nested-container-frames

Pass

Audited by Gen Agent Trust Hub on Aug 26, 2026

Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTION
Full Analysis
  • [OBFUSCATION]: The file demo/index.html uses Base64 encoding to bundle the primary HTML payload for the design demo. The encoded string is decoded at runtime and loaded into a sandboxed iframe using srcdoc.
  • [INDIRECT_PROMPT_INJECTION]: The demo prompt in demo/PROMPT.md instructs the agent to process external inputs such as product briefs and reference URLs, creating a surface for potential indirect prompt injection attacks.
  • Ingestion points: demo/PROMPT.md (Product brief and Fidelity target sections).
  • Boundary markers: None identified in the provided prompts.
  • Capability inventory: The skill is capable of generating complex HTML, CSS, and JavaScript code.
  • Sanitization: No explicit instructions for sanitizing or escaping the ingested content are present.
  • [EXTERNAL_DOWNLOADS]: The demo configuration in demo/source.json and demo/index.html references external runtime dependencies from well-known services, including Tailwind CSS and the Iconify icon library, which are used to render the design reference.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 26, 2026, 01:38 PM
Security Audit — agent-trust-hub — nested-container-frames