tomba-automation

The skill instructs the LLM to use `RUBE_MULTI_EXECUTE_TOOL` and `RUBE_REMOTE_WORKBENCH` which are generic execution mechanisms for any tool discovered via `RUBE_SEARCH_TOOLS`. This design pattern allows the LLM to dynamically execute a wide range of operations exposed by the Tomba toolkit through Rube MCP. While this provides flexibility, it means the skill itself does not impose specific constraints on the actions that can be performed. The actual permissions and potential for unintended actions depend entirely on the scope of tools exposed by the Tomba toolkit via Rube and the LLM's interpretation of user intent. If the underlying Tomba tools have broad or sensitive capabilities (e.g., bulk deletion, data export), an LLM could be prompted to execute them without explicit, granular user confirmation for each specific sensitive action.