lp-design

Pass

Audited by Gen Agent Trust Hub on Jun 16, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill instructions and processes are focused on professional design tasks, such as generating CSS custom properties and WCAG contrast verification. No malicious code or obfuscation is present.\n- [PROMPT_INJECTION]: The skill incorporates a knowledge-ingestion mechanism to read local design principles, which represents a standard surface for indirect prompt injection in agentic workflows.\n
  • Ingestion points: Markdown files located in ../landing-page/knowledge/ are read via the read tool to inform design logic (SKILL.md).\n
  • Boundary markers: None; the skill does not explicitly use delimiters to separate ingested knowledge data from the agent's core instructions.\n
  • Capability inventory: The skill has access to tools including bash, write, and edit (SKILL.md frontmatter).\n
  • Sanitization: Content from the local knowledge files is used as-is without validation or escaping.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 16, 2026, 10:10 PM
Security Audit — agent-trust-hub — lp-design