best-practices

Pass

Audited by Gen Agent Trust Hub on Mar 15, 2026

Risk Level: SAFE
Full Analysis
  • [PROMPT_INJECTION]: No instructions were found that attempt to override agent safety filters or system prompts.
  • [DATA_EXFILTRATION]: No patterns for accessing sensitive files like SSH keys or environment variables were detected.
  • [EXTERNAL_DOWNLOADS]: Code examples reference well-known services such as polyfill.io and trusted-cdn.com for illustrative purposes.
  • [COMMAND_EXECUTION]: Provides examples of legitimate development commands such as npm audit and npm update for security maintenance.
  • [CREDENTIALS_UNSAFE]: No hardcoded credentials or secrets were found; only generic placeholders like 'abc123' are used in examples.
  • [SAFE]: The skill presents an indirect prompt injection surface as it is intended to audit user-provided code. Ingestion points: User-supplied web code (SKILL.md). Boundary markers: Absent. Capability inventory: No executable scripts or filesystem write access within the skill itself. Sanitization: The skill correctly recommends the use of libraries like DOMPurify for sanitizing untrusted HTML in web applications.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 15, 2026, 01:31 AM
Security Audit — agent-trust-hub — best-practices