#AI Agent
1114 related articles
MCP Server Deep Dive: Transforming AI from Assistant to Autonomous DevOps Agent
MCP (Model Context Protocol) is Anthropic's open standard dubbed the 'USB-C port for AI.' This deep dive covers its 3-layer architecture, Resource/Tools/Prompts primitives, real-world DevOps incident response, and security best practices.
Pine Computer: A Cloud Computer Built for AI Agents to Actually Get Things Done
Pine Computer is a cloud computer built for AI agents — it reads app structure (not pixels) to handle research, form-filling, and no-API portals. Multi-tenant, bring-your-own-model, with human takeover support.
Opposable: Letting AI Control Real Phones — Giving Claude a "Thumb"
Opposable lets Claude Code, Codex, and MCP agents see, tap, and type on real iPhones and Androids. Common tasks need no model; payments require human confirmation.
Melete: An Always-On Personal AI Agent With Its Own Computer
Melete is an always-on personal AI agent with its own computer, persistent memory, and autonomous execution — keeping tasks moving even when you're offline.
MCP Is Seriously Underrated: The Hidden Value of How AI Agents Connect to Tools
MCP is being re-evaluated. This article explains why early Token cost criticisms are outdated, and why MCP beats CLI on workflow reliability and key security for AI Agents.
Drawing Arrows on Screen: A New Way to Visualize AI Agent Actions
A Hacker News-featured project lets AI agents draw arrows, boxes, and labels on screen. Explore how on-screen annotations improve agent observability and human-AI collaboration.
OpenAI Agents SDK Guardrails Explained: Input, Output, and Tripwire
A deep dive into OpenAI Agents SDK Guardrails: how input and output guardrails work, their execution order, and how the Tripwire mechanism triggers interception to build safe, cost-efficient AI Agents.
Building Your First AI Agent: Only 4 Core Components Required
Build your first AI Agent with just 4 core components: model, instructions, tools, and state — plus a run loop. Based on OpenAI's official guide.
Getting Started with LangChain: Build Your First AI Agent in Just a Few Lines with create_agent()
LangChain tutorial: compare bare OpenAI SDK vs. create_agent() with a weather example. Learn to build your first AI Agent and master hidden agent loop mechanics.
16 Leading AI Agents Explained: What They Actually Do
A breakdown of 16 leading AI agents — ChatGPT Work, Claude, Manus, Codex, Devin, Cursor, GitHub Copilot and more — covering their design, real capabilities, and key differences across four categories.
Laravel AI SDK v1.0 in Practice: Building an Intelligent Agent with Tools, Memory, and MySQL
Step-by-step Laravel AI SDK v1.0 tutorial: build an intelligent agent with tool calling, conversation memory, and MySQL for an e-commerce customer service bot.
OpenAI Agents SDK in Practice: Implementing Human-in-the-Loop Approval
A complete guide to implementing Human-in-the-Loop approval with OpenAI Agents SDK: pause tool calls with needs_approval, capture interruptions, handle approve/reject decisions, and serialize RunState for async workflows.
Hermes Agent Deep Configuration Guide: The Open-Source AI Agent for Everyone
Hermes Agent by Nous Research is an open-source AI agent with a desktop app, Telegram integration, remote deployment, and multi-user dashboard. Full setup guide inside.
Claude Cowork Hands-On Tutorial: Let AI Take Over Your Daily Computer Tasks
Complete Claude Cowork hands-on tutorial: learn the three Claude product lines, master desktop file operations, Chrome automation, Connectors, reusable Skills, and Dispatch for remote phone control.
Why AI Agents Need Skills: A Modular Approach to Eliminating Context Bloat
Agent Skills solve AI Agent context window bloat through modular, on-demand knowledge packages. Learn how skill.md's Front Matter/Body structure, smart skill selection, and portability reduce token costs and hallucinations.
Before Deploying an AI Agent, You Must Understand These Safety Switches
Three essential safety layers before deploying AI Agents: Permission Mode, the critical difference between tools and allowedTools, and zero-token-cost Hooks for hard interception.
9 Open-Source AI Agent Tools Roundup: From Workflow Orchestration to Skills Distribution
A roundup of 9 open-source AI Agent tools covering workflow orchestration, Skills distribution, model gateways, memory management, and UI generation for production-ready Agents.
Hermes AI Agent for Automated Penetration Testing: A Deep Dive into Skills, Memory, and Cron Jobs
How to turn Hermes AI Agent into an authorized pentest and bug bounty assistant: cloud VPS deployment, Telegram integration, custom Skills, persistent Memory, and Cron Jobs for automated subdomain monitoring.
Why Does Your AI Customer Service Keep Saying 'Please Be Patient'? Because It Never Checks Real-Time Status
Why does AI customer service always say 'please wait'? Because it only queries a knowledge base and never reads real-time status. Learn how financial AI agents should combine policy rules with live transaction data.
Plan-and-Patch: Agentic Planning and Local Repair with Diffusion Language Models
Plan-and-Patch uses diffusion language models for agentic planning and local repair. Diffusion planners hit 53.7% repair success rate — nearly 2x AR models — while cutting latency 39–46%.