How (and what) I'm building this week

2025-04-24 · Source: Ben's Bites · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Software Development & Engineering, Robotics & Autonomous Systems · Depth: Intermediate, medium

Summary

This week's builder's log details the development of an interactive cookbook for AI agents, designed to teach users how to build and deploy websites. The author also released a visualization skill for agents, reverse-engineered from Claude's new features, which has garnered over 200 GitHub stars. Additionally, a redesign of the Ben's Bites Cookbook site is underway, aiming for a cleaner, more efficient platform for helpful documentation. The author shares their current tech stack, including models like GPT 5.4 XHigh and Opus 4.6, CLIs such as Droid and Pi, and agent apps like Codex and Claude Code. Key skills like "frontend-design" and "agent-browser" are highlighted, alongside other tools like exe for virtual servers and here.now for site deployment. The post concludes with a detailed look at the author's `AGENTS.md` file, outlining instructions for agent behavior during development.

Key takeaway

For AI Engineers and developers building with agents, consider adopting an `AGENTS.md` file to standardize your agent's workflow, ensuring consistent practices like creating spec folders and running automated tests. This approach can significantly streamline your development process and improve code quality by guiding agents to follow best practices and reference documentation, ultimately accelerating project delivery and reducing manual oversight.

Key insights

Interactive, agent-driven learning and development workflows enhance building and deployment efficiency.

Principles

• Iterative development improves complex tools.
• Agent-based learning can be highly effective.
• Visualizations aid code system comprehension.

Method

An interactive cookbook guides agents to teach users web development by building and deploying a site, while a custom `AGENTS.md` file standardizes agent behavior for structured development.

In practice

• Use `AGENTS.md` for consistent agent instructions.
• Integrate visualization skills for code learning.
• Employ `agent-browser` for automated site testing.

Topics

• AI Agents
• Generative AI Models
• AI Development Tools
• Automated Web Development
• AI Agent Skills

Code references

• bentossell/visualise
• millionco/react-doctor
• iannuttall/dotagents

Best for: AI Engineer, Software Engineer, Prompt Engineer

Related on AIssential

• AGENTS.md, a README.md for agents. — AGENTS.md and agent skills provide structured, open formats for guiding AI coding agents in software development.
• Agent Skills Work but the Research Shows Most Teams Are Building Them Wrong — Curated, focused agent skills improve AI performance, but require structured organization and security diligence to scale effectively.
• Day 2 - Agentic Engineering/VIbe Coding an AI App from Zero To Production | 🔴 Live — AI agents can rapidly generate functional prototypes and manage project workflows from high-level requirements.
• addyosmani / agent-skills — AI coding agents can achieve production-grade software quality by following structured, verifiable engineering workflows and best practices.
• Building AI Agents That Actually Work: Lessons from Jason Lemkin, Jeanne DeWitt Grosser (Vercel), Amelia Lerutte & Amjad Masad (Replit) — Effective AI agents require robust data foundations, iterative development, and a strategic "buy vs. build" approach.
• My AI Coding Workflow — The author's AI coding workflow leverages a "Guides, Gates, and Guards" model, shifting to Codex for efficiency and cost reduction.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Ben's Bites.