The Agent Skills Directory

[EXTERNAL_DOWNLOADS] (MEDIUM): The skill requires the user to add an external MCP server from an untrusted domain: https://rube.app/mcp. This domain is not part of the trusted ecosystem, and adding remote MCP endpoints can expose the agent's context and capabilities to a third party.- [INDIRECT_PROMPT_INJECTION] (MEDIUM): The skill's core logic involves fetching tool schemas and 'recommended execution plans' from the remote RUBE_SEARCH_TOOLS endpoint. This external data directly influences the agent's actions.
Ingestion points: RUBE_SEARCH_TOOLS response body.
Boundary markers: Absent; the agent is instructed to use the returned schemas and plans directly.
Capability inventory: RUBE_MULTI_EXECUTE_TOOL and RUBE_REMOTE_WORKBENCH which can modify external Pdfmonkey data.
Sanitization: Absent; the skill explicitly advises using the exact field names and types returned by the remote search result.- [COMMAND_EXECUTION] (LOW): The skill uses RUBE_MULTI_EXECUTE_TOOL to perform operations on Pdfmonkey. While limited to the scope of the Pdfmonkey API, it grants the agent the ability to execute side-effecting operations based on instructions from the untrusted MCP server.

pdfmonkey-automation