extract-design-system
Warn
Audited by Gen Agent Trust Hub on Oct 22, 2026
Risk Level: MEDIUMEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill is a stub that initiates the installation of a full skill and CLI from a third-party GitHub repository (
arvindrk/extract-design-system) usingnpx skills add. It also requires downloading the Chromium browser vianpx playwright install chromiumat runtime. - [COMMAND_EXECUTION]: The skill uses
npxto install and run theextract-design-systemCLI, which executes shell commands and performs headless browser operations. - [INDIRECT_PROMPT_INJECTION]: The skill processes arbitrary data from public websites to generate design token files. This provides an attack surface where a malicious website could embed instructions designed to influence the agent's behavior.
- Ingestion points: Public website URLs (HTML/CSS content) via
SKILL.mdinstructions. - Boundary markers: None identified in the provided files to separate untrusted web content from agent instructions.
- Capability inventory: Shell command execution (
npx), network crawling, and file system writes (tokens.json,tokens.css,.extract-design-system/). - Sanitization: No evidence of validation or sanitization of the extracted web data before it is used to generate project files.
Audit Metadata