operate-hermes-agent-gatewaylisted

# Operate Hermes Agent Gateway Treat the messaging gateway as a long-running service with external identities, inbound events, persistent sessions, and delivery obligations. ## Define the Deployment 1. Name the platform or API surface. 2. Name the Hermes profile that owns its config, credentials, memory, sessions, skills, and gateway state. 3. Choose the host and terminal backend. 4. Define authorized users, chats, workspaces, or webhook senders before enabling actions. 5. Define delivery behavior for replies, media, scheduled work, and failures. 6. Define process supervision and logs through Hermes-supported or OS-supported service paths; do not invent a second gateway manager. ## Keep Gateway Surfaces Distinct - Messaging gateway: `hermes gateway` and its platform adapters. - API server: OpenAI-compatible HTTP endpoint hosted by the gateway. - Inbound webhooks: authenticated event routes that trigger agent work. - Outbound webhooks: signed delivery of Hermes events to a configured receiver; they are not inbound prompt endpoints. - A2A platform: peer-agent discovery and tasks; hand its protocol lifecycle to `operate-a2a-agent-integration`. - TUI gateway: JSON-RPC host protocol, not the messaging service. - Nous Tool Gateway: hosted tool backends, not message transport. ## Configure Safely - Use a dedicated profile when a bot identity or authorization boundary differs. - Keep platform tokens and API server keys in private configuration. - Use separate secrets for i