The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill requires the configuration of an external MCP server at https://rube.app/mcp to function. This server provides tool definitions and schemas dynamically at runtime from a third-party domain.
[COMMAND_EXECUTION]: The inclusion of the RUBE_REMOTE_WORKBENCH tool indicates that the skill facilitates the execution of code or commands within a remote environment managed by the service provider.
[DATA_EXFILTRATION]: Workflow execution involves routing authentication and operational data through the rube.app infrastructure via RUBE_MANAGE_CONNECTIONS and RUBE_MULTI_EXECUTE_TOOL.
[PROMPT_INJECTION]: The skill exhibits an indirect prompt injection attack surface.
Ingestion points: Data is ingested through RUBE_SEARCH_TOOLS which fetches tool schemas and plans from a remote server.
Boundary markers: No instructions are provided for the agent to treat search results as untrusted or to ignore instructions embedded within the schemas.
Capability inventory: The skill possesses high-privilege tools including RUBE_MULTI_EXECUTE_TOOL and RUBE_REMOTE_WORKBENCH for command execution.
Sanitization: No validation or sanitization of the remote tool definitions is described before execution.

waiverfile-automation