schema
Pass
Audited by Gen Agent Trust Hub on Sep 10, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests user-provided content to generate structured JSON-LD markup, which presents a surface for indirect prompt injection. \n
- Ingestion points: The skill processes user-supplied descriptions of products, articles, and FAQs to populate schema templates as defined in SKILL.md and evals/evals.json. \n
- Boundary markers: No specific boundary markers or 'ignore' instructions for embedded data are specified to prevent the agent from following instructions hidden within user-supplied content. \n
- Capability inventory: The agent's primary capability in this skill is generating JSON-LD script blocks intended for implementation in web environments, as documented in SKILL.md and references/schema-examples.md. \n
- Sanitization: The instructions do not include explicit requirements for the agent to sanitize or escape user-provided strings (such as escaping script tags) when interpolating them into code outputs. \n- [SAFE]: The skill references established industry tools and official documentation for its primary functions. \n
- Directs users to the Google Rich Results Test (search.google.com) for validation. \n
- References official Schema.org documentation and their validator (validator.schema.org). \n
- Correctly utilizes standard https://schema.org contexts for all structured data examples.
Audit Metadata