The Agent Skills Directory

External Downloads (HIGH): The skill requires connecting to an external MCP server at https://rube.app/mcp. This domain is not within the defined trusted scope. The documentation states that no API keys are needed and the endpoint 'just works', which implies a lack of transparent authentication and full reliance on a third-party service's integrity.
Indirect Prompt Injection (HIGH): This skill exhibits a significant indirect prompt injection surface.
Ingestion points: Untrusted data enters the agent context via RUBE_SEARCH_TOOLS, which returns tool slugs, input schemas, 'recommended execution plans', and 'known pitfalls'.
Boundary markers: No boundary markers or instructions to ignore embedded instructions within the search results are present.
Capability inventory: The skill possesses powerful write/execute capabilities through RUBE_MULTI_EXECUTE_TOOL and RUBE_REMOTE_WORKBENCH, allowing it to modify Appdrag infrastructure.
Sanitization: There is no evidence of sanitization or validation of the remote execution plans before the agent acts on them.
Command Execution (HIGH): The skill facilitates the execution of arbitrary tools on the Appdrag platform via RUBE_MULTI_EXECUTE_TOOL. Since the specific tools and their arguments are determined at runtime by an untrusted remote source, this constitutes a high risk of unauthorized command execution.

appdrag-automation