mbo-aerzte-pruefung

Pass

Audited by Gen Agent Trust Hub on Jun 28, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill ingests untrusted user input in the form of medical conduct descriptions, which presents a surface for indirect prompt injection.
  • Ingestion points: The "konkretes ärztliches Verhalten" (specific medical behavior) input field in SKILL.md.
  • Boundary markers: The skill does not implement specific delimiters or instructions to treat the ingested case data as untrusted content.
  • Capability inventory: The skill orchestrates researcher and drafting agents for legal analysis and documentation; it does not request high-privilege system access or unauthorized network operations.
  • Sanitization: There are no explicit validation or sanitization mechanisms defined for the user-provided data.
  • [SAFE]: The skill references official German regulatory websites and established legal research portals (e.g., gesetze-im-internet.de, bundesaerztekammer.de) which are well-known and trusted sources for the stated purpose of medical law compliance.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 28, 2026, 08:40 AM
Security Audit — agent-trust-hub — mbo-aerzte-pruefung