The Agent Skills Directory

[No Code] (SAFE): The skill consists of text-based instructions and metadata only. There are no scripts or binary files, eliminating the risk of local code execution.\n- [Indirect Prompt Injection] (LOW): The skill possesses an indirect prompt injection surface. Evidence: (1) Ingestion points: Data enters the context via RUBE_SEARCH_TOOLS output. (2) Boundary markers: No delimiters or 'ignore embedded instructions' warnings are present in the provided templates. (3) Capability inventory: The skill can trigger subprocesses via RUBE_MULTI_EXECUTE_TOOL and RUBE_REMOTE_WORKBENCH. (4) Sanitization: No sanitization, escaping, or validation of the fetched tool schemas is defined.\n- [External Dependency] (SAFE): The skill references an external MCP server (rube.app/mcp). While this is an untrusted third-party source, referencing it is the fundamental purpose of the skill and does not involve the direct download of executable scripts or binaries.

zenserp-automation