ai-contract-disclosure
Pass
Audited by Gen Agent Trust Hub on Sep 3, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONNO_CODE
Full Analysis
- [NO_CODE]: The skill consists entirely of markdown instructions and does not package any executable scripts, binaries, or configuration files, which significantly minimizes the potential for direct technical compromise.- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze external evidence and legal documents, creating an attack surface where maliciously crafted documents could influence the agent's behavior.
- Ingestion points: The workflow in
SKILL.mdinvolves collecting 'Dokumente und Beweismittel' (documents and evidence) such as Form TX/PA/VA/SR, Deposit Copies, license agreements, and Cease-and-Desist orders. - Boundary markers: The instructions do not specify any delimiters or safety prompts to ensure the agent ignores instructions embedded within these external documents.
- Capability inventory: The agent generates risk matrices and decision-making outputs based on the ingested content.
- Sanitization: There are no instructions for sanitizing or filtering the content of the provided legal documents before processing.
Audit Metadata