visual-testing-advanced
Warn
Audited by Gen Agent Trust Hub on Feb 16, 2026
Risk Level: MEDIUMPROMPT_INJECTION
Full Analysis
- Prompt Injection (MEDIUM): Vulnerability to indirect prompt injection via untrusted web content. • Ingestion points: Browser automation tasks that navigate to and render external URLs (e.g., page.goto). • Boundary markers: Absent; there are no instructions to isolate or ignore potential prompt injection content embedded in the HTML/CSS of tested pages. • Capability inventory: Browser automation and integration with a 'quality gate' agent that can block or approve CI/CD pipelines based on visual diffs. • Sanitization: None; visual data is processed directly without filtering.
- Metadata Poisoning (LOW): The dependencies field in the YAML frontmatter is empty, despite the skill's code requiring @playwright/test and @percy/playwright.
Audit Metadata