convertapi-automation

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • Unverifiable Dependencies & Remote Code Execution (LOW): The skill directs users to add an external MCP server endpoint (https://rube.app/mcp). While this is functional for the skill, the domain is not on the trusted organizational list, constituting an untrusted external dependency.- Indirect Prompt Injection (LOW): The skill relies on dynamically retrieved data from RUBE_SEARCH_TOOLS to determine its actions, which could be exploited to influence agent behavior.
  • Ingestion points: Data enters the context via the results of RUBE_SEARCH_TOOLS (SKILL.md).
  • Boundary markers: Absent; there are no instructions to the agent to disregard instructions found within the tool schemas.
  • Capability inventory: The skill provides tools for execution like RUBE_MULTI_EXECUTE_TOOL and RUBE_REMOTE_WORKBENCH (SKILL.md).
  • Sanitization: No sanitization is performed on the dynamic tool schemas.- Dynamic Execution (LOW): The skill uses RUBE_MULTI_EXECUTE_TOOL to execute tools that are identified and described at runtime via an external discovery service.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 06:36 PM