infographic-design
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute a local Python utility (
scripts/check.py) to perform preflight checks on generated SVG files. This process is used to ensure the output is technically sound (e.g., checking for XML parse errors, broken references, and CSS variable consistency) before delivery to the user. - [INDIRECT_PROMPT_INJECTION]: The skill includes instructions to "review existing graphics" and provide modification suggestions. This capability involves processing user-supplied SVG files, which technically constitutes an indirect prompt injection surface. The risk is minimized by the skill's specific focus on design analysis and the lack of high-privilege tool access that could be abused by instructions embedded in a design file.
- [DYNAMIC_EXECUTION]: The provided validation scripts utilize the standard Python
xml.etree.ElementTreemodule to parse and analyze SVG content. While XML parsing of untrusted data has inherent risks, the usage here is focused on structural analysis of design elements (shapes, text, and styles) rather than executing embedded code or accessing external entities.
Audit Metadata