anklageschrift-zerlegen
Pass
Audited by Gen Agent Trust Hub on Sep 3, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No malicious patterns detected. The skill instructions are focused entirely on legal analysis and document drafting using official German legal frameworks (StPO, AktO, RiStBV).
- [SAFE]: Data exposure and exfiltration check. The skill references legitimate, well-known legal databases (gesetze-im-internet.de, dejure.org, openJur) for verifying statutes and case law. No unauthorized network activity or credential exposure is present.
- [SAFE]: Prompt injection analysis. The skill contains specific defensive instructions for the AI, such as 'keine Modellwissen-Zitate' (no model-knowledge citations) and 'keine Literatur- oder Datenbankfundstellen erfinden' (do not invent citations), which are designed to prevent the model from generating plausible but false information. No instructions to bypass safety filters were found.
- [SAFE]: Dependency analysis. The skill does not include or download any external code, scripts, or packages.
Audit Metadata