architecture-diagram

Pass

Audited by Gen Agent Trust Hub on May 9, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill's template loads html2canvas and jsPDF from a well-known content delivery network (jsDelivr). These are legitimate, widely-used libraries for browser-based rendering and PDF generation. The implementation correctly includes Subresource Integrity (SRI) hashes to ensure that only the verified, untampered versions of these scripts are executed in the browser.
  • [PROMPT_INJECTION]: The skill provides a surface for indirect prompt injection by interpolating user-provided architectural requirements into an HTML template. 1. Ingestion points: User architectural descriptions provided during interaction. 2. Boundary markers: No explicit instructions or delimiters are used to wrap the interpolated user data. 3. Capability inventory: The skill generates HTML files that include SVG and JavaScript for user-initiated export features. 4. Sanitization: No explicit sanitization or escaping of user input is specified for the text labels rendered within the diagram. This represents a standard surface for static report generation tools.
  • [DATA_EXFILTRATION]: No unauthorized exfiltration or exposure of sensitive data was identified. Browser APIs such as navigator.clipboard and URL.createObjectURL are used appropriately for the 'Copy' and 'Download' features triggered by user interaction.
Audit Metadata
Risk Level
SAFE
Analyzed
May 9, 2026, 07:42 PM