anerkennung-ects-und-ausland

Pass

Audited by Gen Agent Trust Hub on Sep 3, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to process external, untrusted academic and administrative documents including study regulations, university notices, and correspondence. \n
  • Ingestion points: Documents such as Studien-/Prüfungsordnung, Bescheid, and E-Mail are ingested as part of the legal assessment workflow. \n
  • Boundary markers: The instructions do not define specific delimiters or instructions to ignore embedded commands within the provided documents. \n
  • Capability inventory: The agent generates sensitive legal outputs such as appeals (Widerspruch), action plans, and internal memos, which could be influenced by adversarial content in the input. \n
  • Sanitization: The workflow does not specify methods for sanitizing or validating the text content of the provided documents.\n- [EXTERNAL_DOWNLOADS]: The skill directs the agent to verify legal information using external databases such as gesetze-im-internet.de, dejure.org, and openJur. These are well-known and reputable legal information services in the German jurisdiction and are used appropriately for evidence verification.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 3, 2026, 08:35 AM