skills/mtent/speshul-skills/schema/Gen Agent Trust Hub

schema

Pass

Audited by Gen Agent Trust Hub on Aug 28, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill involves a vulnerability surface where the agent is instructed to "Inspect the rendered page" to extract information. This untrusted data ingestion point could potentially be used for indirect prompt injection if a target webpage contains malicious instructions. However, the skill provides mitigating instructions to verify all data against visible content and authoritative sources, and to perform rigorous syntax and semantic validation.
  • Ingestion points: SKILL.md (inspecting rendered pages)
  • Boundary markers: Not explicitly defined for the external page content ingestion.
  • Capability inventory: The skill provides guidelines for structured data generation; no dangerous shell, network, or file-writing tools are introduced by the skill scripts.
  • Sanitization: Instructions include mandatory validation of JSON syntax, type semantics, and search-feature eligibility.
  • [SAFE]: The skill and its reference documents consist solely of instructional text and documentation. No scripts, executables, or network-enabled code were found. External URLs in examples point to well-known placeholder domains (example.com, example.test) and official documentation sources (schema.org).
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 28, 2026, 11:56 AM
Security Audit — agent-trust-hub — schema