magic-ui-generator

Pass

Audited by Gen Agent Trust Hub on Mar 23, 2026

Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface where external component descriptions could be used to influence the logic of the generated code integrated into the project. * Ingestion points: Component descriptions provided during the requirement analysis phase (SKILL.md). * Boundary markers: No specific delimiters or "ignore embedded instruction" warnings are defined to isolate untrusted input. * Capability inventory: The skill is capable of writing functional TypeScript code and managing project dependencies (SKILL.md). * Sanitization: There is no mention of sanitization or validation of the generated output before project integration.
  • [EXTERNAL_DOWNLOADS]: The skill instructs the agent to install external UI libraries. * Evidence: Mentions ensuring that lucide-react and framer-motion are installed via package managers (SKILL.md).
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 23, 2026, 05:49 AM