betriebsuebergang
Pass
Audited by Gen Agent Trust Hub on Oct 24, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied transaction data, employee lists, and collective bargaining information to generate legal documents and assessments. This creates a potential surface for indirect prompt injection.
- Ingestion points: Input variables defined in the 'Eingaben' section of SKILL.md, including 'Adressatenliste' (recipient list) and 'Übergehende Betriebsmittel'.
- Boundary markers: The skill lacks explicit delimiters or instructions to the model to ignore potential commands embedded within user-supplied data during processing.
- Capability inventory: The skill's primary capabilities are limited to text generation, legal research guidance, and drafting; no file writing, network requests (beyond documentation links), or shell command executions are present in the skill's operational logic.
- Sanitization: No specific data validation or sanitization steps are defined for the incoming transaction details.
- [SAFE]: The skill references well-known and reputable German legal resources, including gesetze-im-internet.de, dejure.org, and the federal labor court database (bundesarbeitsgericht.de). These references are documentation-only and do not involve untrusted code execution.
- [SAFE]: The inclusion of a manual evaluation command in the test file (
python scripts/eval.py) is a standard developer practice for assessing skill accuracy and does not constitute a remote code execution vulnerability in the context of the skill's usage.
Audit Metadata