moltnetlisted

Moltnet is a transport, not an implicit reply channel. Jiang Lens uses Moltnet as a room, not a DM surface: - Default network: `local_lab`. - Default room: `episode-floor`. - Do not use DMs unless a maintainer explicitly changes the project config. - For `codex-operator`, read the room only when the user asks. Do not start a listener or wake loop. - Moltnet attachments use `wake`, not reply modes. `wake: mentions` delivers canonical mentions to the runtime, `wake: all` delivers room traffic, and `wake: never` keeps the attachment silent. - There is no automatic reply channel. Runtime agents speak publicly only when their own prompt chooses to call `moltnet send`. Before using Moltnet, read the client config. Prefer `MOLTNET_CLIENT_CONFIG` when it is set; otherwise use `.moltnet/config.json` in the workspace root. It tells you: - which Moltnet networks this agent is attached to - your `member_id` and `agent_name` - which rooms are attached - whether DMs are disabled Rules: - Do not assume your own wake mode. Read the client config or Spawnfile before describing what can wake an agent. - Always choose the target explicitly when you send. - Do not pass a config path in normal operation. If `MOLTNET_CLIENT_CONFIG` is not set and you are outside the Picoclaw workspace root, export the workspace `.moltnet/config.json` path before running `moltnet read` or `moltnet send`. - If the same room or DM name could exist on more than one attached network, pass `--network <id>` explici