event-driven-analyzer

Pass

Audited by Gen Agent Trust Hub on Jun 16, 2026

Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
  • Local Script Execution: The skill incorporates a Python script, scripts/event_math.py, to handle deterministic financial calculations. This script uses standard libraries and processes data from local JSON files or standard input, which is a common and safe pattern for this type of analytical tool.
  • Data Source Discipline: The instructions emphasize a strict source hierarchy, prioritizing primary documents like SEC filings and court records over third-party commentary. This focus on authoritative data helps maintain the integrity of the generated reports.
  • Information Handling: The skill includes explicit instructions to flag potential Material Non-Public Information (MNPI) for compliance review. This proactive approach to data sensitivity aligns with industry best practices for financial analysis.
  • Workflow Configuration: A preflight script, user_context_preflight.py, is used to initialize the workflow. This process is documented as a standard setup step to ensure the agent has the necessary context before beginning analysis.
  • Indirect Data Ingestion: As the skill processes external financial data (e.g., filings and transcripts), it possesses a potential attack surface for indirect prompt injection. However, the presence of comprehensive quality checks and a requirement for primary-source verification effectively manages this consideration.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 16, 2026, 06:12 AM
Security Audit — agent-trust-hub — event-driven-analyzer