All Articles
Multi-Agent Design with Claude Code: Phase Splitting and Context Engineering in Practice
A practical guide to designing multi-agent systems with the Claude Agent SDK. Learn phase-splitting for context optimization, session memory externalization, and automated feedback learning — with complete, production-ready code examples.
Claude API Multi-Agent Design Patterns: Implementation and Operations for Production Systems
A deep-dive into production-grade multi-agent design patterns using the Claude API. Covers orchestrator/subagent architecture, parallel execution, circuit breakers, context compression, and token budget management — with working code.
Understanding Claude in Chrome's MCP Tools: From take_snapshot to lighthouse_audit
A practical breakdown of Claude in Chrome's Playwright-based MCP tools. Learn when to use take_snapshot vs take_screenshot, how lighthouse_audit differs from performance tracing, and which tool combinations work best for common automation tasks.
Building a Production Customer Support AI Agent with Claude API: Hallucination Prevention, Multilingual Support, and Escalation Design
A complete guide to designing and implementing a production-grade customer support AI agent using Claude API. Covers grounding to prevent hallucinations, automatic multilingual responses, human escalation with context handover, and cost optimization — with working code throughout.
Build an English Learning Environment with Claude Code: Combining output-styles, skills, and hooks
Learn how to combine Claude Code's output-styles, skills, and hooks to create an English learning environment. This guide covers switching to English responses, auto-recording grammar corrections, and optional TTS readout — all without sacrificing productivity.
Claude Stops Mid-Response: A Complete Troubleshooting Guide
Why does Claude stop responding mid-generation? This guide diagnoses every cause — max_tokens limits, rate limits, timeouts, context overflow — and provides production-ready code patterns to fix each one.
Gemini 2.5 Pro vs Claude Sonnet 4.6: Real Coding Task Comparison for Japanese Developers
A hands-on comparison of Gemini 2.5 Pro and Claude Sonnet 4.6 across real coding tasks: bug fixing, feature implementation from Japanese specs, and code review. Find out which model fits your workflow.
Publishing npm Packages with Claude Code: From API Design to GitHub Actions
A practical guide to designing, building, testing, and publishing npm packages solo with Claude Code. Covers README-first design, TDD workflow, and automated publishing via GitHub Actions.
When Extended Thinking Actually Pays Off — and How to Configure It
A practical guide to Claude's Extended Thinking feature — which problems genuinely benefit from it, how to tune budget_tokens, and how to use the thinking block as a design tool, not just an output.
Why Claude Generates Outdated Code — and How Context7 MCP Fixes It
Claude Code sometimes generates deprecated or outdated API code due to its training data cutoff. Context7 MCP solves this by fetching live documentation from official library sources in real time, dramatically reducing broken code generation.
Claude Keeps Formatting With Markdown — How to Get Plain Text Output Without Asterisks and Bullets
Tired of Claude's responses filled with **bold text**, bullet points, and ## headers? Here are three practical methods to stop Markdown formatting and get clean plain text output every time.
Card Declined on Claude? Here's How to Fix It by Cause
Getting a payment declined error when renewing Claude Pro? This guide breaks down the most common causes — card issuer blocks, regional restrictions, and billing limits — and walks you through fixing each one.