split-layout-technical
Warn
Audited by Gen Agent Trust Hub on Sep 21, 2026
Risk Level: MEDIUMOBFUSCATIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [OBFUSCATION]: The demo/index.html file uses a large Base64-encoded string (encodedHtml) to store the demo's HTML, CSS, and JavaScript. This hides the executable content from static analysis tools.
- [DYNAMIC_EXECUTION]: The skill decodes the Base64-encoded content at runtime and executes it within a sandboxed iframe via the srcdoc attribute.
- [INDIRECT_PROMPT_INJECTION]: The skill defines a workflow in demo/PROMPT.md that processes external 'Product brief' and 'Content anchors' data. This creates a surface for indirect prompt injection where malicious instructions could be embedded in the design metadata.
- Ingestion points: Metadata fields in demo/PROMPT.md.
- Boundary markers: None present.
- Capability inventory: High-fidelity HTML/JS generation.
- Sanitization: No sanitization logic detected.
- [EXTERNAL_DOWNLOADS]: The skill and its demo reference assets and runtime dependencies from well-known services such as Cloudflare, Google, and Supabase.
Audit Metadata