Problem

In multi-agent setups, plugins.slots (e.g., memory) is a global setting — it applies to all agents. There's no way to configure different plugins per agent for the same slot.

Real-world example

I run three agents: a general assistant, a job search agent, and a research agent (mem0claw). I wanted only the research agent to use the @mem0/openclaw-mem0 memory plugin (Mem0 Cloud), while keeping the other two on the default memory-core.

Installing the Mem0 plugin switched the memory slot globally, disabling memory-core and memory-lancedb for all agents — not just the one I intended.

Current workaround

Using the Mem0 Python SDK directly from the agent's workspace (mem0-service.py) while keeping the global slot on memory-core. This works but misses the plugin lifecycle benefits (auto-recall, auto-capture, tool registration).

Proposed solution

Allow per-agent plugin slot overrides in agent config, e.g.:

{
  agents: {
    "mem0claw": {
      plugins: {
        slots: {
          memory: "openclaw-mem0"
        }
      }
    }
  }
}

Agents without overrides would inherit the global plugins.slots config.

Impact

This would enable mixed-memory (or mixed-anything) setups across agents without workarounds, which is increasingly relevant as the plugin ecosystem grows and agents specialize.