software-policy-legal-engineering
Pass
Audited by Gen Agent Trust Hub on Sep 30, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONNO_CODE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill establishes a workflow for ingesting and analyzing untrusted external legal documents, which represents an indirect prompt injection surface.\n
- Ingestion points:
SKILL.mdinstructs the agent to collect and analyze documents such as contracts (Vertragsurkunden), expert opinions (Sachverständigengutachten), and administrative acts (Verwaltungsakte).\n - Boundary markers: Absent; the skill does not specify the use of delimiters or specific instructions to ignore embedded prompts within these documents.\n
- Capability inventory: None; the skill does not include any executable code, scripts, or tool definitions.\n
- Sanitization: Absent; the instructions do not provide methods for sanitizing or validating the contents of the ingested documents.\n- [NO_CODE]: The skill is composed entirely of natural language instructions and markdown formatting. It contains no scripts (Python, JavaScript, shell), binaries, or automated configuration files.
Audit Metadata