The Agent Skills Directory

Dynamic Execution (LOW): The skill utilizes RUBE_REMOTE_WORKBENCH and RUBE_MULTI_EXECUTE_TOOL to perform operations. This involves running tool-based logic on an external workbench, which is the primary intended purpose but carries inherent risks associated with remote execution environments.
Unverifiable Dependencies & Remote Code Execution (LOW): The skill requires connecting to an external MCP server at https://rube.app/mcp. As this domain is not on the trusted sources list, it is treated as an unverifiable external dependency.
Indirect Prompt Injection (LOW): The skill discovery mechanism (RUBE_SEARCH_TOOLS) ingests external tool schemas to determine agent actions, creating an attack surface where a compromised or malicious remote server could influence agent behavior.
Ingestion points: Runtime response data from the rube MCP server tools.
Boundary markers: Absent; the agent is instructed to follow the discovered schemas directly without delimiters or explicit warnings.
Capability inventory: Remote tool execution, connection management, and workbench access via RUBE_MULTI_EXECUTE_TOOL and RUBE_REMOTE_WORKBENCH.
Sanitization: No sanitization or validation of the discovered tool schemas or arguments is performed by the skill instructions.

serply-automation