mieterhoehung-558-bgb
Pass
Audited by Gen Agent Trust Hub on Sep 6, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes raw text from legal notices ("Verlangen im Wortlaut"), which constitutes a surface for indirect prompt injection. An attacker could embed instructions in a fake legal notice to manipulate the agent's output.
- Ingestion points: The skill accepts the full text of a rent increase request as input in SKILL.md.
- Boundary markers: None identified.
- Capability inventory: The skill utilizes a command-line calculator scripts.legal_calc.cli.
- Sanitization: No explicit sanitization or validation of the input text is described.
- [COMMAND_EXECUTION]: The skill demonstrates the use of local Python scripts (scripts.legal_calc.cli and scripts/eval.py) via shell commands for precise legal date calculations and testing. These are functional components of the skill but involve executing code in the host environment.
Audit Metadata