e2e-testing

Warn

Audited by Gen Agent Trust Hub on Apr 5, 2026

Risk Level: MEDIUMCOMMAND_EXECUTIONCREDENTIALS_UNSAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [COMMAND_EXECUTION]: The skill provides a template for CLI testing that uses subprocess.run(command.split()) to execute strings directly from Gherkin feature files. This pattern is susceptible to command injection if the Gherkin scenario contains malicious input.
  • [CREDENTIALS_UNSAFE]: The authenticated_page fixture template contains hardcoded example credentials admin and admin123.
  • [PROMPT_INJECTION]: The skill architecture creates an indirect prompt injection surface. 1. Ingestion points: Gherkin .feature files described in SKILL.md. 2. Boundary markers: Absent for scenario-derived command strings. 3. Capability inventory: subprocess.run (CLI testing) and playwright browser automation in SKILL.md. 4. Sanitization: Absent; command.split() provides no security boundary against injection.
  • [EXTERNAL_DOWNLOADS]: The skill instructions include installing browser binaries via playwright install chromium, which is a standard procedure from a well-known service.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Apr 5, 2026, 06:13 PM