pauschalreise-maengel
Pass
Audited by Gen Agent Trust Hub on Sep 6, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes a local Python module to calculate legal deadlines and compensation sums based on contract dates. Evidence: The command python -m scripts.legal_calc.cli is used within SKILL.md.
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user inputs which is a potential injection vector. 1. Ingestion points: Input fields for travel defect descriptions and service descriptions in SKILL.md. 2. Boundary markers: The skill employs a structured output schema and the Reviewer agent tags unconfirmed data. 3. Capability inventory: The skill utilizes subprocess calls to internal Python scripts as seen in SKILL.md. 4. Sanitization: The multi-agent workflow includes a Reviewer agent responsible for marking unverified information.
Audit Metadata