enhance-web-landing

Pass

Audited by Gen Agent Trust Hub on Oct 26, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill references and provides installation instructions for various industry-standard design systems and libraries, including Google's Material UI, Microsoft's Fluent UI, IBM's Carbon, GitHub's Primer, and Atlassian's Design System. These are reputable resources from trusted organizations.
  • [COMMAND_EXECUTION]: The skill includes standard package management commands (npm install, yarn add) for the aforementioned libraries to facilitate the setup of a modern web development environment.
  • [INDIRECT_PROMPT_INJECTION]: The skill defines an interaction model where it processes user-provided design briefs to generate code. While this represents a potential attack surface for indirect prompt injection, the skill includes detailed logic for 'reading the room' and adhering to strict design guardrails, which serves as a natural constraint on the output.
  • Ingestion points: User-provided design briefs and vibe signals (SKILL.md Section 0).
  • Boundary markers: The instructions mandate a 'Design Read' one-liner before code generation to align context, though no explicit adversarial delimiters are used.
  • Capability inventory: The skill is designed to generate React, Next.js, and Tailwind CSS code. It recommends shell commands for package installation.
  • Sanitization: The skill focuses on design engineering logic and adherence to hard layout rules (e.g., contrast checks, viewport fitting) which naturally sanitize the resulting UI structure.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 26, 2026, 12:53 PM