concept-diagrams
Pass
Audited by Gen Agent Trust Hub on Sep 5, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-provided descriptions to generate SVG content which is then saved to the local filesystem. This creates an attack surface where maliciously crafted user input could potentially influence the resulting HTML/SVG files.
- Ingestion points: User descriptions for diagrams processed in the main workflow (SKILL.md).
- Boundary markers: The workflow relies on a predefined template (templates/template.html) to wrap the SVG content.
- Capability inventory: The skill uses write_file to save output and shell commands (python3, open) to preview diagrams (SKILL.md).
- Sanitization: The instructions do not specify sanitization or escaping of user-provided strings for SVG labels or titles.
- [DYNAMIC_EXECUTION]: The skill provides a mechanism to start a local web server by executing a Python one-liner via the shell.
- Evidence: The SKILL.md file contains a command pattern "python3 -c 'import http.server...'" to serve files from a local directory for previewing purposes.
Audit Metadata