box-automation

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

• Third-party content exposure detected (high risk: 0.80). The SKILL.md workflows (e.g., "Search and Browse Content" and "Upload and Download Files") explicitly require using BOX_SEARCH_FOR_CONTENT and BOX_DOWNLOAD_FILE to fetch and read user-generated Box files and web_links (third-party content) which the agent is expected to interpret and use to decide subsequent actions.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill requires connecting to the Rube MCP endpoint https://rube.app/mcp at runtime (via RUBE_SEARCH_TOOLS) to fetch current tool schemas which directly determine the agent's available instructions and behavior, so this external endpoint is a required runtime dependency that can control prompts.