number-details

Pass

Audited by Gen Agent Trust Hub on Aug 16, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [OBFUSCATION]: The demo loader in demo/index.html contains a large Base64-encoded string (encodedHtml) which decodes to the full HTML structure and initialization logic for the design demo.
  • [DYNAMIC_EXECUTION]: The demo/index.html script decodes bundled HTML at runtime and injects it into a sandboxed iframe using the srcdoc attribute. It also dynamically resolves and injects paths for runtime dependencies.
  • [EXTERNAL_DOWNLOADS]: The skill fetches several runtime scripts and fonts from well-known and trusted services, including Google Fonts, Cloudflare (cdnjs), and the Iconify CDN. These are documented in demo/source.json as verified dependencies.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted 'Product brief' and 'Reference snapshot' data from demo/PROMPT.md to generate code. 1. Ingestion points: demo/PROMPT.md. 2. Boundary markers: Absent. 3. Capability inventory: Full HTML/JavaScript code generation. 4. Sanitization: Absent.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 16, 2026, 07:33 AM
Security Audit — agent-trust-hub — number-details