anhoerungsprotokoll-und-korrektur
Pass
Audited by Gen Agent Trust Hub on Sep 3, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes hearing protocols provided by the user as untrusted data. Evidence chain: 1. Ingestion points: Review of hearing protocols (SKILL.md). 2. Boundary markers: Not specified. 3. Capability inventory: Verification of legal norms via official databases and generation of text outputs. 4. Sanitization: Not specified. This represents a standard document analysis surface.
- [EXTERNAL_DOWNLOADS]: The agent is instructed to verify legal information using official or well-known services such as gesetze-im-internet.de and dejure.org. These are reputable sources within the legal domain and do not contribute to a negative verdict.
Audit Metadata