scandinavian-design
Pass
Audited by Gen Agent Trust Hub on Sep 13, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill includes a
scripts/directory containing Node.js utilities (run-eval.js,snap.js,recon.js, etc.) that automate a Chromium browser using the Playwright library. These tools are used to capture screenshots, inspect DOM elements, and verify design changes on target websites. - [EXTERNAL_DOWNLOADS]: Verification scripts fetch the 'Inter' font family from Google's official font service (
fonts.googleapis.com) to serve as a consistent typography baseline for design previews. - [DYNAMIC_EXECUTION]: The browser automation scripts utilize
page.evaluate()to run JavaScript within the context of the target web pages. This is used to calculate visual metrics such as color contrast, element spacing, and content density. - [INDIRECT_PROMPT_INJECTION]: The skill's auditing tools ingest data from external websites, including HTML structure and text content. While this processes untrusted data, the risk is localized to the design auditing environment.
- Ingestion points: DOM and style extraction logic in
scripts/recon.js,scripts/probe.js,scripts/density.js, andscripts/tints.js. - Boundary markers: None identified for external content ingestion.
- Capability inventory: Browser automation, taking screenshots, and writing results to the local filesystem (
demos/directory). - Sanitization: No specific sanitization of ingested DOM text is performed before it is passed back to the agent environment.
Audit Metadata