browser-testing-with-devtools

Pass

Audited by Gen Agent Trust Hub on Jul 15, 2026

Risk Level: SAFE
Full Analysis
  • [PROMPT_INJECTION]: The skill contains defensive instructions to ignore command-like patterns (such as 'Ignore previous instructions') discovered within web pages, which serves as a mitigation against indirect prompt injection. It also defines clear boundary markers between trusted user instructions and untrusted browser data.
  • [EXTERNAL_DOWNLOADS]: The skill configures the installation of the chrome-devtools-mcp package from the npm registry using npx. This is the standard deployment method for the tool required by the skill.
  • [COMMAND_EXECUTION]: A configuration command is provided to run the MCP server. This is necessary for the skill to communicate with the browser and is performed within the tool's restricted execution environment.
  • [DATA_EXFILTRATION]: While the agent is granted the ability to read browser state (DOM, network traffic, console logs), the skill provides mandatory constraints against reading authentication material like cookies or localStorage, and prohibits making external network calls from the browser context.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 15, 2026, 11:47 AM
Security Audit — agent-trust-hub — browser-testing-with-devtools