openclaw - 💡(How to fix) Fix Feature request: per-agent default cron delivery accountId in config

Problem

When an agent dynamically creates a cron job at runtime (e.g., a background check-in or pipeline monitor job), the delivery.accountId field defaults to "default" unless explicitly set. This causes notifications to route through the default agent's channel (typically the main Orchestrator account) instead of the creating agent's own account/channel.

This is a persistent operational issue in multi-agent setups. AGENTS.md rules can mitigate it, but AGENTS.md context can be truncated in long sessions, making this an unreliable enforcement point.

Expected behavior

A config-level setting that specifies the default accountId for any cron job created by a given agent. For example:

{
  agents: {
    list: [
      {
        id: "recon",
        cron: {
          defaultDelivery: {
            accountId: "recon"
          }
        }
      }
    ]
  }
}

Or at the global default level:

{
  agents: {
    defaults: {
      cron: {
        defaultDelivery: {
          accountId: "<inherit-agent-id>"  // special value to auto-use agent's own account
        }
      }
    }
  }
}

Why this matters

• In multi-agent deployments, each agent has its own channel account (e.g., separate Telegram bots, Teams accounts)
• Agents that self-schedule cron jobs during runs should route notifications through their own account by default
• Without config-level enforcement, this relies on agents reading and retaining instructions from AGENTS.md — which can be truncated in long sessions, causing silent regressions
• A sensible default would be: if an agent creates a cron job, inherit that agent's accountId on the delivery block automatically

Workaround

Currently the only workaround is to include the rule in the agent's AGENTS.md and hope context isn't truncated. This is fragile.

Environment

• OpenClaw 2026.5.26 (10ad3aa)
• Multi-agent setup with 15+ agents across Telegram, Teams, and other channels