Skills
Navigate
skills/proffesor-for-testing/agentic-qe/security-visual-testing/Gen Agent Trust Hub

security-visual-testing

Pass

Audited by Gen Agent Trust Hub on Sep 29, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external content via browser automation, which introduces a potential attack surface for indirect prompt injection.
  • Ingestion points: The vibium go "$TARGET_URL" command in SKILL.md navigates to and ingests content from arbitrary target URLs.
  • Boundary markers: The skill documentation notes a URL validation step that blocks malicious schemes such as javascript:, data:, and file:.
  • Capability inventory: The skill executes CLI tools (aqe, vibium), runs Node.js scripts, and writes screenshot files to the local file system.
  • Sanitization: The skill utilizes a check-injection.js script to scan page content for malicious patterns before screenshots are processed or stored.
  • [COMMAND_EXECUTION]: The skill utilizes several CLI tools and scripts to perform its intended testing functions.
  • Evidence: SKILL.md documents the execution of the vibium browser engine, the aqe testing CLI, and specific Node.js scripts (check-injection.js, visual-diff.js) located within the agent's skill directory.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 29, 2026, 06:06 AM