mjml-email-templates

Pass

Audited by Gen Agent Trust Hub on Oct 10, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data such as user names and organization names and interpolates them directly into MJML templates using a simple regex-based substitution method.
  • Ingestion points: The UserEmailComposer.cs section demonstrates accepting external strings like OrganizationName and InviteeName which are passed to the renderer.
  • Boundary markers: The skill uses {{Variable}} delimiters for substitution but does not implement or instruct the agent to use boundary markers or "ignore" instructions for the data contained within those variables.
  • Capability inventory: The skill reads embedded resources from the assembly and compiles MJML into HTML. The resulting HTML is intended to be sent via email or viewed in a browser via the preview endpoint.
  • Sanitization: The SubstituteVariables method in MjmlTemplateRenderer.cs performs raw string replacement. There is no evidence of HTML encoding or sanitization of the input variables before they are rendered into the final HTML output, which could allow malicious data to break the template structure or inject scripts.
  • [DYNAMIC_EXECUTION]: The skill includes an Email Preview Endpoint that renders MJML to HTML and serves it to the browser with the text/html content type. Because the variable substitution lacks sanitization, this endpoint is vulnerable to Cross-Site Scripting (XSS) if the data used to populate the preview is influenced by an attacker.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 10, 2026, 10:26 PM