running
max-mini-hermes
@max-mini-hermes
Modelgpt-5.5
ProviderOpenAI Codex
ChannelTelegram + Matrix
Agent Fleet Dashboard
Operational snapshot for Matrix, Telegram, Hermes, Claw, and host capacity.
Top Rules
Main thread
max-mini-hermes is the default human-facing router. Long work launches as worker sessions with --source tool and --worktree.
VPS ops
max-vps-hermes is first responder in Contabo/VPS ops rooms. Other agents wait unless mentioned or the VPS lane is blocked.
Project rooms
Shared project rooms stay mention-only. Agent Ops rooms are free-response for that one agent.
Agents
running
max-mini-claw
@max-mini-claw
RuntimeOpenClaw
Gatewayrunning
Configcanonical slug
running
lera-mini-hermes
@lera-mini-hermes
Modelgpt-5.5
ProviderOpenAI Codex
Configcanonical slug
running
lera-mini-claw
@lera-mini-claw
RuntimeOpenClaw
Gatewayrunning
Configcanonical slug
running
max-vps-hermes
@max-vps-hermes
ProviderMiniMax in Matrix
RoleVPS first responder
Configcanonical slug
standby
max-macpro-hermes
@max-macpro-hermes
Modelgpt-5.5
Gatewaystopped
Rolelocal high-context controller
Capacity
| Host | Available | Slots | Pressure | Load |
|---|---|---|---|---|
max-mini | 3295 MB | 2 | ok | 7.65 |
lera-mini | 7195 MB | 2 | ok | 1.34 |
max-vps | 5698 MB | 1 | ok | 4.18 |
Migration Checklist
- Canonical Matrix users created, tokened, and joined to required rooms.
- Runtime configs switched to canonical IDs and allowlists.
- Affected gateways restarted and reported running.
- Own Ops room Matrix send smokes passed for all canonical lanes.
- Stale
maxiclaw,uniclaw,contabo,@hermes, and*-openclawidentities deactivated with history preserved.