shadcn-ui

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (medium risk: 0.65). The required workflow is to install/invoke the shadcn-ui skill and then use its component registry/search features; this typically involves fetching/reading shadcn/ui component registry content from the public upstream repo at runtime, which would be outsider-authored free text ingested into the agent’s LLM context.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.80). The skill explicitly requires running a runtime installer ("npx skills add shadcn-ui/skills") which fetches and executes remote package code (the referenced upstream repository https://github.com/shadcn-ui/ui), so the agent would load external code that can control prompts/behavior.