browser-testing-efficiency

Pass

Audited by Gen Agent Trust Hub on Mar 18, 2026

Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill makes extensive use of the agent-browser tool to execute JavaScript in the browser context via eval and perform UI interactions. This is the primary and expected functionality of the skill for browser testing.
  • [INDIRECT_PROMPT_INJECTION]: As a browser automation skill, it interacts with external web content. This inherently creates a surface for indirect prompt injection if a website under test contains malicious instructions. However, the skill implements standard testing patterns and does not introduce specific vulnerabilities.
  • Ingestion points: Browser content accessed via agent-browser snapshot and the execution environment of agent-browser eval (SKILL.md).
  • Boundary markers: None explicitly defined for untrusted web data.
  • Capability inventory: Browser interaction (click, fill, viewport), screenshot capture (snapshot), and arbitrary JavaScript execution (eval) (SKILL.md).
  • Sanitization: The skill uses template literals to inject variables into JavaScript strings, which could be an injection point if variable sources are untrusted, though typical for testing scripts.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 18, 2026, 02:41 AM
Security Audit — agent-trust-hub — browser-testing-efficiency