credit-collateral-risk-mgmt

Pass

Audited by Gen Agent Trust Hub on Jun 29, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill is highly structured and focuses on business-specific regulatory compliance for bank collateral management.
  • [COMMAND_EXECUTION]: The skill invokes a local script scripts/validate_collateral.py to verify data input schema. Analysis of this script confirms it only uses standard Python libraries (re, sys, datetime) and contains no dangerous system calls or external network operations.
  • [DATA_EXFILTRATION]: While the skill references external data sources (e.g., official government registries like the 'Zhong Deng Wang'), these are well-known official services used for verification purposes. There are no patterns of sending sensitive user data to unauthorized or untrusted domains.
  • [PROMPT_INJECTION]: The instructions are strictly limited to risk assessment logic. No attempts to override system behavior or bypass safety filters were found. The skill includes 'step-gated' mechanisms to ensure the agent follows mandatory compliance checks.
  • [EXTERNAL_DOWNLOADS]: No external dependencies are listed in the requirements file, and no remote code execution patterns (like curl | bash) are present. All referenced files are local to the skill repository.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 29, 2026, 06:40 AM
Security Audit — agent-trust-hub — credit-collateral-risk-mgmt