build-credit-scorecard
Pass
Audited by Gen Agent Trust Hub on Oct 31, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill establishes domain-specific logic for financial risk modeling (WOE binning, IV filtering, and PDO scaling) without any detected malicious patterns or attempts to bypass safety protocols.
- [SAFE]: Explicit security guardrails are included (ASI02 and ASI07) which specifically instruct the agent to sanitize SQL queries and mask sensitive PII (Social Security Numbers), demonstrating proactive safety measures.
- [SAFE]: No obfuscation, remote code execution patterns, or unauthorized network operations were identified in the instructions or metadata.
- [SAFE]: The skill uses a set of standard data analysis tools (
read_file,query_db, etc.) that are consistent with its stated purpose of building financial scorecards.
Audit Metadata