skills/klotzkette/claude-fuer-deutsches-recht/abmahnung-patentverletzung-verteidigung/Gen Agent Trust Hub
abmahnung-patentverletzung-verteidigung
Pass
Audited by Gen Agent Trust Hub on Sep 3, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill defines a process for analyzing external legal documents (patents, claims, court documents) which serves as an ingestion point for untrusted content.
- Ingestion points: Identified in the 'Arbeitsweg' and 'Sofortscan' sections where the agent collects patents, infringement letters, and court filings.
- Boundary markers: None specified in the instructions to separate untrusted data from instructions.
- Capability inventory: None. The skill does not use shell commands, network-writing tools, or dynamic code execution, mitigating the risk of exploitation.
- Sanitization: None specified.
- [SAFE]: No malicious patterns, obfuscation, or unauthorized data access were detected. The skill references well-known legal information portals and official registries for research purposes.
Audit Metadata