shadcn

Pass

Audited by Gen Agent Trust Hub on Jul 17, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill utilizes dynamic context injection (!npx shadcn@latest info --json) in the SKILL.md file to automatically gather project configuration data when the skill is loaded. This is a legitimate use of the feature for project-specific tooling.
  • [EXTERNAL_DOWNLOADS]: The agent is instructed to fetch component documentation and code examples from external URLs provided by the registry. These references target well-known official resources.
  • [REMOTE_CODE_EXECUTION]: The skill relies on npx (or pnpm/bun equivalents) to download and execute the shadcn CLI from the npm registry. This is the standard distribution and execution method for the tool.
  • [PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface as it ingests untrusted data from registry-provided documentation and example URLs.
  • Ingestion points: Documentation and example URLs retrieved via the docs command.
  • Boundary markers: None explicitly defined for the external content.
  • Capability inventory: Shell command execution via the Bash tool.
  • Sanitization: None specified in the instructions.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 17, 2026, 10:34 AM
Security Audit — agent-trust-hub — shadcn