Subagent

[haughty adder]

If I have a cron job for one agent, and I want to have it call another agent to get some data, how can I do this?

[kind lotus]

To help others find answers, you can mark your question as solved via Right click solution message -> Apps -> ✅ Mark Solution

[jovial fossil]

Yep — the clean way is to make the cron agent use the session/subagent tools.

Two common patterns:

1. One-off data fetch from another agent
   Have the cron prompt tell Agent A to spawn Agent B as a sub-agent, e.g. “spawn the weather agent to collect today’s data, then summarize it.” Under the hood that uses sessions_spawn with agentId: "weather". The child runs separately and its result is handed back to the cron/requester session.

   Config-wise, the cron/requester agent needs:

   • sessions_spawn available (tools.profile: "coding" or allow sessions_spawn / group:sessions)
   • permission to target the other agent via subagents.allowAgents, e.g.

{
  agents: {
    list: [
      {

        id: "cron-agent",
        subagents: {
          allowAgents: ["data-agent"]
        }
      },
      { id: "data-agent" }
    ]
  }
}

Then the cron message can be something like:

Ask the data-agent subagent to fetch X. Wait for its result, then produce the final report.

2. Talk to an existing persistent session
   If you want to message an already-existing session instead of spawning a fresh one, use sessions_send. For cross-agent/session visibility you’ll usually need broader session visibility, e.g. tools.sessions.visibility: "all", and for cross-agent targeting enable tools.agentToAgent.

For most cron workflows I’d recommend pattern 1: sessions_spawn + agentId. It’s cleaner, isolated, and cron/subagent completion delivery is designed for this.

Docs:
https://docs.openclaw.ai/tools/subagents
https://docs.openclaw.ai/concepts/session-tool
https://docs.openclaw.ai/automation/cron-jobs