architecture-diagram
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted user descriptions of system architectures and interpolates them into a generated HTML/SVG document saved to the local filesystem. The lack of defined sanitization protocols creates a vulnerability surface where malicious input could be interpreted as code if the file is opened.\n
- Ingestion points: Architecture components, connections, and technology descriptions provided by the user (SKILL.md).\n
- Boundary markers: None; the instructions do not specify delimiters or markers to isolate user input from the HTML structure.\n
- Capability inventory: The skill utilizes the
write_filetool to save content to the user's local directory.\n - Sanitization: No evidence of input validation, filtering, or HTML escaping is present in the instructions or template.\n- [EXTERNAL_DOWNLOADS]: The generated HTML template references and fetches the JetBrains Mono font family from Google's official font delivery network (fonts.googleapis.com).
Audit Metadata