data-model-creation
Pass
Audited by Gen Agent Trust Hub on Oct 3, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user input, such as business entities and descriptions, to generate Mermaid diagrams and call data modeling tools, creating an attack surface for indirect instructions.
- Ingestion points: User-supplied business descriptions, entity names, and relationship cardinality (SKILL.md).
- Boundary markers: Absent. The instructions do not specify the use of delimiters or provide "ignore embedded instructions" warnings when processing user data.
- Capability inventory: The skill uses
manageDataModel(actions: list, get, docs) andmodifyDataModel(action: create) to interact with the database modeling layer (SKILL.md). - Sanitization: No explicit sanitization, escaping, or validation steps for user-provided text are mentioned prior to the generation of schema identifiers or Mermaid content.
Audit Metadata