shadcn-vue
Fail
Audited by Gen Agent Trust Hub on Oct 6, 2026
Risk Level: CRITICALDYNAMIC_CONTEXT_INJECTIONCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [DYNAMIC_CONTEXT_INJECTION]: The skill uses the placeholder syntax
!npx shadcn-vue@latest info --json`` inSKILL.mdto gather project context automatically when loaded. - [COMMAND_EXECUTION]: The instructions direct the agent to use CLI runners like
npxto perform significant project modifications, including initializing configurations and installing source code. - [INDIRECT_PROMPT_INJECTION]: The MCP server documentation in
references/mcp.mddescribes how the agent can interact with external registries, which presents a surface where untrusted registry providers could influence agent behavior. - [DYNAMIC_EXECUTION]: The documentation generator script
scripts/generate-components.tsprogrammatically fetches data from the GitHub API and raw content URLs to populate local files. - [EXTERNAL_DOWNLOADS]: The skill encourages downloading various Node.js packages and UI components from official but external sources like GitHub and the NPM registry.
Recommendations
- CRITICAL: 8 file(s) identified as malware by FileRep - DO NOT USE
Audit Metadata