react-bits
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill provides instructions for the agent to suggest shell commands to the user, including
npx shadcn@latest add @react-bits/...and variousnpm installcommands for dependencies likemotion,gsap, andthree.js. These are standard procedures for modern React development and involve well-known tools. - [EXTERNAL_DOWNLOADS]: The skill refers to external resources including a GitHub repository (
DavidHDev/react-bits) and official documentation sites (reactbits.dev,vue-bits.dev). It facilitates the download of component source code through the@react-bitsregistry namespace using the shadcn CLI. - [INDIRECT_PROMPT_INJECTION]: The skill processes user-provided "design briefs" to map them to specific UI components. This creates a potential surface for indirect injection if a user brief were to attempt to influence the agent's logic; however, the skill's structure is informative and the proposed commands are restricted to a specific library catalog, posing low risk.
- Ingestion points: Design briefs provided by the user in the chat session (SKILL.md).
- Boundary markers: Absent.
- Capability inventory: The skill instructs the agent to propose shell commands (
npx,npm) and provide URLs for previews (SKILL.md). - Sanitization: None explicitly mentioned; the agent relies on its internal logic to map text to a static catalog of 130 components.
Audit Metadata