Skills
Navigate
skills/danielkerridge/claude-code-power-platform-skills/dataverse-web-resources/Gen Agent Trust Hub

dataverse-web-resources

Pass

Audited by Gen Agent Trust Hub on Oct 31, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill documents patterns for ingesting Dataverse record data and rendering it in custom HTML interfaces. This creates a potential attack surface where malicious content stored in Dataverse fields could attempt to execute scripts or influence the agent's context.
  • Ingestion points: Data is fetched from Dataverse tables via Xrm.WebApi calls in resources/html-dashboards.md and resources/js-form-scripts.md.
  • Boundary markers: No specific delimiters or "ignore embedded instructions" warnings are used when processing record data retrieved from the API.
  • Capability inventory: The documented scripts have access to the full Xrm client API, which includes Dataverse record operations (Xrm.WebApi), navigation control (Xrm.Navigation), and side pane management (Xrm.App.sidePanes).
  • Sanitization: While most examples demonstrate safe practices like using textContent for UI updates, some template patterns in resources/ribbon-command-bar.md use innerHTML. This represents a potential Cross-Site Scripting (XSS) risk if the source data is attacker-controlled.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 31, 2026, 05:50 PM