salgadoia-web-design-engineer-pro
Pass
Audited by Gen Agent Trust Hub on Sep 19, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill has a significant attack surface for indirect prompt injection during the website redesign process. • Ingestion points: The skill uses
scripts/scrape-cascade.mjsto fetch content from user-provided URLs which is then stored in.brief/scrape.json. • Boundary markers: There are no specific delimiters or instructions to ignore embedded commands within the scraped data before it is analyzed by the agent. • Capability inventory: The agent utilizesBash,Write, andEdittools to perform complex project creation and file modifications. • Sanitization: Scraped content is directly transformed into JSON and markdown reports without explicit sanitization steps mentioned in the instructions. - [COMMAND_EXECUTION]: The skill relies on extensive use of shell commands for project setup and management. • Evidence: The skill automates project creation using
npx create-next-app, installs various dependencies vianpm install, and executes local Node.js scripts for scraping and build verification using theBashtool.
Audit Metadata