e2e-testing-patterns

Pass

Audited by Gen Agent Trust Hub on Sep 14, 2026

Risk Level: SAFE
Full Analysis
  • [PROMPT_INJECTION]: No evidence of prompt injection or instructions to bypass safety guardrails. The skill uses standard instructional language focused on testing workflows.
  • [DATA_EXFILTRATION]: No sensitive data access or unauthorized network exfiltration patterns detected. Code snippets use standard placeholders like 'user@example.com' and reference environment variables for secrets, which is a recommended security practice.
  • [OBFUSCATION]: No obfuscated code, hidden URLs, or suspicious character encoding techniques were found. All code snippets are provided in clear, readable TypeScript.
  • [REMOTE_CODE_EXECUTION]: No dangerous remote code execution patterns like piped shell commands from external URLs were found. References to package installations (e.g., axe-core/playwright) target well-known, legitimate testing libraries.
  • [COMMAND_EXECUTION]: Shell commands mentioned (e.g., npx playwright test) are standard CLI usage for the documented testing frameworks and do not involve unsanitized user input.
  • [DYNAMIC_CONTEXT_INJECTION]: No dynamic context injection markers (!command) were found in the skill metadata or content.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 14, 2026, 07:39 PM
Security Audit — agent-trust-hub — e2e-testing-patterns