brand-onboarding

Pass

Audited by Gen Agent Trust Hub on Mar 31, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: Employs Playwright to automate web browsing, capture screenshots, and read page content from user-provided URLs and Instagram handles.
  • [EXTERNAL_DOWNLOADS]: Fetches data and images from external websites and Instagram profiles, saving them to the local assets/ and context/ directories.
  • [PROMPT_INJECTION]: Susceptible to indirect prompt injection as it processes untrusted content from the web. * Ingestion points: Website homepage, about pages, and Instagram profiles accessed via Playwright (SKILL.md). * Boundary markers: Absent. * Capability inventory: File system writes to create directories and save images, screenshots, and markdown documentation (SKILL.md). * Sanitization: Absent. No validation or filtering of extracted web content is specified.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 31, 2026, 07:51 PM
Security Audit — agent-trust-hub — brand-onboarding