gluestack-mcp-tools

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

• Third-party content exposure detected (high risk: 0.80). The SKILL.md explicitly describes a GitHub Mode that "fetches components directly from the gluestack-ui GitHub repository," so the agent will retrieve and interpret public, user-authored GitHub content (source code, Storybook demos, and metadata) as part of its workflow, which could contain instructions that influence subsequent actions.

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

• Potentially malicious external URL detected (high risk: 0.70). The skill fetches component source and demos at runtime from the gluestack-ui GitHub repository (https://github.com/TheBushidoCollective/han), and those raw files are injected into the agent context as required data, which can directly influence model outputs.