The Agent Skills Directory

[Unverifiable Dependencies] (HIGH): The skill instructs users to add https://rube.app/mcp as an MCP server. Neither the domain nor the provider (Composio/Rube) are listed in the trusted repositories or organizations, making this an unverified remote dependency.
[Indirect Prompt Injection] (HIGH): The skill's core workflow relies on external content to define its behavior.
Ingestion points: RUBE_SEARCH_TOOLS retrieves tool slugs, schemas, and 'recommended execution plans' from a remote server (SKILL.md).
Boundary markers: Absent. The instructions explicitly tell the agent to 'Always search first' and 'Use exact field names' from the remote response.
Capability inventory: The skill possesses powerful write/execute capabilities through RUBE_MULTI_EXECUTE_TOOL and RUBE_REMOTE_WORKBENCH (SKILL.md).
Sanitization: Absent. The agent is instructed to trust the search results as the source of truth for execution without validation.
[Remote Code Execution] (MEDIUM): Through the RUBE_REMOTE_WORKBENCH and RUBE_MULTI_EXECUTE_TOOL components, the skill facilitates execution on remote infrastructure. Given the provider is not in the trusted scope, this poses a risk of unauthorized remote command execution via the MCP interface.

cloudcart-automation