Blog

Notes from a
one-person studio.

Longer writing than the ship log — how the apps are designed, built, and kept alive. Newest first.

17 posts

2026.04.28 Deep Dive: How Claude Code Works Under the Hood (Part 10) — Worktree Isolation, MCP & The Complete Architecture The final installment covering worktree isolation for concurrent agents, MCP plugin architecture, and a full retrospective on Claude Code's architecture from simple loop to multi-agent platform. claude-code

2026.04.25 Deep Dive: How Claude Code Works Under the Hood (Part 9) — Agent Teams & Coordination Protocols How Claude Code orchestrates multiple AI agents as persistent teammates with file-based messaging, structured protocols, autonomous task claiming, and graceful shutdown handshakes. claude-code

2026.04.22 Deep Dive: How Claude Code Works Under the Hood (Part 8) — Task Systems & Background Execution How Claude Code manages persistent task DAGs, background execution with daemon threads, and cron scheduling to handle complex, long-running work that survives restarts. claude-code

2026.04.19 Deep Dive: How Claude Code Works Under the Hood (Part 7) — Error Recovery That Actually Works How Claude Code handles failures gracefully with classified recovery strategies, budget-based retries, and visible state machines instead of hidden catch blocks. claude-code

2026.04.16 Deep Dive: How Claude Code Works Under the Hood (Part 6) — Memory & System Prompts How Claude Code remembers across sessions and assembles its system prompt. Covers memory categories, CLAUDE.md layering, the prompt build pipeline, and what makes AI agent memory actually useful. claude-code

2026.04.13 Deep Dive: How Claude Code Works Under the Hood (Part 5) — Permissions & Hooks The safety architecture of Claude Code explained: the four-stage permission pipeline, three permission modes, hook lifecycle events, and how control flow stays with the agent while hooks observe. claude-code

2026.04.10 Deep Dive: How Claude Code Works Under the Hood (Part 4) — Skills & Context Compaction How Claude Code manages knowledge efficiently through lazy skill loading and a multi-lever context compaction strategy. Learn the architecture behind context window management in AI agents. claude-code

2026.04.07 Deep Dive: Planning & Subagents — How Claude Code Stays Focused on Complex Tasks How Claude Code uses structured planning with TodoWrite, single-task constraints, nag injection, and subagents with isolated context to tackle complex coding tasks without losing focus. claude-code

2026.04.05 Vibe Coding with Claude Code: How I Build iOS Apps Faster with AI How I use Claude Code as an indie developer to ship iOS apps faster — from brainstorming features to writing Swift code, debugging, and automating repetitive tasks. claude-code

2026.04.04 Deep Dive: Tool Use & Dispatch Maps — How Claude Code Executes Your Commands Safely How Claude Code's tool system works under the hood — from dispatch maps and purpose-built tools to path sandboxing and security. A practical guide to AI agent tool architecture. claude-code

2026.04.03 Mastering Claude Code (Part 6): Advanced Features — Planning Mode, Checkpoints, Plugins & CLI Mastery The complete guide to Claude Code's advanced features: planning mode for complex tasks, checkpoints for safe experimentation, plugins for bundled workflows, and CLI mastery for automation. claude-code

2026.04.01 Deep Dive: The Agent Loop — The Heart of How Claude Code Actually Works A technical deep dive into the agent loop that powers Claude Code. Learn how the core cycle of prompt, tool execution, and write-back creates an autonomous AI coding assistant in under 30 lines of code. claude-code

2026.03.31 Mastering Claude Code (Part 5): Hooks — Event-Driven Automation That Runs Itself Learn how to set up Claude Code hooks — automatic shell commands that trigger on events like tool use, session start, file changes, and task completion. claude-code

2026.03.27 Mastering Claude Code (Part 4): MCP Protocol — Connecting AI to Your Entire Dev Stack How to use the Model Context Protocol (MCP) to connect Claude Code to GitHub, databases, file systems, and any external API for a truly integrated AI workflow. claude-code

2026.03.24 Mastering Claude Code (Part 3): Skills & Subagents — Building Automated AI Workflows How to create reusable skills and specialized subagents in Claude Code to automate code reviews, testing, documentation, and other development tasks. claude-code

2026.03.20 Mastering Claude Code (Part 2): Slash Commands & Memory — Your Personalized AI Workflow Learn how to create custom slash commands and configure Claude Code's memory system to build a personalized, context-aware development workflow. claude-code

2026.03.17 Mastering Claude Code (Part 1): Getting Started — From Zero to Productive in 15 Minutes A practical beginner's guide to Claude Code — what it is, how to set it up, and how to start using it for real development tasks in under 15 minutes. claude-code

Where the writing becomes software.

Every post here eventually turns into a version number. See what actually shipped.

Browse the ship log