delegating-to-omp-rpc
Pass
Audited by Gen Agent Trust Hub on Jul 17, 2026
Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill documentation includes examples of executing shell commands using the omp-rpc utility to manage agent sessions. This includes starting sessions with specific working directories and model selections, and sending task descriptions to the sub-agent.
- [PROMPT_INJECTION]: The skill facilitates an indirect prompt injection surface by passing task instructions to a secondary agent. 1. Ingestion points: The omp-rpc send command in SKILL.md takes task descriptions as input. 2. Boundary markers: Instructions are typically enclosed in quotes within the command line. 3. Capability inventory: The sub-agent has the capability to perform coding and research tasks and execute shell tools. 4. Sanitization: The skill explicitly notes that destructive shell commands are auto-denied by the tool and suggests using throwaway containers for high-risk tasks.
Audit Metadata