# The AI Engineer Get $1M+ in software deals, the exact agent skills I use to build software, and one practical AI email each week. This is one page of public article previews, not the complete archive. Follow Next page to continue. Summaries are not the original full articles. ## Design and Planning Are the Key Skills for Building Software with AI Coding Agents DevFeed: [Design and Planning Are the Key Skills for Building Software with AI Coding Agents]() Original publisher: [Read original article]() Author: Owain Lewis Published: 2026-08-14T12:12:26Z Content type: article Language: en Sources: [The AI Engineer]() Topics: [AI-assisted coding](), [AI Development](), [Software Engineering](), [Development]() Tags: [agents](), [ai](), [ai-agents](), [ai-coding](), [coding-agents](), [design](), [development-process](), [requirements](), [verification]() ### AI overview The article argues that AI coding agents can handle much of software implementation, testing, and deployment. It says engineers should focus most on design and planning: defining requirements, making technical decisions, and breaking work into tasks agents can execute reliably. ### Source excerpt How to design better software and turn it into tasks AI agents can execute reliably. ## Build your own coding agent DevFeed: [Build your own coding agent]() Original publisher: [Read original article]() Author: Owain Lewis Published: 2026-08-05T20:06:19Z Content type: tutorial Language: en Sources: [The AI Engineer]() Topics: [Agent Harness](), [coding](), [AI-assisted coding](), [AI Development](), [Code](), [Go Language](), [Shell](), [Language models](), [Open Source](), [codex](), [Claude Code]() Tags: [agent](), [agent-harness](), [agents](), [ai](), [ai-development](), [claude](), [claude-code](), [code](), [codex](), [coding](), [commands](), [development](), [go](), [open-source]() ### AI overview This tutorial explains the basic architecture of coding agents and implements a minimal Go-based agent called Micro Neo. The agent connects to an existing AI model and supports reading files, editing files, and running shell commands. It also describes the harness loop that sends tasks and tool results between the user, model, and local machine. ### Source excerpt Why I built my own coding agent and lessons learned ## Why the author replaced Hermes Agent and OpenClaw with a simpler personal AI assistant DevFeed: [Why the author replaced Hermes Agent and OpenClaw with a simpler personal AI assistant]() Original publisher: [Read original article]() Author: Owain Lewis Published: 2026-07-18T10:16:21Z Content type: opinion Language: en Sources: [The AI Engineer]() Topics: [Artificial Intelligence](), [OpenClaw](), [Messaging](), [coding]() Tags: [agent](), [ai](), [coding-agents](), [messaging](), [openclaw](), [personal]() ### AI overview The author explains why Hermes Agent and OpenClaw became frustrating to use as a daily AI personal assistant, focusing on their complexity and overhyped features. The article describes the personal assistant pattern, messaging gateways, and the author's simpler solution built around organisational tasks and existing coding agents. ### Source excerpt How I built my own personal AI Chief of Staff without all the complexity. ## Using Herdr to coordinate multiple coding agents across projects DevFeed: [Using Herdr to coordinate multiple coding agents across projects]() Original publisher: [Read original article]() Author: Owain Lewis Published: 2026-07-10T18:25:31Z Content type: tutorial Language: en Sources: [The AI Engineer]() Topics: [AI-assisted coding](), [Terminal](), [Command-line interface](), [Claude Code](), [codex]() Tags: [agents](), [ai](), [claude-code](), [codex](), [coding-agents](), [how-to](), [terminal]() ### AI overview The article describes using Herdr, a terminal multiplexer, to organize coding agents across projects with workspaces, panes, and tabs. It presents an orchestrator-worker pattern in which a primary agent starts and manages worker agents, including Claude Code, Codex, Pi, and OpenCode, for tasks such as implementing GitHub issues. ### Source excerpt Here's how I use Herdr to manage agents across projects without losing track of what each one is doing. ## We are all AI agent managers now DevFeed: [We are all AI agent managers now]() Original publisher: [Read original article]() Author: Owain Lewis Published: 2026-07-03T11:08:20Z Content type: opinion Language: en Sources: [The AI Engineer]() Topics: [AI Agent](), [systems]() Tags: [agent](), [ai-agent](), [ai-agents](), [engineering](), [systems]() ### AI overview An engineering manager reflects on how working with AI agents changes the role of developers. The article argues that the work increasingly involves defining goals, quality standards, constraints, and systems that enable agents to produce results, while retaining human direction because agents make frequent mistakes and lack human intelligence and creativity. ### Source excerpt The work is no longer writing the code. The work is designing and scaling agent systems. ## Examples of Agent Goals, Loops, and Scheduled Workflows in Codex and Claude Code DevFeed: [Examples of Agent Goals, Loops, and Scheduled Workflows in Codex and Claude Code]() Original publisher: [Read original article]() Author: Owain Lewis Published: 2026-07-02T07:01:00Z Content type: article Language: en Sources: [The AI Engineer]() Topics: [codex](), [Loop Engineering](), [Automation](), [Claude Code](), [Pull Request](), [GitHub Issues](), [ci]() Tags: [agent](), [ai-agents](), [automation](), [ci](), [claude-code](), [codex](), [github-issues](), [logs](), [review](), [scheduled](), [tests](), [workflow]() ### AI overview The article explains how prompts, goals, recurring runs, and workflows differ, then presents examples of using agents for recurring engineering tasks. It emphasizes evidence, stop rules, tests, CI, logs, issues, pull requests, and human review as safeguards because agents can misunderstand context or make mistakes. ### Source excerpt Practical ways to combine prompts, goals, and scheduled runs in Codex and Claude Code. ## I tested nested subagents in Claude Code DevFeed: [I tested nested subagents in Claude Code]() Original publisher: [Read original article]() Author: Owain Lewis Published: 2026-06-17T14:03:13Z Content type: opinion Language: en Sources: [The AI Engineer]() Topics: [Claude Code](), [context window](), [Code](), [GitHub](), [Go](), [Pull Request](), [Code review]() Tags: [claude-code](), [coding](), [go](), [pipeline](), [pull-request](), [recursion](), [review](), [workflow]() ### AI overview The article tests nested subagents in Claude Code, including recursion depth and a three-level workflow of orchestrator, workers, and sub-workers. In a Go project, the workflow implemented a change, ran checks and tests, reviewed the result, fixed minor issues, and opened a pull request. ### Source excerpt Claude Code now let's you run nested subagents. But is it useful? ## How I'm building autonomous agent loops DevFeed: [How I'm building autonomous agent loops]() Original publisher: [Read original article]() Author: Owain Lewis Published: 2026-06-13T14:02:17Z Content type: tutorial Language: en Sources: [The AI Engineer]() Topics: [Loop Engineering](), [Artificial Intelligence](), [GitHub Issues](), [Claude Code](), [codex](), [GitHub]() Tags: [agent](), [ai](), [autonomous](), [claude-code](), [codex](), [github-issues](), [loops](), [workflow]() ### AI overview This guide explains how to build autonomous agent loops in which agents read project state, perform work, update that state, and repeat on a schedule or in response to events. It presents GitHub Issues as a control plane and describes manager and worker loops built with Claude Code and Codex, including risk classification, ticket management, and safeguards against destructive actions. ### Source excerpt A guide on loop engineering and autonomous AI loops ## How Codex coordinates parallel coding tasks through a coordinator thread DevFeed: [How Codex coordinates parallel coding tasks through a coordinator thread]() Original publisher: [Read original article]() Author: Owain Lewis Published: 2026-06-06T12:54:48Z Content type: article Language: en Sources: [The AI Engineer]() Topics: [codex](), [ai-coding](), [Git](), [Pull Request](), [NEO]() Tags: [ai-agents](), [ai-coding](), [also](), [codex](), [git]() ### AI overview The article describes using a Codex coordinator thread to create and manage multiple worker threads for parallel coding tasks. Each worker uses a separate Git branch and worktree, follows an implementation and testing workflow, receives review feedback, and opens a draft pull request. ### Source excerpt How I let OpenAI Codex run a team of AI agents from a single Codex coordinator thread. ## I let AI agents run my entire GitHub workflow DevFeed: [I let AI agents run my entire GitHub workflow]() Original publisher: [Read original article]() Author: Owain Lewis Published: 2026-05-29T08:34:57Z Content type: tutorial Language: en Sources: [The AI Engineer]() Topics: [GitHub](), [AI-assisted coding](), [Command-line interface](), [Claude Code]() Tags: [ai-coding-agents](), [cli](), [github]() ### AI overview A tutorial describing the author's workflow for using AI coding agents with the GitHub CLI to manage a Go command-line project, including repositories, tickets, reviews, merges, and a release pipeline. ### Source excerpt Give an agent the GitHub CLI and it can do far more than write code. ## The 7 skills I actually use every day with AI coding agents DevFeed: [The 7 skills I actually use every day with AI coding agents]() Original publisher: [Read original article]() Author: Owain Lewis Published: 2026-05-15T11:24:18Z Content type: opinion Language: en Sources: [The AI Engineer]() Topics: [AI-assisted coding](), [Loop Engineering](), [Command-line interface](), [codex]() Tags: [agents](), [ai-coding-agents](), [cli](), [codex](), [coding-agents](), [skills](), [workflow]() ### AI overview An opinion article argues that productive work with AI coding agents does not require large collections of skills, MCP servers, plugins, or complex configuration. It presents a small set of preferred practices, including clarifying requirements before coding and breaking work into reviewable tasks. ### Source excerpt My favourite Codex skills for building software. ## Pi versus Claude Code: An Honest Comparison of a Minimalist AI Coding Agent DevFeed: [Pi versus Claude Code: An Honest Comparison of a Minimalist AI Coding Agent]() Original publisher: [Read original article]() Author: Owain Lewis Published: 2026-04-18T17:00:15Z Content type: comparison Language: en Sources: [The AI Engineer]() Topics: [AI-assisted coding](), [Open Source](), [Claude Code](), [codex](), [TypeScript]() Tags: [bash](), [claude-code](), [code-reviews](), [coding-agents](), [extension](), [open-source](), [reviews](), [tests]() ### AI overview The article compares Pi, an open-source minimalist coding agent, with Claude Code and Codex. It examines Pi's four built-in tools, broad model support, editable system prompt, and TypeScript extensions, including a workflow for writing, reviewing, fixing, testing, and verifying code. It also discusses trade-offs such as the lack of built-in MCP, to-do tracking, sub-agents, and hooks. ### Source excerpt I spent a week with the minimalist AI coding agent Pi. Here's my honest take. ## Six RAG strategies, explained simply (with code). DevFeed: [Six RAG strategies, explained simply (with code).]() Original publisher: [Read original article]() Author: Owain Lewis Published: 2026-04-03T08:36:55Z Content type: tutorial Language: en Sources: [The AI Engineer]() Topics: [Retrieval Augmented Generation (RAG)](), [Database](), [Large Language Model]() Tags: [database](), [embeddings](), [full-text-search](), [rag](), [retrieval](), [tutorials](), [vector-search]() ### AI overview This tutorial explains retrieval-augmented generation and compares retrieval strategies, including loading complete documents, full-text search, and vector search. It presents Postgres as a practical option for implementing these approaches and discusses their trade-offs. ### Source excerpt Six Ways to Retrieve Data for an LLM ## A Four-Layer Workflow for Reviewing AI-Generated Code DevFeed: [A Four-Layer Workflow for Reviewing AI-Generated Code]() Original publisher: [Read original article]() Author: Owain Lewis Published: 2026-03-27T17:33:31Z Content type: tutorial Language: en Sources: [The AI Engineer]() Topics: [AI-assisted coding](), [Code review](), [Claude Code](), [coding](), [Pull Request](), [Shell]() Tags: [ai](), [ai-agents](), [ai-code-review](), [claude-code](), [code-review](), [coding](), [eslint](), [security](), [tests]() ### AI overview The article presents a four-layer workflow for reviewing AI-generated code: automated checks, local AI review, CI review on pull requests, and human review. It emphasizes using hooks in Claude Code to run linters and security scanners, executing the code, reading the diff, and using a fresh context for a second AI opinion. ### Source excerpt How to write better code when using AI agents ## How I delegate work to a team of AI agents DevFeed: [How I delegate work to a team of AI agents]() Original publisher: [Read original article]() Author: Owain Lewis Published: 2026-03-19T14:42:29Z Content type: tutorial Language: en Sources: [The AI Engineer]() Topics: [AI Agent](), [ai-coding](), [TypeScript](), [Claude Code](), [Pull Request](), [bug](), [Command-line interface]() Tags: [ai-agent](), [ai-coding](), [ai-coding-agents](), [architecture](), [bug](), [claude-code](), [latency](), [pull-request](), [typescript]() ### AI overview The article describes a proof-of-concept TypeScript worker that polls a task management system, delegates smaller tasks to headless Claude Code, runs checks, and opens pull requests. It presents this as an agent control plane and explains why a pull-based design can reduce internet exposure, with latency as the trade-off. ### Source excerpt Building systems for delegating work to AI agents ## Seven stages of building software with AI across the development lifecycle DevFeed: [Seven stages of building software with AI across the development lifecycle]() Original publisher: [Read original article]() Author: Owain Lewis Published: 2026-03-06T13:15:47Z Content type: article Language: en Sources: [The AI Engineer]() Topics: [Artificial Intelligence](), [AI-assisted coding](), [Software](), [Code]() Tags: [ai](), [ai-agents](), [ai-coding](), [code](), [evals](), [technical](), [workflow]() ### AI overview This article presents a seven-stage software lifecycle--requirements, design, task breakdown, building, review, deployment, and monitoring--and explains how AI can support each stage, not only code generation. It emphasizes planning and documented requirements, describing how they helped guide an AI agent and generate evaluations on a client project. ### Source excerpt Real prompts for planning, building, reviewing, and shipping software with AI agents ## Using OpenAI Codex Automations to Scan for Bugs and Fix Code on a Schedule DevFeed: [Using OpenAI Codex Automations to Scan for Bugs and Fix Code on a Schedule]() Original publisher: [Read original article]() Author: Owain Lewis Published: 2026-02-21T15:17:09Z Content type: tutorial Language: en Sources: [The AI Engineer]() Topics: [codex](), [Automation](), [ai-coding](), [coding](), [bug]() Tags: [ai](), [ai-agents](), [ai-coding](), [automation](), [bug](), [codex](), [coding](), [config](), [how-to](), [openai]() ### AI overview The article explains how to configure OpenAI Codex Automations to run coding tasks on a schedule. It describes a paired workflow in which one automation scans for bugs and creates Linear tickets, while another fixes the tickets and opens pull requests, using persistent memory to avoid duplicate reports and track prior work. ### Source excerpt How to create AI agents that work for you 24/7. ## Claude Code agent teams explained DevFeed: [Claude Code agent teams explained]() Original publisher: [Read original article]() Author: Owain Lewis Published: 2026-02-12T17:25:21Z Content type: article Language: en Sources: [The AI Engineer]() Topics: [Claude Code](), [Artificial Intelligence](), [Code]() Tags: [agent](), [agents](), [c](), [claude-code](), [compiler](), [rust](), [teams](), [terminal]() ### AI overview The article explains Claude Code's Agent Teams feature, in which independent Claude Code sessions have separate context windows, can communicate directly, and coordinate through a shared task list. It contrasts this with subagents and describes agent-to-agent review loops for longer-running work. ### Source excerpt Surprising lessons from building an app with a team of agents ## How to build an AI agent workflow with Linear, MCP, and CLAUDE.md DevFeed: [How to build an AI agent workflow with Linear, MCP, and CLAUDE.md]() Original publisher: [Read original article]() Author: Owain Lewis Published: 2026-02-07T16:36:17Z Content type: tutorial Language: en Sources: [The AI Engineer]() Topics: [Claude Code](), [Model Context Protocol (MCP)](), [Code](), [Pull Request](), [Terminal]() Tags: [agent](), [agents](), [claude-code](), [code](), [context](), [developer](), [how-to](), [linear](), [markdown](), [mcp](), [project](), [review](), [terminal](), [workflow]() ### AI overview This tutorial explains how to manage AI coding tasks through a Linear board connected to Claude Code with MCP. A CLAUDE.md file defines the workflow, enabling the agent to read tickets, implement changes, run builds, review code, and open pull requests. The approach depends on clear tickets and provides status visibility as the number of tasks grows. ### Source excerpt How to build your agent control plane ## How I code with AI agents (spec-driven development) DevFeed: [How I code with AI agents (spec-driven development)]() Original publisher: [Read original article]() Author: Owain Lewis Published: 2026-01-30T12:24:42Z Content type: tutorial Language: en Sources: [The AI Engineer]() Topics: [Spec Driven Development](), [Development](), [Code](), [AI Agent](), [Claude Code]() Tags: [agents](), [claude-code](), [code](), [guide](), [spec-driven-development]() ### AI overview An opinionated guide to spec-driven development with AI coding agents. It explains how to define a reviewed, version-controlled specification before execution, distinguishes specs from PRDs and design documents, and outlines when each document is useful. ### Source excerpt An opinionated guide to writing code with AI agents like Claude Code.