data-act-nutzerdatenzugang
Pass
Audited by Gen Agent Trust Hub on Sep 6, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied text ("Wortlaut des Zugangsverlangens") that is analyzed by agents to perform legal drafting. The absence of strict boundary markers or sanitization logic creates a surface where an adversary could provide a legal request containing instructions designed to bias the AI's legal analysis or override its internal logic.
- Ingestion points: User-provided legal request text defined as input in
SKILL.md. - Boundary markers: The instructions do not specify the use of delimiters or 'ignore' instructions for the ingested text.
- Capability inventory: The skill utilizes a local Python script for calculations and accesses legal references.
- Sanitization: No sanitization or validation of the user-supplied text is described.
- [COMMAND_EXECUTION]: The skill documentation includes instructions for running a local Python module (
scripts.legal_calc.cli) to perform automated date calculations based on input parameters. - [EXTERNAL_DOWNLOADS]: The skill provides links to official and well-known legal repositories, including EUR-Lex and the German federal justice portal (gesetze-im-internet.de), to support its legal research functions.
Audit Metadata