skills/klotzkette/claude-fuer-deutsches-recht/ki-arbeitsauftrag-mahnwesen-honorar/Gen Agent Trust Hub
ki-arbeitsauftrag-mahnwesen-honorar
Pass
Audited by Gen Agent Trust Hub on Sep 29, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process external documents, facts, and evidence, which represents a potential indirect prompt injection surface.
- Ingestion points: Untrusted data enters the context through the "Material erfassen" (capture material) and "Dokumente und Beweismittel sammeln" (collect documents and evidence) workflow steps.
- Boundary markers: The skill provides structural instructions to "sichtbar zwischen Tatsache, Behauptung und Bewertung trennen" (clearly separate facts, claims, and evaluations), which acts as a logical boundary for the agent.
- Capability inventory: The skill is strictly instructional and does not utilize high-risk tools such as network requests, arbitrary shell commands, or persistent file writes.
- Sanitization: The skill explicitly includes a "Halluzinations-Sperre" (hallucination block) that mandates the verification of every legal statement against actual statutes and case law before delivery.
Audit Metadata