Inter-agent delegation failing: sessions_spawn with agentId blocked in sandboxed multi-agent setup
Running a multi-agent setup on Mac Mini M4 with 6 registered agents — one orchestrator agent and 5 specialist agents. All agents show up correctly in openclaw agents list with their own workspaces and SOUL.md files.
When the orchestrator agent tries to delegate to a specialist agent using sessions_spawn with agentId: agent-id-here, getting this error:
agentId is not allowed for sessions_spawn (allowed: none)
Current tool deny list in openclaw.json:
"browser", "exec", "process", "apply_patch"
Questions:
1. Is sessions_spawn with agentId the correct mechanism for inter-agent delegation?
2. Is the sandbox blocking this — and if so, what’s the specific config to allow it?
3. What’s the correct way for one registered agent to delegate a task to another registered agent and receive its output?
Running OpenClaw on macOS with Docker sandbox enabled, agents communicating via Telegram.
When the orchestrator agent tries to delegate to a specialist agent using sessions_spawn with agentId: agent-id-here, getting this error:
agentId is not allowed for sessions_spawn (allowed: none)
Current tool deny list in openclaw.json:
"browser", "exec", "process", "apply_patch"
Questions:
1. Is sessions_spawn with agentId the correct mechanism for inter-agent delegation?
2. Is the sandbox blocking this — and if so, what’s the specific config to allow it?
3. What’s the correct way for one registered agent to delegate a task to another registered agent and receive its output?
Running OpenClaw on macOS with Docker sandbox enabled, agents communicating via Telegram.
The AI that actually does things. Emails, calendar, home automation - all from your favorite chat app. New shell, ready to help. The lobster way.
131,281Members
Resources
Similar Threads
Was this page helpful?
