Self-improving agent architecture

Self-Improving Agent Systems: Weekly Trio Reports and Automated Audits

What happens when your AI agents review their own work every week? They find things you’d never catch manually — duplicated code, exposed credentials, and missed optimization opportunities. The Weekly Trio Every Sunday at 03:00, three specialized agents wake up and run an autonomous improvement cycle: Cody (Code Architect) — Scans for technical debt, duplicated logic, and missing tests Max (Red Team Lead) — Audits for credential exposure, permission violations, and security drift Molty (Research Lead) — Identifies workflow gaps and builds missing tooling This isn’t a demo. This is production. Here’s what happened this week. ...

OAuth and voice integration illustration

OAuth Gateway Migration and Voice Responses: Making AI Assistants More Natural

OAuth tokens expire. Every week, I’d manually re-authorize Google Workspace integrations. It worked, but it wasn’t sustainable for an autonomous system. This week, I finally fixed that — and added voice responses along the way. The OAuth Problem My AI assistant (Clawd) pulls calendar events for daily briefs. The old GOG skill required manual re-auth every 7 days. That’s: Click authorization link Sign in to Google Grant permissions Copy token back to config Not exactly “set it and forget it.” ...

AI and technology illustration

OpenClaw: Building a Personal AI Assistant the Lobster Way

After years of cobbling together scripts, cron jobs, and manual SSH sessions, I found OpenClaw — a framework for building personal AI assistants that actually understand your infrastructure. Here’s how I turned it into Clawd, my homelab orchestrator. What is OpenClaw? OpenClaw is an open-source framework for running personal AI assistants. Unlike cloud-based options (ChatGPT, Claude, Gemini), OpenClaw runs on your hardware with your choice of LLM backend. Key features: ...

Server room and automation illustration

Automating with Clawd: Building an AI Assistant for Homelab

Managing a homelab is a constant battle between “set it and forget it” and “everything is on fire.” After years of manual script maintenance, I decided to build Clawd — a hierarchical AI assistant that orchestrates my entire infrastructure. The Problem Every homelab owner knows the drill: SSH into multiple machines to check status Remember which scripts need to run when Monitor services across Proxmox VMs, Docker containers, and NAS Keep track of updates, backups, and security alerts My solution? Build an AI that understands my infrastructure and can execute tasks on my behalf. ...

Server rack and homelab infrastructure

Homelab Audit Q1 2026: Security Hardening & Hardware Reality

Maintaining a homelab is a constant cycle of automation and observation. This week, during a deep audit with Clawd, we pushed several significant updates to the core infrastructure to improve reliability and security. 1. Platform Hardening & Signal Migration With the release of OpenClaw v2026.2.24, delivery policies for heartbeats changed. To comply with the new security standard (blocking automated pushes in DMs), we migrated all system alerts and heartbeats to a dedicated Signal Alert Group. This ensures our 1-on-1 chat stays focused on high-level decision-making while technical logs stream to the audit group. ...

Bash Terminal

Bash Scripting: A High-Tier Syntax Guide

Master the Shell Bash isn’t just a shell; it’s the glue that holds the modern homelab together. Whether you’re automating Nextcloud syncs or managing Proxmox VMs, knowing your way around a .sh file is essential. 1. The Shebang Always start your scripts with a proper shebang. #!/bin/bash is standard, but #!/usr/bin/env bash is more portable across different Linux distributions. #!/usr/bin/env bash 2. Variables Variables are case-sensitive and should not have spaces around the =. ...