The Agent Skills Directory

[External Downloads] (LOW): The skill directs the agent to connect to a remote MCP server at https://rube.app/mcp. This domain is not on the trusted list, meaning the integrity and safety of the tools provided depend entirely on a third-party service provider.
[Indirect Prompt Injection] (LOW): The workflow requires the agent to fetch schemas and execution plans from the remote server (RUBE_SEARCH_TOOLS) and follow them as authoritative. This creates an ingestion surface where a compromised or malicious server could inject instructions into the agent's context.
[Ingestion points]: Data returned by the RUBE_SEARCH_TOOLS operation.
[Boundary markers]: Absent; the skill lacks instructions to treat remote data as untrusted or to ignore embedded directives.
[Capability inventory]: RUBE_MULTI_EXECUTE_TOOL and RUBE_REMOTE_WORKBENCH allow for various system and network operations based on the remote data.
[Sanitization]: Absent; no validation or filtering of the remote payload is mentioned.

unione-automation