ui-test

Pass

Audited by Gen Agent Trust Hub on Jul 19, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONDATA_EXFILTRATIONREMOTE_CODE_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [SAFE]: No malicious code or patterns were identified. All actions taken by the skill align with its primary purpose as a UI testing tool.\n- [EXTERNAL_DOWNLOADS]: The skill requires the installation of the 'browse' CLI and dynamically loads the 'axe-core' accessibility library from 'cdnjs.cloudflare.com'. These are standard dependencies for the tool's intended QA functionality and are provided by the vendor or well-known services.\n- [DATA_EXFILTRATION]: The skill facilitates the synchronization of local browser cookies to Browserbase cloud browsers using a 'cookie-sync' script. This is the intended mechanism for testing authenticated pages on deployed sites and is a documented feature.\n- [COMMAND_EXECUTION]: The skill uses the 'browse' CLI to perform interactions like clicking and taking screenshots, and the 'Agent' tool to coordinate parallel test execution among sub-agents.\n- [REMOTE_CODE_EXECUTION]: The skill executes JavaScript in the browser context via 'browse eval' to perform deterministic audits. It also loads the 'axe-core' library from a trusted CDN during accessibility testing.\n- [PROMPT_INJECTION]: The skill interacts with external web content (DOM snapshots and accessibility trees). It manages the risk of indirect prompt injection by using structured assertion protocols, evidence-based reporting, and fixed sub-agent step budgets.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 19, 2026, 05:54 AM
Security Audit — agent-trust-hub — ui-test