design-taste-frontend
Fail
Audited by Gen Agent Trust Hub on Nov 2, 2026
Risk Level: CRITICALEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: Automated security scanners have positively identified malicious components within the skill's resource library.
- The file 'library/sites.json' is flagged with a confirmed malware reputation detection ('FileRepMalware').
- Reference URLs provided in the skill's study material, including 'https://thedeepend.space' and 'https://extension.textmode.art', are identified as blacklisted by threat intelligence services.
- [COMMAND_EXECUTION]: The skill performs automated control of local software to execute measurement and auditing tasks.
- Spawns headless browser instances (Chrome, Edge, or Brave) using 'child_process.spawn' in 'scripts/lib/cdp.mjs'. While this is central to the skill's function, it represents a high-privilege automated behavior.
- [EXTERNAL_DOWNLOADS]: The skill performs numerous network operations to fetch external content and recommends the use of third-party scripts.
- Fetches image and video data from 'api.openverse.org', 'commons.wikimedia.org', 'api.artic.edu', 'collectionapi.metmuseum.org', 'archive.org', and 'images-api.nasa.gov'.
- Instructs the agent to include third-party libraries (GSAP, Lenis) from 'cdn.jsdelivr.net' in the generated frontend code, creating a dependency on external CDNs.
- [INDIRECT_PROMPT_INJECTION]: The core auditing functionality creates a vulnerability surface for indirect prompt injection.
- Ingestion points: 'scripts/lib/probe.mjs' extracts raw text content, including headings and body text, from arbitrary external websites or local files.
- Capability inventory: The AI agent using this skill has high-privilege capabilities, including command execution (browser automation), local file writing, and network access.
- Boundary markers: Absent. Extracted text is provided to the agent for critique without protective delimiters or instructions to ignore embedded commands.
- Sanitization: Absent. The probe script does not filter or validate the content retrieved from target pages before processing.
Recommendations
- CRITICAL: 1 file(s) identified as malware by FileRep - DO NOT USE
- AI detected serious security threats
- Contains 2 malicious URL(s) - DO NOT USE
Audit Metadata