dataverse-metadata
Pass
Audited by Gen Agent Trust Hub on Oct 27, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes Dataverse metadata and solution XML files which could theoretically contain malicious instructions if the source environment is compromised. However, the risk is inherent to the metadata management purpose of the skill.
- Ingestion points: Dataverse environment schema (EntityDefinitions), local XML configuration files (e.g.,
Other/Solution.xml), and user-supplied names for publishers and solutions. - Boundary markers: The instructions do not define specific delimiters to separate data from instructions during metadata processing.
- Capability inventory: The skill utilizes a Dataverse SDK and Web API for creating and modifying schema, the
pacCLI for solution management, and is instructed to perform filesystem commits of generated XML. - Sanitization: No explicit sanitization of external metadata values is described before processing or interpolation into prompts.
Audit Metadata