (35% bank, 25% accounting, 20% bureau, etc.) and a mandatory contradiction detection engine. It doesn't just read data; it looks for "lies" between data sources—like when a QuickBooks file shows growth that isn't reflected in the bank deposits. It's an essential tool for FinTech developers and lending teams who need consistent, auditable risk assessments before a human making the final call.

Supported tools

The skill is architected to work with data from major financial aggregators including Plaid, Finicity, QuickBooks, Xero, Experian, and Equifax. It produces standardized JSON and narrative outputs compatible with modern lending LOS (Loan Origination Systems).

What the output looks like

You receive a multi-layered briefing containing a weighted risk score, a per-source confidence matrix, a flagged list of data contradictions, and a specific advisory recommendation (e.g., Approve with Conditions). Every briefing includes a self-critique section highlighting potential blind spots like seasonality or thin files.