akteneinsicht-vollstaendig
Pass
Audited by Gen Agent Trust Hub on Sep 3, 2026
Risk Level: SAFENO_CODEINDIRECT_PROMPT_INJECTION
Full Analysis
- [NO_CODE]: The skill consists entirely of markdown instructions (SKILL.md) and lacks any executable scripts, binary files, or package management configuration. No automated logic is performed outside of standard language model reasoning.- [INDIRECT_PROMPT_INJECTION]: 1. Ingestion points: The skill is designed to process external legal documents, case facts, and administrative decisions (e.g., Bescheide, Protokolle) mentioned in the 'Arbeitsweg' and 'Kaltstartfragen' sections. 2. Boundary markers: While the skill uses specific formatting instructions, it does not define strict delimiters or 'ignore instructions' warnings for untrusted case data provided by users. 3. Capability inventory: The skill has no associated scripts or tools, meaning it cannot execute shell commands, perform file system writes, or make network calls. 4. Sanitization: There are no explicit sanitization or filtering instructions for the ingested text, though the risk is minimal given the absence of executable capabilities.- [SAFE]: All identified legal resources (gesetze-im-internet.de, dejure.org) are well-known, trusted official or public repositories for German legal texts, representing standard research functionality for this skill's domain.
Audit Metadata