How-Tos & Practical Guides

AIO Sandbox from Agent-Infra packages everything an AI agent needs to operate — browser, shell, filesystem, MCP server, VSCode, and Jupyter — into a single Docker container. Here’s how to get it running in under 5 minutes. Prerequisites Docker installed and running (get Docker) Port 8080 available on your machine ~2GB free disk space for the container image Step 1: Pull and Run the Container docker run --security-opt seccomp=unconfined --rm -it -p 8080:8080 ghcr.io/agent-infra/sandbox:latest The --security-opt seccomp=unconfined flag is required for browser automation to work inside the container. The first run will pull the image (~1-2GB), subsequent starts are fast. ...

If you want to understand how a complete agentic AI system actually fits together — not from a marketing diagram, but from working Python code — nanobot is one of the best educational repositories available right now. Built by HKUDS and actively maintained (last commit March 2026), it’s an ultralight OpenClaw-inspired personal agent framework that clocks in at roughly 4,000 lines of Python. No heavy dependencies, no framework magic — just the core subsystems laid bare. ...

Most tutorials about AI agents end with something that produces output. This one is about something different: an agent that produces income. Developer Eliott Reich documented how they built an AI agent that earns real money — not through speculation, not through selling the agent itself, but through autonomous task completion that generates actual revenue. Here’s a breakdown of how the system works and how you can build one. The Core Concept: Agents as Economic Actors The insight behind a money-earning agent is simple but consequential: if an agent can complete tasks that have economic value, and if those tasks can be reliably discovered and delivered, then the agent earns money as a byproduct of working. ...

Claude Code’s Auto Mode is one of the most practically useful features Anthropic has shipped for autonomous development workflows — and one of the least understood. This guide explains exactly what Auto Mode does, how its safety classifier works, when to use it versus manual mode, and what configuration patterns will keep your codebase intact. What Is Claude Code Auto Mode? Auto Mode is a Team-tier feature that gives Claude Code permission to auto-approve certain actions without prompting you for confirmation. That might sound alarming if you’ve worked with AI agents before — but the key is that “certain actions” is a carefully bounded category, enforced by a separate Sonnet 4.6 classifier model that runs before each action is executed. ...

Zhipu AI released GLM-5.1 on March 27, 2026, and the benchmark numbers are legitimately surprising. On Claude Code’s own coding evaluation, GLM-5.1 scores 45.3 — that’s 94.6% of Claude Opus 4.6’s 47.9. On SWE-bench-Verified, it hits 77.8 (open-source state of the art). On Terminal Bench 2.0, it posts 56.2. And it’s available via OpenRouter at a fraction of Opus pricing. This guide walks you through connecting GLM-5.1 to OpenClaw via OpenRouter and configuring it intelligently for coding-heavy agent workloads. ...

Your LangGraph agent works perfectly in development. Then it hits production and you discover the problem every agent developer eventually hits: when the process restarts, your agent remembers nothing. In-memory state is fine for demos and local testing. For production agents — especially those handling multi-step workflows that can span hours, serve concurrent users, or need to resume after infrastructure failures — you need persistent state. This guide walks through adding Aerospike Database 8 as a durable memory store for your LangGraph agent. ...

If your Claude Code usage limits are draining faster than you expect, you’re not imagining it and you’re not hitting a bug. Anthropic confirmed this week that usage consumed during peak hours counts at an accelerated rate against your monthly limit. The peak window: 5:00 AM to 11:00 AM Pacific Time, Monday through Friday. This guide covers what that means for your usage, how to track where your limit is going, and the practical strategies that actually help. ...

Irish AI startup Jentic just launched Jentic Mini — a free, open-source, self-hosted API execution firewall specifically designed to sit between your OpenClaw agents and the external APIs they call. It handles credentials, permissions, and access control so your agents don’t have to. If you’re running OpenClaw agents that interact with external services — and especially given the recent GhostClaw malware campaign targeting AI agent skill systems — adding an execution firewall layer is no longer optional. It’s operational security. ...

The Silverfort researchers who disclosed the ClawHub ranking-manipulation vulnerability found that attackers could push a malicious skill to the #1 spot in a category using nothing more than unauthenticated HTTP requests to inflate download counts. Snyk’s ToxicSkills study independently identified 1,467 vulnerable or malicious skills across the registry. If you use ClawHub skills in your OpenClaw deployment — especially if you have auto-install or auto-upgrade enabled — this guide will walk you through a complete audit. ...

Figma just made a significant move: the design canvas is now open to AI coding agents via a native MCP (Model Context Protocol) server. As of this week, agents like Claude Code, Cursor, VS Code Copilot, Codex, and Warp can read your Figma files, understand the design structure, and generate code that maps directly to your actual components — not a screenshot approximation, but the live design graph. This is currently in free beta. Here’s how to get connected. ...