The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill instructions direct users to add 'https://rube.app/mcp' as a remote Model Context Protocol (MCP) server. This setup allows the external server to provide tool definitions that the agent executes, effectively granting a third-party control over the agent's available actions.
[DATA_EXFILTRATION]: By using the rube.app MCP server as an intermediary, sensitive Miro data—such as board details, item content, and user email addresses (via the MIRO_SHARE_BOARD tool)—is potentially exposed to the third-party service provider.
[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because it ingests untrusted data from external sources.
Ingestion points: Board content and sticky note data are retrieved through the MIRO_GET_BOARD_ITEMS tool in SKILL.md.
Boundary markers: Absent. The skill provides no instructions to use delimiters or to disregard instructions found within Miro board items.
Capability inventory: The skill can share boards (MIRO_SHARE_BOARD), create/modify items (MIRO_CREATE_ITEMS_IN_BULK), and manage account connections (RUBE_MANAGE_CONNECTIONS).
Sanitization: Absent. There is no requirement or instruction to sanitize data retrieved from Miro before it is processed by the agent's context.
[EXTERNAL_DOWNLOADS]: The skill relies on an external, unverified domain ('rube.app') to host the MCP server configuration and tool logic, which is not associated with the primary service (Miro) or a known trusted vendor.

miro-automation