lookup-dataverse-table

Pass

Audited by Gen Agent Trust Hub on Aug 30, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [Indirect Prompt Injection Surface]: The skill retrieves and displays records from Microsoft Dataverse tables, which is a standard method for data ingestion.
  • Ingestion points: Data from external tables is brought into the agent's context via the read_query tool in SKILL.md.
  • Capability inventory: The agent can query Dataverse and manage user-specific field preferences using memory tools.
  • Boundary markers: The instructions do not explicitly require wrapping retrieved data in delimiters (such as triple backticks or specific tags) to differentiate record content from system instructions.
  • Sanitization: The skill prioritizes user privacy by hiding internal GUIDs and identifiers, though it does not specify sanitization or escaping for the textual content of the records themselves.
  • Context: When agents process content from external or shared databases, it is a standard practice to ensure the data is handled strictly as information to be displayed rather than instructions to be followed.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 30, 2026, 11:56 PM
Security Audit — agent-trust-hub — lookup-dataverse-table