🎙️ This week on How I AI: How to build your own AI developer tools with Claude Code

2025-05-13 · Source: Lenny's Newsletter · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Software Development & Engineering, Emerging Technologies & Innovation · Depth: Intermediate, extended

Summary

CJ Hess, a software engineer at 10X, demonstrates his AI-native development tool, Flowy, which he built to enhance his AI engineering workflow, particularly with Claude Code. Flowy addresses the limitations of ASCII flowcharts and markdown-based UI mockups by generating visual diagrams and wireframes from JSON files. Hess utilizes custom skills within Claude Code to enable the AI to understand and interact with Flowy's proprietary JSON schema, allowing for iterative design and development. He showcases a workflow where Claude Code generates animation timing and user flow diagrams, which can then be edited visually in Flowy, with changes reflected back in the JSON for Claude to interpret. Hess also employs model-to-model evaluation using Codeex to review Claude's generated code, identifying discrepancies and suggesting refactoring improvements, highlighting a shift towards AI-managed documentation and rapid prototyping.

Key takeaway

For AI Engineers seeking to optimize their development cycles and improve code quality, consider building custom AI-native tools and integrating model-to-model evaluation. This approach allows for more intuitive visual planning and rapid prototyping, while a secondary AI can act as a "staff engineer" to identify subtle code smells and suggest architectural improvements, significantly reducing manual review time and preventing technical debt from "vibe coding" practices.

Key insights

Custom AI-native dev tools and model-to-model evaluation enhance AI engineering workflows and code quality.

Principles

• AI agents can manage and update living documentation.
• Visual representations improve human-AI communication.
• Model-to-model evaluation catches subtle code discrepancies.

Method

Develop custom AI-native tools with proprietary schemas, create specific AI skills for interaction, and use model-to-model evaluation for code review and refactoring suggestions, enabling rapid, iterative development.

In practice

• Build custom tools for specific workflow gaps.
• Use AI to generate and refine skill documentation.
• Employ a second AI for code review and refactoring.

Topics

• AI Engineering Workflows
• Claude Code
• AI-Native Dev Tools
• Model-to-Model Evaluation
• AI-Powered UI/UX

Best for: AI Engineer, Machine Learning Engineer, MLOps Engineer

Related on AIssential

• v2.1.87 — Claude Code updates significantly enhance AI model capabilities, user experience, and extensibility through plugins and agent teams.
• DIY dev tools: How this engineer created “Flowy” to visualize his plans and accelerate coding — Custom AI-native dev tools enhance planning and coding efficiency by enabling visual iteration and model-to-model evaluation.
• We're sharing our internal AI engineering cheatsheets — Pre-made engineering decisions and architectural guidelines accelerate AI development and improve output quality.
• Use model-to-model comparison to improve code quality — Comparing AI-generated code with a dedicated AI reviewer improves code quality and architectural consistency.
• v2.1.80 — Recent Claude Code updates enhance agent capabilities, user experience, and system stability through new features and critical bug fixes.
• How I built a Claude AI workflow that generates images and video on demand — Integrate AI visual generation directly into VS Code for a more efficient, browser-free workflow.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Lenny's Newsletter.