owl-protocol-automation

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • EXTERNAL_DOWNLOADS (LOW): The skill references an external MCP endpoint at https://rube.app/mcp. While necessary for the skill's functionality, this domain is not on the trusted sources list and acts as a remote dependency for extending agent capabilities.
  • COMMAND_EXECUTION (LOW): The skill instructs the agent to use RUBE_MULTI_EXECUTE_TOOL and RUBE_REMOTE_WORKBENCH to perform operations. These tools execute logic based on parameters and schemas retrieved at runtime.
  • Indirect Prompt Injection (LOW): The workflow relies on ingesting external data (tool schemas and execution plans) which could be used to influence agent behavior if the source is compromised.
  • Ingestion points: Data returned by the RUBE_SEARCH_TOOLS function as described in SKILL.md.
  • Boundary markers: Absent; the instructions do not define delimiters or provide 'ignore embedded instructions' warnings for the fetched tool metadata.
  • Capability inventory: The skill possesses the ability to execute complex operations via RUBE_MULTI_EXECUTE_TOOL and RUBE_REMOTE_WORKBENCH (SKILL.md).
  • Sanitization: Absent; the skill does not outline any validation or sanitization steps for the dynamically retrieved tool definitions.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 06:43 PM