shadcn
Pass
Audited by Gen Agent Trust Hub on Oct 15, 2026
Risk Level: SAFEDYNAMIC_CONTEXT_INJECTIONREMOTE_CODE_EXECUTIONINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [DYNAMIC_CONTEXT_INJECTION]: The skill leverages the dynamic context injection syntax (
!) inSKILL.mdto execute the commandnpx shadcn@latest info --json. This allows the agent to retrieve real-time project configuration, such as Tailwind version, aliases, and installed components, without manual user input. - [REMOTE_CODE_EXECUTION]: The skill guides the agent through installing code from both official shadcn/ui registries and third-party GitHub repositories using
npx shadcn@latest add. To mitigate supply chain risks, the skill explicitly instructs the agent to use--dry-run,--diff, and--viewflags to audit component source code and CSS changes before installation. - [INDIRECT_PROMPT_INJECTION]: The skill presents an attack surface by processing data from documentation URLs and third-party registry item payloads. However, it implements security guidelines by instructing the agent to fetch documentation from vendor-verified domains and requiring manual verification of all external files and dependencies.
- [COMMAND_EXECUTION]: The skill provides a reference for executing various CLI commands and utility scripts, such as
showcase-coverage.mjs, to manage the UI development lifecycle. These tools are project-specific and operate within the scope of a standard frontend development environment.
Audit Metadata