urheberstrafrecht
Pass
Audited by Gen Agent Trust Hub on Sep 30, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: No security risks or malicious patterns were identified. The skill promotes security best practices by requiring live verification of legal norms and case law via official and well-known governmental domains (e.g., gesetze-im-internet.de, dejure.org, and eur-lex.europa.eu).
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process external evidence provided by users, such as platform posts and document logs, which constitutes a potential surface for instructions hidden within that data.
- Ingestion points: Evidence files, platform posts, and promptlogs mentioned in the Sofortstart and Document collection sections.
- Boundary markers: The skill includes instructions to prioritize live verification against official sources over user-provided or model-internal data, which mitigates the risk of acting on malicious instructions in evidence.
- Capability inventory: Textual generation of legal memos, risk matrices, and defense notes.
- Sanitization: Not applicable as the skill performs natural language analysis rather than command execution.
Audit Metadata