google-structured-data
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill provides examples for generating structured data from page content or API responses, which constitutes an ingestion surface for untrusted data if the source content is attacker-controlled.
- Ingestion points: Data extraction from the DOM (e.g.,
document.querySelector) and external APIs (e.g.,fetch('https://api.example.com/recipes/123')) as described inSKILL.mdandreferences/javascript-generation.md. - Boundary markers: Absent in the provided code snippets.
- Capability inventory: Injects metadata into the document head via
<script type="application/ld+json">elements. - Sanitization: The examples use
JSON.stringifyto maintain structural integrity, but do not demonstrate specific sanitization of string values against prompt injection or XSS (though JSON-LD script tags generally do not execute as HTML). - [DYNAMIC_EXECUTION]: The skill documents routine web development patterns for runtime generation and injection of JSON-LD scripts, including manual DOM manipulation (
document.createElement('script')) and framework-specific implementations like React'sdangerouslySetInnerHTML. These patterns are consistent with the skill's stated purpose of facilitating dynamic SEO metadata generation.
Audit Metadata