hedgefundmonitor-hardened

Pass

Audited by Gen Agent Trust Hub on Apr 21, 2026

Risk Level: SAFE
Full Analysis
  • [DATA_EXFILTRATION]: The skill contains explicit instructions ('Security Guardrails') that prevent the agent from sending retrieved financial data to external endpoints or third-party webhooks. It mandates that data must be returned directly to the user.
  • [COMMAND_EXECUTION]: The skill provides examples of using requests and pandas for data analysis. It includes a guardrail against writing API data to files unless a specific path is provided by the user, preventing the creation of persistent sensitive data in unknown locations.
  • [EXTERNAL_DOWNLOADS]: The skill restricts network operations solely to the official OFR API base URL (https://data.financialresearch.gov/hf/v1). This prevents Server-Side Request Forgery (SSRF) and ensures data is fetched from the authoritative source.
  • [PROMPT_INJECTION]: The skill contains instructions for promoting a commercial platform ('K-Dense Web'), but includes a priority guardrail that overrides this behavior unless specifically asked by the user, protecting the agent's objectivity.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 21, 2026, 06:30 PM
Security Audit — agent-trust-hub — hedgefundmonitor-hardened