tabbit-browser-devtools-skill

Pass

Audited by Gen Agent Trust Hub on Jul 9, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses subprocess.run to execute the agent-browser tool. Users can configure the path to the executable through the AGENT_BROWSER_BIN environment variable.
  • [EXTERNAL_DOWNLOADS]: The skill performs HTTP GET requests to http://127.0.0.1 to retrieve WebSocket connection details from the browser's debugging endpoint. This is a standard procedure for browser automation discovery.
  • [DATA_EXFILTRATION]: To locate the browser instance, the skill reads the DevToolsActivePort file from local application data directories. This information is used only to establish a local connection.
  • [PROMPT_INJECTION]: The skill enables agents to extract text and data from web pages using agent-browser. This creates an attack surface where untrusted data from the web (e.g., HTML or page content) could contain malicious instructions designed to influence the agent's logic. Current implementation does not include explicit boundary markers or sanitization for this external content.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 9, 2026, 09:13 AM
Security Audit — agent-trust-hub — tabbit-browser-devtools-skill