visualizations
Pass
Audited by Gen Agent Trust Hub on Sep 19, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill invokes a local Node.js script (
scripts/visualizations/generate-visual.js) to handle the image generation process. This is the primary mechanism of the skill. - [INDIRECT_PROMPT_INJECTION]: The skill accepts user-provided concepts through the
$ARGUMENTSvariable and incorporates them into a structured prompt for an external AI model. - Ingestion points: User input is gathered in Step 1 and used in Step 4 of the prompt construction.
- Boundary markers: The instructions include a 'Spelling protection' phase (Step 3) to minimize and shorten text, which provides a natural constraint, though no formal technical delimiters (like XML tags) are applied to the user-supplied concept.
- Capability inventory: The skill performs file system writes to save the generated PNG in
projects/visualizations/and executes a subprocess vianode. - Sanitization: There are no explicit sanitization steps mentioned to filter out control characters or injection attempts from the user-provided concept before it is interpolated into the CLI command.
Audit Metadata