webapp-testing

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADS
Full Analysis
  • [External Downloads] (SAFE): The skill instructs the user to install the Playwright framework via npm (npm init playwright@latest). While this downloads external code, Playwright is a reputable industry-standard library, and the skill itself is hosted on a trusted GitHub organization (anthropics), which qualifies for a severity downgrade per [TRUST-SCOPE-RULE].
  • [Indirect Prompt Injection] (LOW): This category is flagged because the skill navigates to and interacts with external or local web pages (e.g., page.goto). This creates a surface where an attacker-controlled web page could provide instructions to the agent.
  • Ingestion points: page.goto, page.locator, page.content().
  • Boundary markers: None explicitly defined in the provided snippets.
  • Capability inventory: Browser navigation, screenshot capture, form interaction.
  • Sanitization: None.
  • Assessment: This is an inherent risk of browser automation tools; however, the severity is low given the intended testing use case and the agent's internal safety guardrails.
  • [Data Exposure] (SAFE): The skill includes functionality to capture screenshots (page.screenshot). This is a legitimate feature for a testing toolkit and does not involve unauthorized exfiltration of sensitive system files.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 05:03 PM