design-agentic-infrastructure

Pass

Audited by Gen Agent Trust Hub on Aug 7, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADS
Full Analysis
  • [SAFE]: The skill functions as a consultant, using provided Markdown files for logic and decision-making.
  • [EXTERNAL_DOWNLOADS]: The skill performs live web searches to obtain current market pricing and tool options. This is restricted to distilled results via a subagent to maintain security.
  • [INDIRECT_PROMPT_INJECTION]: The skill has an ingestion surface for user requirements and external search results.
  • Ingestion points: User-provided brief (Step 1) and search results (Step 5).
  • Boundary markers: The design template uses clear section headers, though explicit prompt delimiters for user input are not detailed.
  • Capability inventory: Limited to live web search via a subagent.
  • Sanitization: Instructions explicitly require subagents to distill information and avoid raw page content to prevent prompt injection.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 7, 2026, 04:08 AM
Security Audit — agent-trust-hub — design-agentic-infrastructure