regulatory-affairs-head

Pass

Audited by Gen Agent Trust Hub on Sep 14, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill includes specialized Python utilities, scripts/regulatory_pathway_analyzer.py and scripts/regulatory_tracker.py, designed to be executed by the agent or user to manage regulatory data. These scripts utilize standard libraries for data processing and maintain local JSON records; they do not exhibit any behavior suggestive of malicious command execution, such as spawning unauthorized shells or accessing sensitive system paths.
  • [INDIRECT_PROMPT_INJECTION]: The skill operates on medical device profiles and submission data, which are processed by the included scripts.
  • Ingestion points: Data enters the system via command-line arguments and local JSON input files (regulatory_submissions.json).
  • Boundary markers: The scripts employ structured data parsing and strict validation against predefined Enum classes, providing inherent boundaries between external data and code logic.
  • Capability inventory: The skill is limited to standard console output and local file reading/writing; no network or arbitrary execution capabilities are present.
  • Sanitization: Inputs are verified against specific data models and type-safe enums, ensuring that data processed by the agent remains within defined logical constraints.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 14, 2026, 06:47 PM
Security Audit — agent-trust-hub — regulatory-affairs-head