shadcn
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill instructs the agent to use
npx shadcn@latestandnext-themesfrom the npm registry, which are standard and well-known developer tools. - [COMMAND_EXECUTION]: Provides instructions for running shell commands to initialize the library and manage components using npm, pnpm, and yarn.
- [INDIRECT_PROMPT_INJECTION]: The skill defines UI templates (Tables, Forms, Cards) that ingest and display data, which is an inherent attack surface for indirect prompt injection if the rendered data contains malicious instructions for an observing agent.
- Ingestion points: UI component templates in
SKILL.mdandresources/form-patterns.md(e.g.,PostsTable,PostForm) process application data for display. - Boundary markers: No explicit delimiters or instructions are provided in the UI patterns to separate data from potential instructions.
- Capability inventory: The skill facilitates component installation and file-writing via standard agent capabilities.
- Sanitization: Standard React auto-escaping is used, but no specific mitigation for indirect prompt injection in rendered content is highlighted.
Audit Metadata