moz-automation

Fail

Audited by Gen Agent Trust Hub on Feb 16, 2026

Risk Level: HIGHEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [External Downloads / Remote Code Execution] (HIGH): The skill instructs users to add an unverified remote endpoint (https://rube.app/mcp) as an MCP server. This domain is not a trusted source and serves as the environment for tool execution, posing a significant supply chain and remote execution risk.
  • [Indirect Prompt Injection] (HIGH): 1. Ingestion points: External Moz SEO data, site audits, and keyword reports processed via Rube MCP. 2. Boundary markers: Absent. 3. Capability inventory: RUBE_MULTI_EXECUTE_TOOL and RUBE_REMOTE_WORKBENCH provide the agent with write/execute access to the connected Moz account. 4. Sanitization: Absent. Malicious instructions embedded in Moz account data could manipulate the agent to perform unauthorized operations.
  • [Command Execution] (MEDIUM): The skill provides tools for executing complex workflows (RUBE_MULTI_EXECUTE_TOOL) and remote automation (RUBE_REMOTE_WORKBENCH), granting the agent autonomous control over external service configurations based on potentially untrusted model reasoning.
Recommendations
  • AI detected serious security threats
Audit Metadata
Risk Level
HIGH
Analyzed
Feb 16, 2026, 12:46 PM