setup
Warn
Audited by Gen Agent Trust Hub on Feb 15, 2026
Risk Level: MEDIUMEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTION
Full Analysis
- REMOTE_CODE_EXECUTION (MEDIUM): The skill frequently uses
npx assistant-ui@latest, which downloads and executes the latest version of the CLI tool from the npm registry at runtime. This introduces a dependency on the integrity of the npm package and the assistant-ui maintainers. - EXTERNAL_DOWNLOADS (MEDIUM): The
addcommand fetches component definitions fromhttps://r.assistant-ui.com/{name}.json. This external registry acts as an ingestion point for code that is subsequently added to the user's project. - COMMAND_EXECUTION (LOW): The skill provides numerous commands for project initialization (
init,create) and component management (add). These are standard development operations but grant the agent significant influence over the local file system and project configuration. - INDIRECT_PROMPT_INJECTION (LOW): The skill reads from external sources like
https://www.assistant-ui.com/llms.txtand registry JSON files. While these are intended for documentation and configuration, they represent a surface where untrusted data could influence agent behavior during setup.
Audit Metadata