openclaw-orchestrator
Pass
Audited by Gen Agent Trust Hub on Oct 15, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill instructs the installation of several Node.js packages, including @agent-relay/orchestrating-agent-relay, agent-relay, and @agent-relay/openclaw, which are necessary for the skill's functionality.\n- [COMMAND_EXECUTION]: Provides various shell commands for setting up the environment, spawning agents with specific tasks, and managing the lifecycle of the relay nodes. It includes a specific setup command using npx -y @agent-relay/openclaw@latest setup.\n- [INDIRECT_PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection as it handles task prompts for workers and system event messages that could be populated with untrusted data.\n
- Ingestion points: The --task argument in agent-relay node agent spawn and the --text argument in openclaw system event in SKILL.md.\n
- Boundary markers: There are no explicit instructions or delimiters provided to prevent the agent from interpreting instructions that might be embedded within these text inputs.\n
- Capability inventory: The skill allows for the creation of multiple autonomous agents that can communicate over channels, execute tasks, and report events back to the system.\n
- Sanitization: The instructions do not mention any sanitization, filtering, or validation steps for the content passed into these commands.
Audit Metadata