automated-reconciliation
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The analyzed skill implements a standard financial reconciliation engine. It utilizes internal deterministic rules, rounding tolerances, and subset-sum algorithms (
scripts/calculate.py) to process transaction data locally inside the environment. - [SAFE]: Data ingestion is limited to local file system targets specified within test suites (
evals/files/), posing no exfiltration or external remote code risks. - [SAFE]: No obfuscation, prompt manipulation vulnerabilities, dynamic code injection surfaces, or privilege escalation vectors were observed in the instructions or scripts.
Audit Metadata