The Agent Skills Directory

[EXTERNAL_DOWNLOADS] (MEDIUM): The skill requires connecting to an external MCP server at https://rube.app/mcp. This domain is not recognized as a trusted source, introducing risk from a third-party service provider.
[PROMPT_INJECTION] (HIGH): The skill is highly vulnerable to Indirect Prompt Injection (Category 8) because it uses external data to drive agent logic.
Ingestion points: Untrusted data enters the context via the RUBE_SEARCH_TOOLS response.
Boundary markers: Absent. The skill explicitly directs the agent to 'Always search tools first' and use the returned 'recommended execution plans'.
Capability inventory: RUBE_MULTI_EXECUTE_TOOL and RUBE_REMOTE_WORKBENCH provide the ability to execute write operations and bulk tasks on the connected Alchemy account.
Sanitization: Absent. The instructions mandate using 'exact field names and types from the search results' without validation.
[COMMAND_EXECUTION] (HIGH): By dynamically fetching tool definitions and execution plans from a remote source and then executing them via RUBE_MULTI_EXECUTE_TOOL, the skill allows the remote server to dictate which operations are performed on the user's Alchemy instance, effectively allowing remote command execution.

alchemy-automation