Field notes

Practical writing on AI-assisted work.

Essays, implementation notes, and research from building project systems that work with ChatGPT, Claude, Cursor, and production teams.

Featured

Introducing the Agiflow CLI: Scaling AI Agents Across Machines

GitHub Actions was never built for the fast closed loop an agent needs — going back, redoing a step, fixing its own work. Local agent fan-out solved the loop on one laptop and broke on two. The Agiflow CLI is the convenience wrapper we use internally to drive workflow locks, work units, and artifacts through the Agiflow API — so agents on different machines can pull the same backlog without stepping on each other.

Category: Best Practices
Date: May 7, 2026
Read: 8 min read

All articles

11 posts

May 3, 2026AI Architecture
Multi-Agent Orchestration with Claude and Codex: Role Separation, Handoff Contracts, and Verification Gates
Architect multi-agent code systems that stay coherent. Learn role separation patterns, handoff contracts, and verification gates to prevent coordination failures.
18 min read
Nov 30, 2025Best Practices
Roadmap to Build Scalable Frontend Applications with AI: Atomic Design System, Token Efficiency, and Design Systems
Learn how to architect frontend applications that scale with AI assistance. Discover how atomic design methodology, component libraries, and design systems dramatically reduce token consumption while ensuring consistent, maintainable codebases.
18 min read
Nov 16, 2025Cost Optimization
Token Efficiency in AI-Assisted Development: Why Your Tool Integration Architecture Matters More Than You Think
Optimized tool integration can reduce token consumption by 81% compared to baseline approaches. A controlled study reveals how architectural decisions—not protocol choices—determine scalability and cost in production AI systems.
25 min read
Oct 26, 2025Best Practices
Coordinating Multi-Task AI Workflows with Work Units: Building Complete Features in One Session
AI coding assistants excel at single tasks but struggle with multi-task feature coordination. Learn how work units + Project MCP enable structured AI agent workflows, reducing feature delivery time by 40% and maintaining context across 5-8 related tasks.
22 min read
Oct 18, 2025Technology
Claude Code Internals: Reverse Engineering Prompt Augmentation Mechanisms
Deep dive into Claude Code internals by instrumenting network traffic to understand how CLAUDE.md, output styles, slash commands, skills, hooks, and sub-agents actually work under the hood.
24 min read
Oct 12, 2025Best Practices
How to Enforce Architectural Patterns When AI Generates Your Code (Without Breaking Your Team)
AI coding assistants are shipping features 3x faster—but silently breaking architectural patterns. Learn how to reduce violations by 92% with MCP-based validation and feedback loops.
20 min read
Oct 5, 2025Best Practices
Scaling AI-Assisted Development: The Intelligent Scaffolding Approach
From "can AI write code?" to "can AI write code that scales?" Discover how intelligent scaffolding systems solve the consistency crisis in AI-assisted development.
15 min read
Aug 3, 2024Agentic Workflow
Optimise Cost and Speed of Agentic Workflow - Reversed Engineering Approach
Autonomous Agents show immense potential in solving complex problems, but they can also be brittle, prone to hallucinations, expensive, and slow to run.
8 min read
Jul 14, 2024Testing
Effective Practices for Mocking LLM Responses During the Software Development Lifecycle
Learn best practices for mocking LLM responses during development to improve testing efficiency and reduce API costs while maintaining code quality.
7 min read
Jun 8, 2024Technology
Challenges of Product Analytics in the Era of Generative AI
Explore the unique challenges and opportunities that generative AI brings to product analytics, and how to adapt traditional analytics approaches for AI-powered products.
10 min read

Community

Discuss the notes with the team.

Ask implementation questions, share what you are building, and follow new Agiflow research.

Join Discord

Introducing the Agiflow CLI: Scaling AI Agents Across Machines

All articles

Multi-Agent Orchestration with Claude and Codex: Role Separation, Handoff Contracts, and Verification Gates

Roadmap to Build Scalable Frontend Applications with AI: Atomic Design System, Token Efficiency, and Design Systems

Token Efficiency in AI-Assisted Development: Why Your Tool Integration Architecture Matters More Than You Think

Coordinating Multi-Task AI Workflows with Work Units: Building Complete Features in One Session

Claude Code Internals: Reverse Engineering Prompt Augmentation Mechanisms

How to Enforce Architectural Patterns When AI Generates Your Code (Without Breaking Your Team)

Scaling AI-Assisted Development: The Intelligent Scaffolding Approach

Optimise Cost and Speed of Agentic Workflow - Reversed Engineering Approach

Effective Practices for Mocking LLM Responses During the Software Development Lifecycle

Challenges of Product Analytics in the Era of Generative AI

Discuss the notes with the team.