Cookin' with Gas
How one engineer runs a 24/7 multi-agent AI stack on bare metal. Opinionated. Dogfooded. Broken-and-fixed in production.
58 production-tested recipes for running long-lived AI agents in tandem with daily coding harnesses. Not a framework. A record of what is actually deployed, what broke, and how it was fixed.
Cron Patterns
The three-layer cron stack I actually run: systemd timers for OS plumbing, OpenClaw cron for single-shot agent tasks, n8n schedule triggers for multi-step workflows. Pick the wrong layer and you'll…
I · AI agent stackMulti-Model Orchestration
How to run multiple AI models in one OpenClaw setup, assign each to the right task tier, and stop burning expensive tokens on work that doesn't need them.
Updated 2026-06-05 IV · SecuritySecurity Hardening: Linux Host for OpenClaw
Practical hardening runbook for an Ubuntu 24.04 machine running OpenClaw as an always-on AI agent. This covers firewall configuration, SSH lockdown, fail2ban, and service binding to reduce attack…
Updated 2026-04-19 III · Self-hosted infrastructureBackup & Recovery
How to protect your OpenClaw workspace, configuration, and memory from data loss. Encrypted backups, restore testing, and disaster recovery planning.
Updated 2026-05-31AI agent stack
Multi-model orchestration. One orchestrator, many models, escalation lanes. Most of these guides assume OpenClaw as the agent framework, but the patterns generalize.
13 recipes Chapter IIAutomation
Anything scheduled, hooked, or sandbox-wrapped. The decision tree for "where does this go" lives in cron-patterns.md - start there.
7 recipes Chapter IIISelf-hosted infrastructure
The homelab. Hypervisor decisions, container layout, network storage, off-host backup, integration with daily-driver desktop, and surviving upgrades.
7 recipes Chapter IVSecurity
Defense in depth across host, agents, network, and outbound boundary. Plus what to do when the agent does something it shouldn't.
6 recipes Chapter VKnowledge management
How durable knowledge flows: cards, indexes, handoffs, sync. The goal is "future-me can pick up cold."
6 recipes Chapter VIHardware & host
The physical layer. Choosing the box, partitioning the disk, deciding what the host OS owns vs what gets virtualized.
3 recipes Chapter VIITools
Index of tools shipped from this stack - MCPs, dashboards, helpers - and how they fit together.
5 recipes Chapter VIIIPublishing
Reserved for docs about sanitizing and validating artifacts before they leave the host.
1 recipe Chapter IXPhilosophy
The opinions. Why this stack is shaped the way it is. What I will not do. Why this body of work is a cookbook and not a tool.
4 recipes Chapter XPlans
1 recipe Chapter XISkills
Public, sanitized skills pulled from a real always-on agent stack.
5 recipes