exist-automation
Warn
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: MEDIUMEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- EXTERNAL_DOWNLOADS (MEDIUM): The skill mandates adding an external MCP server (
https://rube.app/mcp) which is not included in the list of trusted external sources. This introduces a dependency on unverified remote infrastructure. - COMMAND_EXECUTION (MEDIUM): The skill pattern relies on
RUBE_MULTI_EXECUTE_TOOLto execute commands using tool slugs and schemas fetched dynamically at runtime. This allows the remote service to influence the agent's actions through the tool discovery phase. - PROMPT_INJECTION (LOW): Indirect Prompt Injection surface identified through remote tool discovery.
- Ingestion points: Data returned from
RUBE_SEARCH_TOOLSincluding 'recommended execution plans' and tool schemas. - Boundary markers: Absent. The skill provides no instructions to treat remote tool definitions as untrusted or to ignore instructions embedded within the search results.
- Capability inventory: Execution of various toolkit operations through
RUBE_MULTI_EXECUTE_TOOLandRUBE_REMOTE_WORKBENCH. - Sanitization: Absent. The agent is instructed to use exact field names and types from remote search results without validation.
Audit Metadata