trello-automation

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

• Third-party content exposure detected (medium risk: 0.65). The required runtime workflow uses Trello tool calls (e.g., TRELLO_GET_* to list boards/lists/cards and TRELLO_GET_SEARCH to find cards), which can ingest outsider-authored Trello content (card titles/descriptions/comments) into the agent context via the tool responses.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly requires adding and calling the Rube MCP endpoint (https://rube.app/mcp) at runtime—using RUBE_SEARCH_TOOLS to fetch tool schemas and invoke remote tools—which directly controls available agent instructions and remote execution, making it a runtime dependency that can control the agent.