figma-logos-svgl
Pass
Audited by Gen Agent Trust Hub on Jun 25, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONREMOTE_CODE_EXECUTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: Retrieves brand logos and API metadata from the SVGL service (api.svgl.app and svgl.app), which is a well-known repository for SVG assets.
- [COMMAND_EXECUTION]: Executes curl commands via bash to interact with the SVGL API and download SVG files.
- [REMOTE_CODE_EXECUTION]: Incorporates externally fetched SVG data into dynamic JavaScript code blocks that are executed within the Figma environment via the use_figma tool.
- [PROMPT_INJECTION]: The skill exposes a surface for indirect prompt injection by ingesting untrusted SVG content from a remote API and interpolating it into a sensitive tool call.
- Ingestion points: External SVG content is fetched via curl from URLs like https://svgl.app/library/{filename}.svg.
- Boundary markers: None. The fetched SVG string is directly assigned to a variable in the generated script.
- Capability inventory: The skill uses the use_figma tool, which allows execution of JavaScript code within the Figma context.
- Sanitization: No comprehensive sanitization or validation of the SVG structure is performed before processing.
Audit Metadata