facebook-automation

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

• Third-party content exposure detected (high risk: 1.00). The SKILL.md explicitly requires using RUBE_MANAGE_CONNECTIONS / the "facebook" toolkit (and RUBE_SEARCH_TOOLS) to list and read Facebook pages, posts, and user comments—user-generated, public third-party content that the agent is expected to interpret and act on (e.g., reply/hide comments), allowing indirect instruction injection.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill requires adding and calling the MCP server at https://rube.app/mcp during runtime (via RUBE_SEARCH_TOOLS and RUBE_MANAGE_CONNECTIONS) to fetch tool schemas that directly control agent behavior, so this external URL is a runtime dependency that can influence prompts/instructions.