vereinsrechtliche-sanktion
Pass
Audited by Gen Agent Trust Hub on Jun 28, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION]: The skill does not contain instructions aimed at bypassing safety guardrails or overriding agent behavior. While it processes untrusted user data (bylaws and case facts), its capabilities are limited to legal analysis and document drafting within the agent context.
- [EXTERNAL_DOWNLOADS]: The skill references trusted official repositories for legal research, including gesetze-im-internet.de, cas.tas-cas.org, and hudoc.echr.coe.int. These are well-known and reputable sources for law and jurisprudence.
- [COMMAND_EXECUTION]: The test configuration includes a call to a local evaluation script (eval.py). This is a standard developmental practice for skill validation and does not involve remote code execution or suspicious shell commands.
Audit Metadata