webapp-testing

Pass

Audited by Gen Agent Trust Hub on Jul 18, 2026

Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The helper script scripts/with_server.py uses subprocess.Popen with shell=True to execute commands provided in the --server argument. This allows the execution of arbitrary shell strings, which is a powerful capability if the inputs are influenced by untrusted content.\n- [DATA_EXFILTRATION]: The script examples/console_logging.py records all browser console messages to a file at /mnt/user-data/outputs/console.log. This can lead to the exposure of sensitive information, such as API keys or session tokens, if the web application under test logs them to the console.\n- [PROMPT_INJECTION]: The skill processes data from external web applications, which are untrusted sources, and has access to shell execution tools. This creates an indirect prompt injection surface.\n
  • Ingestion points: page.content(), page.locator().all(), and browser console event handlers in Playwright scripts.\n
  • Boundary markers: None; content is treated as trusted strings without delimiters.\n
  • Capability inventory: Shell execution via the scripts/with_server.py wrapper.\n
  • Sanitization: None implemented.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 18, 2026, 01:48 PM
Security Audit — agent-trust-hub — webapp-testing