frontend-design

Pass

Audited by Gen Agent Trust Hub on May 20, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill's instructions are focused entirely on creative direction and aesthetic standards for frontend development. No malicious code execution, network exfiltration, or credential harvesting patterns were found.
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to process user-provided frontend requirements, which creates a potential surface for indirect prompt injection if user inputs contain instructions intended to deviate from the skill's purpose.
  • Ingestion points: User-provided frontend requirements, components, or application specifications (SKILL.md).
  • Boundary markers: Absent; there are no specific delimiters to isolate user input from the skill's system instructions.
  • Capability inventory: Generates functional frontend code (HTML/CSS/JS, React, Vue).
  • Sanitization: Not explicitly defined within the skill; relies on the underlying model's safety guardrails.
Audit Metadata
Risk Level
SAFE
Analyzed
May 20, 2026, 12:17 PM
Security Audit — agent-trust-hub — frontend-design