# Publications — Ken Imoto

> Auto-generated from `src/content/{blog,books}/*.md` on each build.
> Do not edit by hand — see `scripts/generate-llms.mjs`.

## Books

### English

#### Chaos Engineering in Practice — From Netflix's monkey to AI-driven experiments

A 14-chapter book on Krkn-AI / Harness / Steadybit / Dynatrace, Chaos Engineering 2.0, and the operational practices around running chaos in production without making the news. The manuscript is finished and the book is being prepared for publication.

- LP: https://kenimoto.dev/books/chaos-engineering-guide/
- Published: 2026-12-31

#### Debugging Engineering — Ten habits, translated into prompts, layered with CLAUDE.md and hooks

Ten debugging habits — the rationale, the original-language formulations, and the chapter on translating each into Claude Code prompts. The manuscript is finished and the book is being prepared for publication.

- LP: https://kenimoto.dev/books/debugging-engineering/
- Published: 2026-12-31

#### Local LLM or Claude Code? — 96% of Requests Failed Locally, Half of the Steps Didn't: Measured AI Agent Routing with Ollama

Local LLM vs Claude Code, measured on 550 requests and steps: 96% of requests failed locally, half of single steps didn't. Rules, logprobs, thresholds.

- LP: https://kenimoto.dev/books/local-llm-step-routing/
- Kindle: https://www.amazon.com/dp/B0HLLC2Z2R
- Published: 2026-10-01

#### Cognitive Bias for Software Engineers — 20 Biases That Distort Estimates, Code Reviews, and Team Decisions

Cognitive bias for software engineers, written from both sides: the side that uses a psychological technique and the side that refuses to be fooled by it. 15 chapters covering the planning fallacy in estimates, framing in code review, anchoring in meetings, psychological safety and groupthink on teams, and the biases the AI era added. The countermeasures are mechanisms, not reminders to be careful. Built on Kahneman and Tversky, Cialdini, Edmondson, and empirical software engineering research.

- LP: https://kenimoto.dev/books/engineer-psychology-tricks/
- Kindle: https://www.amazon.com/dp/B0HH99235S
- Published: 2026-08-31

#### Why Do Some Words Stay With You Forever? — 100 Engineering Quotes Decoded

From Linux trenches to software philosophy, debugging, leadership, and architecture. 100 quotes from the engineering world, each unpacked through original context and modern application — a philosophy book for engineers, decoded one line at a time.

- LP: https://kenimoto.dev/books/engineer-it-quotes/
- Kindle: https://www.amazon.com/dp/B0GX3CS95M
- Published: 2026-05-01

#### The Practical Knowledge Graph Guide — Structure Your Data, Sharpen Your AI

RAG alone won't make your AI smart. Relationships come from structure: Knowledge Graph, GraphRAG, Neo4j, RDF vs Property Graph, Tree-sitter. 15 chapters.

- LP: https://kenimoto.dev/books/knowledge-graph-practical-guide/
- Kindle: https://www.amazon.com/dp/B0GX2Z73JV
- Published: 2026-04-28

#### The 300ms Threshold — Why Talking to AI Feels Wrong

A voice AI latency book. People reply about 200ms after you stop; with 300, 500, and 800ms as design targets, it shows how to shrink a 525ms cascade.

- LP: https://kenimoto.dev/books/voice-ai-300ms-ux/
- Kindle: https://www.amazon.com/dp/B0GYQ4L2KP
- Published: 2026-04-25

#### Systematizing AI Code Review — The 3-Layer Model for 60% Faster Reviews

Code review time keeps inflating because humans are doing the mechanical work. This book splits the job across three layers — hooks (machines), AI (first pass), and humans (architectural judgment) — and shows the actual implementation that cut review time by 60%. Includes integrated operation of CodeRabbit, GitHub Copilot, and Claude review, AGENTS.md policy design, GitHub Actions pipelines, and an autoFixable feedback loop, all backed by a Next.js + TypeScript reference project.

- LP: https://kenimoto.dev/books/harness-code-review/
- Kindle: https://www.amazon.com/Systematizing-AI-Code-Review-Propel-Lab-ebook/dp/B0GX71K2BV
- Published: 2026-04-21

#### LLMO Quickstart — AI Search Optimization for Engineers

Get LLMO (AI Search Optimization) running in 30 minutes. 8 chapters distill the essentials: writing llms.txt, the minimal JSON-LD patterns, and measuring AI citation rate. The fastest path for engineers who already know SEO to start getting cited by AI.

- LP: https://kenimoto.dev/books/llmo-quickstart/
- Kindle: https://www.amazon.com/dp/B0GHJLJ611
- Published: 2026-04-10

#### MCP Security in Practice — What OWASP Won't Tell You About AI Tool Integrations

MCP security guide covering the OWASP MCP Top 10, token cost measurements, file upload issues verified across 7 services, and lessons from running tax automation on MCP in production. Read this before you ship Model Context Protocol to production.

- LP: https://kenimoto.dev/books/mcp-security-practice/
- Published: 2026-03-25

#### Harness Engineering — From Using AI to Controlling AI

The Harness Engineering book: 5 interpretations from OpenAI, Anthropic, LangChain, Martin Fowler, and academia, merged into one system. The first systematic guide that distills the 6 building blocks, the AGENTS.md/CLAUDE.md/hooks implementation patterns, and Self-Evolving Agents — the practical reference for the 2026 keyword.

- LP: https://kenimoto.dev/books/harness-engineering-guide/
- Kindle: https://www.amazon.com/dp/B0FZNL8D1V
- Published: 2026-03-15

#### Why ChatGPT Ignores Your Website — The LLMO Practical Guide

Google SEO won't get you cited by ChatGPT or Perplexity. This book maps LLMO (LLM Optimization) — the patterns and implementations for AI search era. From llms.txt and JSON-LD to citation-rate KPIs, this is the field guide for being seen by AI search engines.

- LP: https://kenimoto.dev/books/llmo-ai-search-optimization/
- Kindle: https://www.amazon.com/dp/B0GQ2L5HQY
- Published: 2026-02-20

#### Turning LLMs from Liars into Experts — Context Engineering in Practice

Same question, wildly different answers? It is the context, not the prompt. Benchmarks show up to 4.6x quality gain across 5 stages, RAG, MCP, CLAUDE.md.

- LP: https://kenimoto.dev/books/context-engineering/
- Kindle: https://www.amazon.com/dp/B0GHRJX36Z
- Published: 2026-02-15

#### Practical Claude Code — Context Engineering for Modern Development

A field guide to Claude Code from a year in production: CLAUDE.md patterns from 2 to 100+ lines, Plan Mode, team operations, MCP, supply-chain security.

- LP: https://kenimoto.dev/books/claude-code-mastery/
- Kindle: https://www.amazon.com/dp/B0GPWPKNB7
- Published: 2025-09-01

### Japanese (日本語)

#### カオスエンジニアリング実践ガイド — Netflixの猿からAI駆動の実験へ

Krkn-AI / Harness / Steadybit / Dynatrace の全景、Chaos Engineering 2.0、本番でカオスを回しつつニュースにならないための運用手法を14章でまとめたBook。出版準備中。

- LP: https://kenimoto.dev/ja/books/chaos-engineering-guide/
- Published: 2026-12-31

#### Claude Code で動かすマイコン — AI エンジニアが組み込みに踏み込む実装入門

USB シリアル経由で LED を点けるところから始めて、MCP サーバーでデバイスを公開し、UART/I2C/SPI を AI から見える形に抽象化する。AI 側と組み込み側の語彙を互いに翻訳する12章。出版準備中。

- LP: https://kenimoto.dev/ja/books/claude-code-hardware-control/
- Published: 2026-12-31

#### デバッグの技法 — 10の思考武器と現場の物語

前提を疑う、再現を取る、境界で切り分ける。10の思考武器を実際の障害事例で追い、Claude Code のプロンプトに変換するところまで扱うBook。出版準備中。

- LP: https://kenimoto.dev/ja/books/debugging-engineering/
- Published: 2026-12-31

#### エンジニアのための多層防御設計 — 穴のないチーズは存在しない

「壊れない」ではなく「壊れ方を選ぶ」。スイスチーズモデル、Circuit Breaker、Bulkhead、Blast Radius、SOLID、テストと可観測性まで、設計で守る型を16章でまとめたBook。出版準備中。

- LP: https://kenimoto.dev/ja/books/multi-layer-defense-design/
- Published: 2026-12-31

#### トレースとログの教科書 — フロントエンドとバックエンドで何が違うのか

同じシグナルがフロントとバックで違う意味を持つ。その非対称性を軸に、OpenTelemetry の基礎、Datadog RUM とのハイブリッド実装、SPA・API・マイクロサービス別の設計チェックリストを13章と付録でまとめたBook。出版準備中。

- LP: https://kenimoto.dev/ja/books/observability-frontend-backend/
- Published: 2026-12-31

#### 人流分析実践ガイド — Pythonで読み解く都市の動き

国交省のオープンデータと GeoPandas で OD分析・滞在分析を回す。AIによるリアルタイム計測、防災や出店判断への応用まで、全コード検証済みの14章。出版準備中。

- LP: https://kenimoto.dev/ja/books/people-flow-analysis-guide/
- Published: 2026-12-31

#### なぜNVIDIAはロボットに全賭けしたのか — エンジニアのためのフィジカルAI入門

World Model、VLA、Sim-to-Real で技術スタックを整理し、LeRobot × MuJoCo のハンズオンで実際に動かす。実機を持たないエンジニアがフィジカルAIに入るための16章。出版準備中。

- LP: https://kenimoto.dev/ja/books/physical-ai-intro/
- Published: 2026-12-31

#### リファクタリングの教科書 — React / TypeScript / Python 実践パターン

Martin Fowler の古典的カタログを React・TypeScript・Python の3言語で現代化した20章。コードの不吉な臭いを見分け、Strangler Fig で段階的に置き換え、静的解析とAIツールで検出を自動化するところまで扱うBook。出版準備中。

- LP: https://kenimoto.dev/ja/books/refactoring-patterns/
- Published: 2026-12-31

#### 声が届く範囲を設計する

特殊なデバイスは使わない。什器の配置と吸音、マスキング音だけで、隣の人に声が届かないカウンターを作る。音波の物理と法規制から、Zoom/Meet の設定と運用チェックリストまで13章。出版準備中。

- LP: https://kenimoto.dev/ja/books/speech-privacy-design/
- Published: 2026-12-31

#### ビデオ通話アプリの自動テスト — Playwrightで2台分のテスターを置き換える

2デバイス・ネットワーク・CPU という3つの環境依存を、仮想カメラと CDP で切り分ける。2ブラウザ同期テスト、Docker + GitHub Actions の CI、Flaky Test 対策までを12章でまとめたBook。出版準備中。

- LP: https://kenimoto.dev/ja/books/webrtc-e2e-playwright/
- Published: 2026-12-31

#### ローカルLLMで足りる仕事、Claude Codeに任せる仕事 — 判定AI「Jev」流に振り分ける、AIエージェントの実測と設計

Claude CodeとローカルLLMの併用を実測。依頼単位では96%回せず、1ステップずつなら半分。Ollamaの振り分け・機密の門番・しきい値を扱う本。

- LP: https://kenimoto.dev/ja/books/local-llm-step-routing/
- Kindle: https://www.amazon.co.jp/dp/B0HLGHJ95X
- Published: 2026-10-01

#### AIくさい文章から脱出する技術 — 6モデル180サンプルの実測でわかった「バレる文章」の正体と差別化の手順

「AIっぽい文章」の正体を、6つのLLMに同一テーマで日本語を書かせた180サンプル×16指標の実測で特定する実践書。英語圏では ChatGPT 登場後に \"delves\" が28倍に増えたという研究があるが、日本語で同じ測定をした先行例は見つからなかったので自分で測った。実測の上位は「適切な」0.59回/記事、「重要です」0.34、「本記事では」0.27。逆に、AIっぽいと感じていた「浮き彫りにする」は180サンプルに1回も出てこない。語彙・構造・リズムの3方向から差別化する手順、Claude Code Skill での自動監査、商業出版のプロ校正が見る3層、そして「人間が書いたのにAI判定される」問題までを全26章で扱う。2026年3月の初版に、5ヶ月後の追試で判明した初版自身の誤り3件の訂正を加えた増補改訂版。実験コードと全サンプルは GitHub で公開している。

- LP: https://kenimoto.dev/ja/books/ai-text-slop-escape/
- Kindle: https://www.amazon.co.jp/dp/B0H75RMJXT
- Published: 2026-08-23

#### 8000行でわかる大規模言語モデル — Karpathyのnanochatで追う、事前学習から推論エンジンまでの全工程

Andrej Karpathyが公開したnanochat(実装約8,000行)を、工程の順番に上から読み切るコードリーディング本。tokenizer訓練→事前学習→SFT→強化学習(GRPO)→評価(CORE)→推論エンジン(KVキャッシュ)の全工程を、実コードを引用しながら追う。2019年に$43,000だったGPT-2級モデルの訓練が2026年には$48まで下がった、その中身が何なのかをリポジトリ内の出典まで遡って確認する。GPUがなくてもruncpu.shでMacBook/ノートPCだけで全工程を一周する第8章まで、全8章+付録。Zennで全章無料公開。

- LP: https://kenimoto.dev/ja/books/nanochat-code-reading/
- Kindle: https://www.amazon.co.jp/dp/B0HCH7VTBX
- Zenn: https://zenn.dev/kenimo49/books/nanochat-code-reading
- Published: 2026-07-31

#### RTX 4070でQwen 35Bを2.8倍速くする — ローカルLLM 高速化・チューニングの実測ノート llama.cpp 実践ガイド

12GBのRTX 4070で --cpu-moe を測ると12.2→34.6 tok/s。自分で測った数字だけのローカルLLM高速化ガイド、全10章。

- LP: https://kenimoto.dev/ja/books/local-llm-qwen-4070/
- Kindle: https://www.amazon.co.jp/dp/B0HCH7MGW1
- Zenn: https://zenn.dev/kenimo49/books/local-llm-qwen-4070
- Published: 2026-07-03

#### AIに選ばれる店をつくる — 店舗オーナーのための AI Native MEO 実践

検索が「Googleで探す」から「AIに相談して店を決める」へ移るいま、店舗集客の主戦場も変わった。本書はLLMO / AI Search Optimization を専門とするエンジニアが書く『AI Native MEO』時代の店舗集客の実践書。月1.5万〜3万円のMEO代行費の中身を分解し、ClaudeやChatGPTのプロンプト(実物同梱)で自分の手で回す方法を、GBP最適化7手法・業種別口コミ返信・業者ツール32社カタログ・飲食/美容/治療院の公開事例・副業MEO代行・景表法/薬機法対応まで全17章+付録で解説する。

- LP: https://kenimoto.dev/ja/books/store-owner-ai-meo-llmo/
- Kindle: https://www.amazon.co.jp/dp/B0H3VX2YZC
- Published: 2026-06-03

#### エンジニアの心理トリック大全 — 認知バイアスから読み解くエンジニアの実務心理学

経験のあるエンジニアなら覚えがあるはずです。技術的な妥当性を説明したのに「今のままで困ってなくない?」と返される。3週間と見積もった作業に6週間かかる。何時間もかけてコードレビューをしたのに、次のPRも地雷原のまま。原因の多くは技術力の不足ではありません。人間の脳に組み込まれた認知バイアスが、判断を静かに歪めています。本書は20の認知バイアスと心理テクニックをエンジニアの実務に引き付けて解説します。全15章を4部構成で、基礎編(システム1/2 + AI時代の新バイアス)、対自分編(見積もり/技術選定/デバッグ/キャリア)、対人編(コードレビュー/説得/1on1/会議/交渉)、チーム編(心理的安全性/リモートワーク/採用/ダークパターン)と、個人の頭の中からチームの力学まで順に降りていきます。Zennで全章無料公開。

- LP: https://kenimoto.dev/ja/books/engineer-psychology-tricks/
- Kindle: https://www.amazon.co.jp/dp/B0HCJB8MF3
- Zenn: https://zenn.dev/kenimo49/books/engineer-psychology-tricks
- Published: 2026-05-20

#### エンジニアリング100の言葉 — なぜその一文は記憶に残るのか

Linuxの現場、ソフトウェア哲学、デバッグ、リーダーシップ、アーキテクチャ。エンジニアリングの世界に残る100の名言を、原典の文脈と現代の応用を交えて解説。記憶に残る一文の構造を解きほぐす、エンジニアの哲学書。

- LP: https://kenimoto.dev/ja/books/engineer-it-quotes/
- Kindle: https://www.amazon.co.jp/dp/B0GXR5Q6NW
- Published: 2026-05-01

#### ナレッジグラフ活用大全 — 構造化すれば、AIは賢くなる

RAGだけではAIは賢くなりません。Neo4j・GraphRAG・RDF・Property Graph・Tree-sitterで構造を持たせる実践ガイド15章。

- LP: https://kenimoto.dev/ja/books/knowledge-graph-practical-guide/
- Kindle: https://www.amazon.co.jp/dp/B0GX465PG7
- Published: 2026-04-28

#### 音声AIの300ms — 人はなぜAIとの会話に違和感を覚えるのか

音声AIのレイテンシ設計。人の会話は約200msで回る。300/500/800msを目安に、525msのカスケードをストリーミングと知覚ハックで縮める。

- LP: https://kenimoto.dev/ja/books/voice-ai-300ms-ux/
- Kindle: https://www.amazon.co.jp/dp/B0GX2R5HP4
- Zenn: https://zenn.dev/kenimo49/books/voice-ai-300ms-ux
- Published: 2026-04-25

#### LLMOクイックスタート — エンジニアのためのAI検索最適化入門

LLMO (AI検索最適化) を30分で始めるための入門書。llms.txt の書き方、JSON-LD の最低実装、AI引用率の測定まで、8章で核心だけを凝縮。Google SEO の知識を持つエンジニアが、今日から AI検索に拾われるサイトを作るための短時間ガイド。

- LP: https://kenimoto.dev/ja/books/llmo-quickstart/
- Kindle: https://www.amazon.co.jp/dp/B0GHMCKQPX
- Published: 2026-04-10

#### AIコードレビューを仕組み化する技術 — hooks・AI・人間の3層モデル

AIコードレビューを仕組み化する3層モデル。hooks がフォーマットを強制、CodeRabbit/Copilot/Claude が一次レビュー、人間は設計判断に集中。AGENTS.md とフィードバックループで仕組みが進化する設計を Next.js + TypeScript の実装例で解説。

- LP: https://kenimoto.dev/ja/books/harness-code-review/
- Kindle: https://www.amazon.co.jp/dp/B0GHT7FQ7G
- Zenn: https://zenn.dev/kenimo49/books/harness-code-review
- Published: 2026-04-01

#### MCP実践セキュリティ — 本番導入で躓かないための完全ガイド

MCP (Model Context Protocol) を本番導入する前に読むべきセキュリティガイド。トークンコスト実測、ファイルアップロード問題の7サービス検証、OWASP MCP Top 10、freee確定申告自動化の実体験ベースで、安全に本番運用するための知識を体系化。

- LP: https://kenimoto.dev/ja/books/mcp-security-practice/
- Published: 2026-03-25

#### なぜAI生成UIは全部青いのか — 同質化から脱出する技術

AIに作らせたUIが、どれも同じ見た目になる理由を4モデル×40色の実験データで解明。色認識バイアスと分布の収束という原因を示したうえで、Typography・Color・Motion・Spatial Composition・Backgrounds の5軸で抜け出す方法を、全17章とBefore/Afterの実例で扱う。Anthropic公式 frontend-design Skill の中身も分解する。

- LP: https://kenimoto.dev/ja/books/ai-slop-escape-guide/
- Kindle: https://www.amazon.co.jp/dp/B0GNCNLK9K
- Zenn: https://zenn.dev/kenimo49/books/ai-slop-escape-guide
- Published: 2026-03-22

#### LLMOに最適化されたホームページをゼロから作る — GitHub Pagesで無料公開、AIに見つけてもらうサイト設計

AIに引用されるサイトを、実物のコードとともにゼロから作る。Astro と GitHub Pages で無料構築し、llms.txt・robots.txt の AI クローラー設定、JSON-LD 構造化データ、同一URLに .md を付けて Markdown を返す URL.md パターン、Playwright による LLMO 対応テスト、カスタムドメインと HTTPS まで、10章で全工程を扱う。最後に ChatGPT・Claude・Perplexity に実際に質問して、施策の効果を検証する。

- LP: https://kenimoto.dev/ja/books/llmo-website-builder/
- Zenn: https://zenn.dev/kenimo49/books/llmo-website-builder
- Published: 2026-03-20

#### ハーネス・エンジニアリング — AIを\"使う\"から\"操る\"へ

定義が5つに割れています。OpenAIからMartin Fowlerまで突き合わせ、AGENTS.md/CLAUDE.md/hooksの実装に落とした全14章。

- LP: https://kenimoto.dev/ja/books/harness-engineering-guide/
- Kindle: https://www.amazon.co.jp/dp/B0GF8VGSPC
- Published: 2026-03-15

#### Claude Codeクイックスタート — 30分で始めるAIエージェント開発

Claude Code をこれから触る人向けの入門書。インストールと初期設定、CLAUDE.md を置くと開発体験がどう変わるか、Read や Grep や Bash をどう使い分けるか、そして RAG ではなく Agentic Search が採られた理由まで、5章で扱う。Zenn で全文を無料公開している。

- LP: https://kenimoto.dev/ja/books/claude-code-quickstart/
- Zenn: https://zenn.dev/kenimo49/books/claude-code-quickstart
- Published: 2026-03-10

#### LLMO実践ガイド — なぜChatGPTはあなたのサイトを無視するのか

Google SEO だけでは ChatGPT / Perplexity に届かない。LLMO (LLM Optimization = AI検索最適化) の考え方と実装パターンを、構造化データ・llms.txt・JSON-LD・引用率最適化まで体系的に解説。AI検索時代に「拾われるサイト」を作るための実践書。

- LP: https://kenimoto.dev/ja/books/llmo-ai-search-optimization/
- Kindle: https://www.amazon.co.jp/dp/B0GQ4XPBMJ
- Zenn: https://zenn.dev/kenimo49/books/llmo-ai-search-optimization
- Published: 2026-02-20

#### LLMを「嘘つき」から「専門家」に変える技術 — Context Engineering 実践入門

「同じ質問なのに、AIの回答が全然違う」原因はプロンプトではなくコンテキストにあります。架空の社内ツール3つで独自ベンチマーク実験を行い、コンテキスト戦略でAIの回答品質が最大4.6倍変わることを定量実証。「大規模モデルほど上手に嘘をつく」「小さいモデル+RAG > 大きいモデル単体」など衝撃の結果と、5段階のコンテキスト戦略・RAG・MCP・CLAUDE.mdまで体系的に解説。

- LP: https://kenimoto.dev/ja/books/context-engineering/
- Kindle: https://www.amazon.co.jp/dp/B0GHNC29MF
- Zenn: https://zenn.dev/kenimo49/books/context-engineering
- Published: 2026-02-15

#### 実践Claude Code — コンテキストエンジニアリングで開発が変わる

Claude Code を1年以上、実務で使い込んだ。CLAUDE.md の書き方、Plan Mode 起点の1日の開発フロー、チーム展開、セキュリティ。19章で、現場で得た知見を1冊にまとめた。

- LP: https://kenimoto.dev/ja/books/claude-code-mastery/
- Kindle: https://www.amazon.co.jp/dp/B0GPSVW3BJ
- Zenn: https://zenn.dev/kenimo49/books/claude-code-mastery
- Published: 2025-09-01

### Portuguese (Português)

#### Chaos Engineering na Prática — Do macaco da Netflix aos experimentos guiados por IA

Livro de 14 capítulos sobre Krkn-AI / Harness / Steadybit / Dynatrace, Chaos Engineering 2.0, e as práticas operacionais para rodar chaos em produção sem virar manchete. O manuscrito está pronto e o livro está em preparação para publicação.

- LP: https://kenimoto.dev/pt/books/chaos-engineering-guide/
- Published: 2026-12-31

#### Manual completo de Knowledge Graph — Estruture os dados, torne a IA mais inteligente

O RAG sozinho não deixa sua IA inteligente. Relacionamentos só aparecem com estrutura — Knowledge Graph, GraphRAG, Neo4j, RDF, Property Graph, Tree-sitter, MCP e Emotion AI. O guia prático para dar à IA a inteligência estruturada que falta ao embedding.

- LP: https://kenimoto.dev/pt/books/knowledge-graph-practical-guide/
- Kindle: https://www.amazon.com.br/dp/B0H3FVPXZV
- Published: 2026-05-30

#### Por Que Algumas Palavras Ficam para Sempre? — 100 Frases da Engenharia Decifradas

Da trincheira do Linux à filosofia de software, debugging, liderança e arquitetura. 100 frases do mundo da engenharia, cada uma destrinchada pelo contexto original e pela aplicação moderna — um livro de filosofia pra engenheiros, decifrado uma linha por vez.

- LP: https://kenimoto.dev/pt/books/engineer-it-quotes/
- Kindle: https://www.amazon.com.br/dp/B0H2KR8L6T
- Published: 2026-05-22

#### LLMO Quickstart — Otimização para Busca por IA para Engenheiros

Coloque o LLMO (Otimização para Busca por IA) pra rodar em 30 minutos. 8 capítulos destilam o essencial: escrever llms.txt, os padrões mínimos de JSON-LD e medir taxa de citação por IA. O caminho mais rápido pra engenheiros que já conhecem SEO começarem a ser citados pela IA.

- LP: https://kenimoto.dev/pt/books/llmo-quickstart/
- Kindle: https://www.amazon.com.br/dp/B0H1YG3QDP
- Published: 2026-05-21

#### Manual completo dos truques psicológicos para engenheiros — Vieses cognitivos e técnicas psicológicas traduzidos para a prática do engenheiro

Engenheiro experiente já viveu o roteiro: ouve 'do jeito que está hoje não tá bom?' depois de explicar a pertinência técnica. Estima três semanas e gasta seis. Revisão de código por horas e a próxima PR ainda parece um campo minado. O problema, na maior parte das vezes, não é a habilidade técnica. É o viés cognitivo embutido no cérebro humano distorcendo silenciosamente cada decisão. Este livro traduz vinte vieses cognitivos e técnicas psicológicas para a prática diária do engenheiro, em 15 capítulos divididos em quatro partes: fundamentos (Sistema 1/2 + era da IA), contra si mesmo (estimativas / seleção tecnológica / debug / carreira), com outras pessoas (revisão de código / persuasão / 1:1 / reuniões / negociação) e equipe (segurança psicológica / trabalho remoto / contratação / dark patterns).

- LP: https://kenimoto.dev/pt/books/engineer-psychology-tricks/
- Kindle: https://www.amazon.com.br/dp/B0GZFBC7SP
- Published: 2026-05-20

#### Revisão de Código com Harness Engineering — Automação de revisão na era dos agentes de IA

Transforme a revisão de código de pedido em sistema. Modelo de três camadas (portão automático / revisão por IA / revisão humana) usando hooks, CodeRabbit, GitHub Actions e AGENTS.md. 15 capítulos práticos com exemplos copia-e-cola em Next.js + TypeScript + Prisma.

- LP: https://kenimoto.dev/pt/books/harness-code-review/
- Kindle: https://www.amazon.com.br/dp/B0H2DB9YXD
- Published: 2026-05-20

#### Transformando LLMs de Mentirosos em Especialistas — Engenharia de Contexto na Prática

A mesma pergunta dá respostas diferentes — o problema é o contexto, não o prompt. Benchmarks originais mostram ganho de até 4,6× em qualidade. O sistema completo de Engenharia de Contexto: 5 estágios, RAG, MCP, CLAUDE.md, Agentic RAG.

- LP: https://kenimoto.dev/pt/books/context-engineering/
- Kindle: https://www.amazon.com.br/dp/B0H12378CJ
- Published: 2026-05-08

#### Harness Engineering — De Usar IA a Controlar IA

Harness Engineering, integrada nas 5 interpretações de OpenAI, Anthropic, LangChain, Martin Fowler e da academia. O primeiro guia sistemático que destila os 6 componentes, os padrões de implementação de AGENTS.md/CLAUDE.md/hooks e os Self-Evolving Agents — a referência prática para a palavra-chave de 2026.

- LP: https://kenimoto.dev/pt/books/harness-engineering-guide/
- Kindle: https://www.amazon.com.br/dp/B0GZ4CYY6W
- Published: 2026-05-08

#### Practical Claude Code — Engenharia de Contexto que Transforma seu Desenvolvimento

Guia de campo para Claude Code após 1 ano em produção: padrões de CLAUDE.md de 2 a 100 linhas, Plan Mode, operação em equipe e segurança. 19 capítulos.

- LP: https://kenimoto.dev/pt/books/claude-code-mastery/
- Kindle: https://www.amazon.com.br/dp/B0GZHB7J3T
- Published: 2026-05-05

### Spanish (Español)

#### Convirtiendo LLMs de Mentirosos en Expertos — Ingeniería de Contexto en la Práctica

¿Tu IA miente con seguridad? El problema no es el modelo — es el contexto. En tres herramientas internas ficticias, este libro mide la calidad de respuesta con cinco estrategias de contexto y demuestra que un modelo pequeño con RAG bien diseñado supera a un modelo más grande operando solo. A partir de esos datos construye el sistema completo de Ingeniería de Contexto: estrategia en 5 etapas, RAG, MCP, CLAUDE.md y Agentic RAG.

- LP: https://kenimoto.dev/es/books/context-engineering/
- Kindle: https://www.amazon.com.mx/dp/B0H12JR228
- Published: 2026-05-09

#### Harness Engineering — De Usar IA a Controlar IA

Harness Engineering, integrada en las 5 interpretaciones de OpenAI, Anthropic, LangChain, Martin Fowler y la academia. La primera guía sistemática que destila los 6 componentes, los patrones de implementación de AGENTS.md/CLAUDE.md/hooks y los Self-Evolving Agents — la referencia práctica para la palabra clave de 2026.

- LP: https://kenimoto.dev/es/books/harness-engineering-guide/
- Kindle: https://www.amazon.com.mx/dp/B0GXSVHZ5G
- Published: 2026-05-09

#### Practical Claude Code — La Ingeniería de Contexto que Transforma tu Desarrollo

Practical Claude Code enseña cómo usar Claude Code y escribir CLAUDE.md a partir de más de un año de uso real en producción. De la filosofía de diseño a los patrones de CLAUDE.md, de los flujos de Plan Mode a la operación en equipo y la seguridad: la guía de campo para ingenieros que quieren dominar Claude Code de verdad.

- LP: https://kenimoto.dev/es/books/claude-code-mastery/
- Kindle: https://www.amazon.com.mx/dp/B0GX3S7VGK
- Published: 2026-05-07

## Blog Articles

### English

- [Claude Code Max vs API vs Qwen: 4h Breakeven](https://kenimoto.dev/blog/claude-code-max-vs-api-vs-qwen-hour-of-day-breakeven/) — 2026-09-09
- [Gemini CLI vs Claude Code: Login Blocked Day 1](https://kenimoto.dev/blog/gemini-cli-vs-claude-code-workspace-account-locked-out/) — 2026-09-08
- [nanochat GRPO Is Just REINFORCE + a 1-Line Regex](https://kenimoto.dev/blog/karpathy-nanochat-grpo-8000-lines-reinforce-regex/) — 2026-09-07
- [2.8x Traffic Spike: How to Identify Bot Traffic](https://kenimoto.dev/blog/crawler-three-types-user-agent-check/) — 2026-09-02
- [Bing Webmaster Tools vs Google: 787 queries](https://kenimoto.dev/blog/google-search-demand-suggest-echo-787-queries/) — 2026-09-02
- [Zenn noindex on Every Post: 8 Weeks Unnoticed](https://kenimoto.dev/blog/zenn-noindex-8-weeks-unnoticed/) — 2026-08-31
- [Harness Engineering Benchmark: 13.7pt Same Model](https://kenimoto.dev/blog/harness-vs-prompt-52-to-66/) — 2026-08-28
- [pgvector vs Qdrant vs Weaviate at 10k: No Cliff](https://kenimoto.dev/blog/pgvector-vs-qdrant-vs-weaviate-10k-agent-memories/) — 2026-08-27
- [I Measured AI Citation Half-Life for 90 Days: 6 of 10 Sources Vanish by Week 4](https://kenimoto.dev/blog/ai-citation-90-days-6-of-10-sources-vanish/) — 2026-08-25
- [Static Call Graphs Miss 61% of Invocations](https://kenimoto.dev/blog/static-call-graphs-miss-61-percent-runtime-traces-ai-code-review/) — 2026-08-25
- [AI Mode Cited My Portuguese 3.4× English. Japanese Also Beat English.](https://kenimoto.dev/blog/ai-mode-portuguese-vs-english-cross-language-llmo/) — 2026-08-23
- [Claude Skills vs Subagents in 2026: Which One To Reach For (7 Decision Rules)](https://kenimoto.dev/blog/claude-skills-vs-subagents-7-decision-rules/) — 2026-08-20
- [Claude Code vs ChatGPT Codex: 30-Day Cost by 7 Task Types](https://kenimoto.dev/blog/claude-code-vs-chatgpt-codex-30-day-cost-per-task-type/) — 2026-08-18
- [Spec-Driven Development with Claude Code: 3 Ways the Spec Itself Broke Us](https://kenimoto.dev/blog/spec-driven-development-claude-code-spec-broke-us/) — 2026-08-15
- [Codex CLI vs Claude Code: 7 Real Tasks, Same Repo, 31 Days Later](https://kenimoto.dev/blog/codex-cli-vs-claude-code-7-real-tasks-31-days/) — 2026-08-14
- [GitHub Copilot Agent Mode vs Claude Code: 8 Tasks, 31 Days, Real Bills](https://kenimoto.dev/blog/github-copilot-agent-mode-vs-claude-code-8-tasks-31-days/) — 2026-08-13
- [Claude Code vs Aider on Same 10 Tasks — One Finished 8, the Other 5](https://kenimoto.dev/blog/claude-code-vs-aider-10-tasks-eight-vs-five/) — 2026-08-12
- [Claude Code vs Qwen3-35B on RTX 4070: 34.6 tok/s Break-even](https://kenimoto.dev/blog/claude-code-vs-qwen3-35b-rtx-4070-benchmark/) — 2026-08-11
- [Karpathy's nanochat on a MacBook: End-to-End runcpu.sh Log from Tokenizer to Chat (2026)](https://kenimoto.dev/blog/karpathy-nanochat-macbook-runcpu-end-to-end-log/) — 2026-08-06
- [Claude setup-token: The 2-Layer Auth Trap (2026)](https://kenimoto.dev/blog/claude-code-two-layer-auth-setup-token/) — 2026-08-05
- [The Shai-Hulud npm attack: why signature verification, npm audit, and --ignore-scripts all failed](https://kenimoto.dev/blog/shai-hulud-npm-supply-chain-why-verification-failed/) — 2026-08-05
- [Claude Code Diagrams: Excalidraw MCP + 3 Rules](https://kenimoto.dev/blog/claude-code-shared-canvas-diagramming/) — 2026-08-04
- [Claude Code: The Operator's Field Guide](https://kenimoto.dev/blog/claude-code/) — 2026-08-04
- [llms.txt vs Claude Skills manifest: I tested 5 AI engines to see which file they actually read](https://kenimoto.dev/blog/llms-txt-vs-claude-skills-manifest-5-ai-engines/) — 2026-08-03
- [When Claude Code's Auto Mode Blocks Only Bash: Investigating the Safety Classifier Outage, Plus a Fail-Open-on-Outage Hook Design](https://kenimoto.dev/blog/claude-code-auto-mode-classifier-fail-open-hook/) — 2026-08-02
- [Chat Models Are Born in the Loss Mask — Reading nanochat's SFT](https://kenimoto.dev/blog/nanochat-sft-loss-mask-where-chat-models-are-born/) — 2026-08-02
- [Five Days of Tracking 346 Aftershocks From 90 km Away — and a Rematch With the 2016 Kumamoto Earthquake](https://kenimoto.dev/blog/uto-m71-aftershock-five-days-fukuoka/) — 2026-08-01
- [nanochat: $48 to Retrain GPT-2 (8×H100, 2026)](https://kenimoto.dev/blog/karpathy-nanochat-gpt-2-training-cost-43000-to-48-h100/) — 2026-07-31
- [JMA Earthquake Prediction Impossible: Day 8 Miss](https://kenimoto.dev/blog/earthquake-prediction-vs-aftershock-forecasting/) — 2026-07-29
- [llmoframework Audit on 30 Dev Blogs: 4 of Top 5 Fail Same 3 Checks](https://kenimoto.dev/blog/llmoframework-audit-30-dev-blogs-top-5-fail-same-3-checks/) — 2026-07-29
- [Cursor Composer vs Claude Code: 400k-Token Repo Benchmarked](https://kenimoto.dev/blog/cursor-composer-vs-claude-code-400k-token-repo/) — 2026-07-28
- [nanochat: I Retrained GPT-2 for $48 in 2026](https://kenimoto.dev/blog/karpathy-nanochat-43k-gpt2-8000-lines/) — 2026-07-26
- [PinchTab Only Shoots One Viewport: Teaching a 9.4k-Star Browser Bridge to Capture Full Pages](https://kenimoto.dev/blog/pinchtab-scroll-shot-full-page-screenshots/) — 2026-07-22
- [pixelshot Read One Tile of an 18,609px Wikipedia Page: the Lazy-Load Trap Under Visual RAG](https://kenimoto.dev/blog/pixelshot-1-tile-wikipedia-lazy-load-trap/) — 2026-07-22
- [The Machine Accent Travels: AI Text Is Rhythmically Monotone in 70/70 Cells Across 3 Languages](https://kenimoto.dev/blog/machine-accent-3-languages-70-70-cells/) — 2026-07-18
- [Article Schema Alone Didn't Make AI Recognize Me as the Author. The Entity Wiring That Did (in 4 JSON-LD Fields).](https://kenimoto.dev/blog/article-schema-alone-author-entity-4-json-ld-fields/) — 2026-07-13
- [OpenCut MCP: Drive a Video Editor from Claude Code (4 Traps)](https://kenimoto.dev/blog/opencut-classic-mcp-4-traps-editor-core-fork/) — 2026-07-13
- [MCP Server Audit: 4 Layers, A-F Grade](https://kenimoto.dev/blog/pre-flight-your-mcp-4-layers-scorecard/) — 2026-07-13
- [The Skill Eval Repo I Didn't Build: 107 SKILL.md Files, 6 Checks, 21 False Positives](https://kenimoto.dev/blog/skill-eval-repo-not-built-107-lint/) — 2026-07-12
- [Ship a product, get a support button for free: an edge-injected overlay on Cloudflare Workers](https://kenimoto.dev/blog/edge-injected-support-overlay-cloudflare-workers/) — 2026-07-11
- [FUNDING.yml but No Sponsor Button? 42-Repo Audit](https://kenimoto.dev/blog/funding-yml-sponsor-button-42-repo-audit/) — 2026-07-11
- [historymap: one YAML file becomes a corporate-style product-history timeline](https://kenimoto.dev/blog/historymap-yaml-corporate-timeline-oss/) — 2026-07-11
- [Claude Code Made Me 40% Slower: 3 Places](https://kenimoto.dev/blog/claude-code-40-slower-3-places/) — 2026-07-06
- [The 7-Step Test That Told Me When to Switch From RAG to GraphRAG](https://kenimoto.dev/blog/graphrag-vs-rag-7-step-switch-test/) — 2026-07-04
- [I Ran Claude Code, Cursor, and Codex Side by Side for 31 Days. Here Is the Real Monthly Bill.](https://kenimoto.dev/blog/claude-code-cursor-codex-31-days-real-monthly-bill/) — 2026-07-03
- [Claude Code vs Cursor: 6 Tasks, 1 Uninstalled](https://kenimoto.dev/blog/claude-code-vs-cursor-6-tasks-uninstalled/) — 2026-07-02
- [Claude Code, Cursor, Codex: 340k Token Breakeven](https://kenimoto.dev/blog/parallel-agents-340k-tokens-breakeven-claude-cursor-codex/) — 2026-07-01
- [Qwen 35B at 2.8x on RTX 4070: One Flag Does It](https://kenimoto.dev/blog/rtx-4070-cpu-moe-flag-2-8x-tokens/) — 2026-06-30
- [Your AI Code Review Burns 80% of the Context Window on Files It Never Needed](https://kenimoto.dev/blog/ai-code-review-80-percent-context-waste/) — 2026-06-29
- [I Made Claude Code Review Only the Blast Radius — Token Bill Dropped 8-49x](https://kenimoto.dev/blog/claude-code-blast-radius-review-tokens-8-49x/) — 2026-06-28
- [Anthropic frontend-design skill: #F4F1EA Named](https://kenimoto.dev/blog/anthropic-frontend-design-skill-rewrite/) — 2026-06-27
- [Claude Code, Cursor, Codex: 412 Decisions/Day](https://kenimoto.dev/blog/claude-cursor-codex-parallel-decision-fatigue-400/) — 2026-06-27
- [Measuring AI Citation Half-Life: A 90-Day Methodology With 3 Real Decay Curves](https://kenimoto.dev/blog/measuring-ai-citation-half-life-90-day-methodology/) — 2026-06-26
- [Multi-Agent Decision Fatigue: I Counted 412 Micro-Choices a Day. The Harness Cut It to 38.](https://kenimoto.dev/blog/multi-agent-decision-fatigue-412-to-38/) — 2026-06-25
- [I Added 3 Numbers to One Paragraph. Perplexity Started Citing It in 11 Days.](https://kenimoto.dev/blog/perplexity-cited-3-numbers-11d/) — 2026-06-24
- [AI Mode Just Hit 1 Billion Users, and Opened a Local-Business LLMO Market Most Engineers Are Ignoring](https://kenimoto.dev/blog/ai-mode-billion-users-local-business-llmo/) — 2026-06-23
- [I Wired My Pages Into Topic Hubs, Not a Flat List: AI Citations Consolidated Onto 4 of Them](https://kenimoto.dev/blog/topic-hubs-ai-citations/) — 2026-06-23
- [AI Search Is Under 1% of My Traffic and 12% of My Signups. That's the LLMO Case I Actually Use.](https://kenimoto.dev/blog/llmo-roi-23x-conversion/) — 2026-06-20
- [GEO's +115% From Statistics Is Domain-Dependent: It Worked for My Tech Posts and Did Nothing for My How-To Pages](https://kenimoto.dev/blog/geo-stats-domain-dependent/) — 2026-06-19
- [AI Reads Your Chunks, Not Your Page: I Promoted 9 Sections from H3 to H2 and Watched Which Ones Got Quoted](https://kenimoto.dev/blog/ai-reads-chunks-not-pages/) — 2026-06-18
- [I Wired Claude Code to Real Hardware Over USB Serial. The MCP Tool Was the Easy Part.](https://kenimoto.dev/blog/claude-code-real-hardware-usb/) — 2026-06-17
- [I Checked What GPTBot Actually Sees on My JS-Rendered Pages. It Was an Empty `<div>`.](https://kenimoto.dev/blog/ai-crawlers-dont-execute-javascript-invisible-pages/) — 2026-06-16
- [AI Wrote 100 Passing Tests. Mutation Testing Says They Caught 58% of Real Bugs.](https://kenimoto.dev/blog/ai-100-tests-mutation-score/) — 2026-06-15
- [AI Finds Your Page Three Ways. I Published the Same Fact in All Three and Timed Which Reached AI First.](https://kenimoto.dev/blog/ai-finds-your-page-three-ways/) — 2026-06-14
- [Claude Code Pricing vs API vs Local GPU in 2026](https://kenimoto.dev/blog/ai-agent-monthly-cost-api-subscription-local-breakeven/) — 2026-06-13
- [Multilingual LLMO: 4 Languages, Wrong Citation](https://kenimoto.dev/blog/ai-cites-wrong-language-version-multilingual-llmo/) — 2026-06-12
- [I Rewrote 12 Pages to Answer the Question in the First Sentence. AI Started Quoting 7 of Them.](https://kenimoto.dev/blog/answer-first-7-of-12-cited/) — 2026-06-11
- [I Rank #1 on Google. On Brave I'm Page 5. My Own AI Agents Can't Find Me.](https://kenimoto.dev/blog/brave-invisible-to-ai-agents/) — 2026-06-10
- [Perplexity Citations Exploded After I Changed 3 Things. Only 1 Was Schema.](https://kenimoto.dev/blog/perplexity-3-changes-1-schema/) — 2026-06-09
- [I Gave Every Page on My Site a .md Twin. The AI Fetchers Stopped Guessing](https://kenimoto.dev/blog/every-page-md-twin-llmo/) — 2026-06-08
- [My Best Page Went Stale in a Month: Why AI Search Rewards Freshness, Not Just Schema](https://kenimoto.dev/blog/ai-search-rewards-freshness/) — 2026-06-07
- [AI Search Splits Your One Question Into Six. My Pages Answered None of Them.](https://kenimoto.dev/blog/query-fanout-ai-citations/) — 2026-06-06
- [I Stopped Adding Context to My Agent and Pruned Tool Outputs Instead — My 3-Hour Task Stopped Forgetting Its Own Plan](https://kenimoto.dev/blog/stopped-adding-context-pruned-tool-outputs-accuracy-returned/) — 2026-06-05
- [AI Citation Decay: What It Is, Measured 9 Weeks](https://kenimoto.dev/blog/ai-citations-half-life-decay/) — 2026-06-04
- [Blast Radius: The File That Breaks Is 2 Hops Out](https://kenimoto.dev/blog/codebase-graph-two-hops/) — 2026-06-03
- [Link-less Brand Mentions Beat Backlinks for AI Visibility — I Read the Ahrefs 75,000-Brand Study So You Don't Have To](https://kenimoto.dev/blog/mentions-beat-backlinks-ai/) — 2026-06-02
- [Your Page Rank Is Invisible to AI — Only Your Passages Get Cited](https://kenimoto.dev/blog/passage-rank-beats-page-rank-ai-citations/) — 2026-06-01
- [I Crosspost to 4 Platforms with rel=canonical Pointing Home. AI Search Still Picks the Copy.](https://kenimoto.dev/blog/crosspost-canonical-ai-picks-the-copy/) — 2026-05-31
- [Claude Code Skills Cost Tokens Even When They Don't Fire. I Measured 5 Skills Across 7 Hours. The Bill Was 18%.](https://kenimoto.dev/blog/skills-loaded-3-never-fired-18/) — 2026-05-30
- [I Cron-Scheduled 7 AI Agents. 2 Silently Failed for 18 Days. Tracing Wouldn't Have Caught It.](https://kenimoto.dev/blog/seven-cron-agents-18d-silent/) — 2026-05-28
- [I Ran 3 Claude Code Sessions in Parallel for 8 Hours. They Overwrote Each Other's Context Twice.](https://kenimoto.dev/blog/three-claude-sessions-parallel-8h-context-overwrite/) — 2026-05-27
- [I Asked 5 AI Search Engines to Cite My Own Blog. Only 3 of 31 Articles Showed Up.](https://kenimoto.dev/blog/five-ai-engines-cite-my-blog-three-of-thirty-one/) — 2026-05-26
- [I Added 11 JSON-LD Schemas. Three Months Later, Only 3 Showed Up in AI Citations.](https://kenimoto.dev/blog/11-json-ld-3-cited-by-ai/) — 2026-05-25
- [I Refactored 100 Functions With Claude. 7 Got Slower in Production.](https://kenimoto.dev/blog/claude-refactor-100-functions-7-slower-production/) — 2026-05-24
- [Claude Code Hooks for TDD: 4 of 10 to 9 of 10](https://kenimoto.dev/blog/claude-code-tdd-test-after-code-six-of-ten/) — 2026-05-23
- [I Added a 4th Agent That Audits My Other Agents. It Caught My Strategist Procrastinating for 3 Weeks.](https://kenimoto.dev/blog/evolver-fourth-agent-caught-strategist-procrastinating/) — 2026-05-22
- [I Translated My Blog Into 4 Languages. Portuguese Got Nearly 4× the Traffic of English.](https://kenimoto.dev/blog/four-languages-thirty-days-portuguese-four-x-traffic/) — 2026-05-21
- [TRM 8,337% LLMO Playbook: 1 of 4 Pillars Worked](https://kenimoto.dev/blog/trm-8337-percent-llmo-pillars-indie-test/) — 2026-05-20
- [Claude Said 'You're Absolutely Right!' 47 Times Last Week. I Was Only Right 11 Times. Claude Was Wrong 36.](https://kenimoto.dev/blog/claude-sycophancy-47-times-measured/) — 2026-05-19
- [I Plugged the Same Site Into 7 AI-Citation Trackers. They Reported 7 Different Numbers.](https://kenimoto.dev/blog/seven-ai-citation-trackers-seven-different-numbers/) — 2026-05-18
- [The 5 AI Crawlers That Hit My Sites Most in 30 Days — What Their Logs Told Me About LLMO](https://kenimoto.dev/blog/five-ai-crawlers-hit-my-site-30-days/) — 2026-05-17
- [I Plugged Claude into a Chaos Engineering MCP Server. It Killed Staging 4 Times Before Finding a Bug We'd Missed for 6 Months.](https://kenimoto.dev/blog/claude-chaos-engineering-mcp-killed-staging-4-times/) — 2026-05-16
- [I Caught Claude Hiding My Bug 3 Times in a Row. Then I Turned 10 Debugging Habits Into Prompts.](https://kenimoto.dev/blog/claude-hid-my-bug-three-times-ten-debugging-prompts/) — 2026-05-15
- [I Gave My Strategist Agent WebSearch. 5 Topics Took 20 Minutes. Splitting It Into 3 Made It 3.](https://kenimoto.dev/blog/three-role-separation-observer-strategist-marketer/) — 2026-05-14
- [I Benchmarked 5 Voice AI Stacks. Only 2 Stayed Under 300ms.](https://kenimoto.dev/blog/five-voice-ai-stacks-only-two-under-300ms/) — 2026-05-13
- [Claude Code Sub-Agents Disagreed on 40% of My PR](https://kenimoto.dev/blog/three-sub-agents-reviewed-same-pr-40-percent-disagreement/) — 2026-05-12
- [llms.txt Audited: 30 Files, 5 Anti-Patterns](https://kenimoto.dev/blog/30-llms-txt-files-5-anti-patterns-already-forming/) — 2026-05-11
- [OpenClaw Hit 250K Stars Faster Than React. I Spent a Day Switching From Claude Code](https://kenimoto.dev/blog/openclaw-vs-claude-code-24h/) — 2026-05-10
- [I Refused to Write Specs Until Claude Code Generated Wrong Code Three Times](https://kenimoto.dev/blog/spec-driven-development-claude-code-three-failures/) — 2026-05-09
- [I Let My Claude Code Agent Run for 24 Hours. The $400 Bill Was the Least Scary Part.](https://kenimoto.dev/blog/autonomous-agent-24-hours-security-lessons/) — 2026-05-08
- [The og:type Bug Three of My Astro Sites Quietly Shipped](https://kenimoto.dev/blog/og-type-double-emit-three-astro-sites/) — 2026-05-08
- [I Stacked 4 More Context Layers on Top of RAG. The Improvement Was 12%.](https://kenimoto.dev/blog/full-context-engineering-rag-80-percent/) — 2026-05-07
- [arXiv 2603.25723 \"Natural-Language Agent Harnesses\": 4 Patterns + 3 Anti-Patterns After 12 Weeks in Prod](https://kenimoto.dev/blog/natural-language-agent-harnesses-arxiv/) — 2026-05-06
- [Claude Code Agents vs Skills: 47 Prompts Later](https://kenimoto.dev/blog/claude-code-skills-reusable-workflow-pattern/) — 2026-05-05
- [Your New Domain's First Week of GA4 Is a Lie: 4 Days of Raw Data from kaoriq.com's Launch](https://kenimoto.dev/blog/new-domain-first-week-ga4-is-a-lie/) — 2026-05-05
- [ChatGPT Codex vs Claude Code Pricing: 3.4x/PR](https://kenimoto.dev/blog/claude-code-vs-chatgpt-codex-official-agents/) — 2026-05-04
- [One Question, Five AI Search Engines, Five Different Answers](https://kenimoto.dev/blog/five-ai-search-engines-architecture-llmo/) — 2026-05-03
- [Is AI Actually Citing Your Site? How to Measure What Google Rankings Can't](https://kenimoto.dev/blog/measure-ai-citations-llmo-kpi/) — 2026-05-02
- [Generative Engine Optimization: 3 of 9 Worked](https://kenimoto.dev/blog/geo-princeton-study-9-ways-ai-cites-you/) — 2026-05-01
- [9 Bugs in My AI Pipeline: None Were the AI's Fault](https://kenimoto.dev/blog/9-bugs-in-my-ai-pipeline/) — 2026-04-30
- [Claude Haiku + RAG Beat Sonnet 11.8 to 5.3](https://kenimoto.dev/blog/cheap-model-won-context-beats-parameters/) — 2026-04-30
- [llms.txt: The File That Decides Whether AI Can Find Your Site](https://kenimoto.dev/blog/llms-txt-ai-find-your-site/) — 2026-04-30
- [Building an Autonomous Content Pipeline with Claude Code](https://kenimoto.dev/blog/hello-world/) — 2026-04-29

### Japanese (日本語)

- [/compactを7日計測、Opus4.7で0発火](https://kenimoto.dev/ja/blog/compact-opus-47-7days-0-fires/) — 2026-10-12
- [マルチエージェントは15倍トークン食う: Anthropic 公式が1行で公表した現実](https://kenimoto.dev/ja/blog/multi-agent-15x-tokens-anthropic-supervisor/) — 2026-10-11
- [AIエージェントのリトライが本番を落とす3つの罠](https://kenimoto.dev/ja/blog/ai-agent-retry-production-3-traps/) — 2026-10-09
- [ハーネスエンジニアリング: 4社4定義の違い](https://kenimoto.dev/ja/blog/harness-engineering-4-definitions-comparison/) — 2026-10-08
- [音声AIを体感350msに縮める7つの知覚ハック](https://kenimoto.dev/ja/blog/voice-ai-350ms-7-perception-hacks/) — 2026-10-07
- [Ollama vs llama.cpp 2.8倍](https://kenimoto.dev/ja/blog/ollama-vs-llama-cpp-engine-choice-rtx-4070/) — 2026-10-06
- [strategy.mdをagentが守る3つの型](https://kenimoto.dev/ja/blog/strategy-md-agent-compliance-rate/) — 2026-10-05
- [見積りを壊す3つのバイアス:AIも私も外す](https://kenimoto.dev/ja/blog/3-estimation-biases-ai-10-minutes/) — 2026-10-04
- [タイトル長さ監査、全525本の74%が上限超過](https://kenimoto.dev/ja/blog/title-lint-audit-525-articles-74-percent/) — 2026-10-03
- [GraphRAG 7ステップで効くのは3つだけ](https://kenimoto.dev/ja/blog/graphrag-7steps-which-3-matter/) — 2026-10-01
- [GraphRAGでコードレビューを7段階に分解](https://kenimoto.dev/ja/blog/code-knowledge-graph-mcp-pr-review-7-steps/) — 2026-09-30
- [Claude Code TDD hook 4例強制](https://kenimoto.dev/ja/blog/claude-code-tdd-hook-enforcement/) — 2026-09-29
- [Jevとは？話題の理由5つとLLM判定との差](https://kenimoto.dev/ja/blog/jev-typesafe-why-trending-vs-llm-as-a-judge/) — 2026-09-28
- [ナレッジグラフは3ヶ月で腐る 保守で見た3パターン](https://kenimoto.dev/ja/blog/knowledge-graph-3-months-decay-3-patterns/) — 2026-09-28
- [autoFixable: 900ルールで前段一掃](https://kenimoto.dev/ja/blog/autofixable-pattern-ruff-900-rules-before-review/) — 2026-09-26
- [WSL2 Tailscale SSHがKEXで停止](https://kenimoto.dev/ja/blog/wsl2-tailscale-ssh-kex-stall-mtu-1280/) — 2026-09-26
- [音声AIの割り込み: VAD限界とKrisp 6M](https://kenimoto.dev/ja/blog/voice-ai-interruption-vad-limit-krisp-6m/) — 2026-09-21
- [個人ナレッジグラフの盲点: Neo4jで橋渡し検出](https://kenimoto.dev/ja/blog/neo4j-infranodus-personal-kg-blind-spots/) — 2026-09-20
- [プロンプトの精度差4.6倍から得た3つの示唆](https://kenimoto.dev/ja/blog/prompt-quality-gap-4-6x-3-lessons/) — 2026-09-20
- [日本語Realtime APIが800msに寄る訳](https://kenimoto.dev/ja/blog/openai-realtime-api-japanese-800ms-vad-tuning/) — 2026-09-18
- [ハーネスエンジニアリング、Fowler実物の3節](https://kenimoto.dev/ja/blog/harness-engineering-fowler-3-sections-existing-repo-audit/) — 2026-09-15
- [Agentic RAGで11.8点:検索を選ばせる](https://kenimoto.dev/ja/blog/agentic-rag-haiku-rag-11-8/) — 2026-09-13
- [GraphRAG論文の72-83%勝率と4段階](https://kenimoto.dev/ja/blog/microsoft-graphrag-paper-c0-c3/) — 2026-09-11
- [AIクローラーは通すbot対策、WAF2段の設計](https://kenimoto.dev/ja/blog/cloudflare-managed-challenge-keep-ai-crawlers/) — 2026-09-09
- [コードをKG化する7ステップと6ツール比較](https://kenimoto.dev/ja/blog/code-kg-7steps-6tools/) — 2026-09-09
- [Kindleが売れない原因を65冊で測り3回外した](https://kenimoto.dev/ja/blog/kindle-sales-cause-65-books-3-misses/) — 2026-09-09
- [GraphRAG企業導入で詰まる3つの壁](https://kenimoto.dev/ja/blog/graphrag-enterprise-3-walls/) — 2026-09-08
- [CLAUDE.md 3階層と「なぜ」契約](https://kenimoto.dev/ja/blog/claude-md-3-scopes-naze-keiyaku/) — 2026-09-07
- [音声AI遅延の3つの崖 200/300/800ms](https://kenimoto.dev/ja/blog/voice-ai-three-cliffs-200-300-800ms/) — 2026-09-06
- [Claude Code Skills 発火率10%](https://kenimoto.dev/ja/blog/claude-code-skills-firing-10-of-101/) — 2026-09-05
- [AIコードレビュー、Diffで依存を見逃す実測](https://kenimoto.dev/ja/blog/ai-code-review-diff-only-dependency-blindspot/) — 2026-09-04
- [MCP tool descriptionの穴7本](https://kenimoto.dev/ja/blog/mcp-description-lint-slipped-invariant-poc/) — 2026-09-04
- [sudoers引数制限の4段階:*は空白を跨ぐ](https://kenimoto.dev/ja/blog/sudoers-narrow-args-4-steps-wildcard-trap/) — 2026-09-03
- [アクセス2.8倍の正体、クローラー3種の見分け方](https://kenimoto.dev/ja/blog/crawler-three-types-user-agent-check/) — 2026-09-02
- [検索ボリューム調べ方: サジェスト787件実測](https://kenimoto.dev/ja/blog/google-search-demand-suggest-echo-787-queries/) — 2026-09-02
- [音声AIの525ms壁を300ms未満にする3手法](https://kenimoto.dev/ja/blog/voice-ai-cascade-525ms-streaming-3-methods/) — 2026-09-02
- [Tailscale CI: Piデプロイ権限3層](https://kenimoto.dev/ja/blog/tailscale-github-actions-pi-deploy-3-layer-permissions/) — 2026-09-01
- [Claude Codeサブエージェントの3階層権限](https://kenimoto.dev/ja/blog/claude-code-sub-agent-3-tier-permissions/) — 2026-08-31
- [Zenn全記事noindex、気づくまで8週間](https://kenimoto.dev/ja/blog/zenn-noindex-8-weeks-unnoticed/) — 2026-08-31
- [PRラベルnit/must機械分別で48h→24h](https://kenimoto.dev/ja/blog/conventional-comments-nit-must-48h-24h-ja/) — 2026-08-30
- [KDP 7枠をBing需要で選んで外した実測](https://kenimoto.dev/ja/blog/kdp-keyword-bing-demand-miss-amazon-suggest/) — 2026-08-30
- [Google Play Booksの20%公開義務](https://kenimoto.dev/ja/blog/play-books-20-percent-preview-and-ai-training/) — 2026-08-29
- [voice AI TTFBを100msに詰めた実測](https://kenimoto.dev/ja/blog/voice-ai-ttfb-100ms-300ms-wall/) — 2026-08-29
- [GraphRAGでLinkedIn 28.6%短縮](https://kenimoto.dev/ja/blog/graphrag-linkedin-28-6/) — 2026-08-28
- [Whisper文字起こしを使わない歌詞同期: 実測誤差0.12秒](https://kenimoto.dev/ja/blog/lyrics-timing-forced-alignment-not-asr/) — 2026-08-27
- [PINとパスワードの違いは桁数ではない: 6桁で足りる理由](https://kenimoto.dev/ja/blog/pin-vs-password-not-about-length/) — 2026-08-26
- [技術書116冊のA+コンテンツを数えたら、個人出版は53冊中11冊しか入れていなかった](https://kenimoto.dev/ja/blog/aplus-content-116-technical-books/) — 2026-08-25
- [AIっぽい表現1位は「適切な」180記事](https://kenimoto.dev/ja/blog/ukibori-zero-in-180-ai-vocabulary-measured/) — 2026-08-25
- [MCP承認ゲートを3層に分けた: allow/ask/denyの運用ログ](https://kenimoto.dev/ja/blog/mcp-approval-gate-3-layers-allow-ask-deny/) — 2026-08-23
- [Claude Code Skills に入れると壊れる処理3種類と、分離先の選び方](https://kenimoto.dev/ja/blog/claude-code-skills-mcp-hook-boundary/) — 2026-08-22
- [音声AIの barge-in 実装で認識精度 22% 減 — VAD 閾値 4 段階実測](https://kenimoto.dev/ja/blog/voice-ai-barge-in-22-percent-accuracy-drop-vad-4-thresholds/) — 2026-08-19
- [「AI経由の流入が増えた」は自分で作れる。導線を入れる前に測り方を分けた](https://kenimoto.dev/ja/blog/ai-referral-self-inflation-measure-first/) — 2026-08-17
- [AI時代の新バイアス3選 — 自動化バイアスで見逃したバグと3ヶ月の実測対策](https://kenimoto.dev/ja/blog/ai-automation-bias-3-months-3-bugs/) — 2026-08-12
- [採譜は動いた。伴奏にはならなかった — OSSでカラオケ音源を作れなかった記録](https://kenimoto.dev/ja/blog/karaoke-oss-transcription-worked-arrangement-missing/) — 2026-08-11
- [CLAUDE.mdの作者は2行、実践者は100行 — 情報密度の非対称論](https://kenimoto.dev/ja/blog/claude-md-2-vs-100-lines-jouhou-mitsudo/) — 2026-08-10
- [MCP Sampling権限昇格を止める防御実装3パターン - Claude Desktopで実測](https://kenimoto.dev/ja/blog/mcp-sampling-defense-3-patterns-claude-desktop-jissoku/) — 2026-08-09
- [OpenVPNはなぜ.ovpnファイルとパスワードだけでVPNに繋がるのか](https://kenimoto.dev/ja/blog/openvpn-config-file-and-authentication-flow/) — 2026-08-06
- [Tailscale・WireGuardを3層比較](https://kenimoto.dev/ja/blog/wireguard-openvpn-tailscale-protocol-layers/) — 2026-08-06
- [Claude Codeの認証は2層構造になっている。setup-tokenとloginは別物だ](https://kenimoto.dev/ja/blog/claude-code-two-layer-auth-setup-token/) — 2026-08-05
- [npmサプライチェーン攻撃Shai-Hulud：署名検証もSBOMも無意味だった理由](https://kenimoto.dev/ja/blog/shai-hulud-npm-supply-chain-why-verification-failed/) — 2026-08-05
- [Claude Codeに図を「描かせる」のをやめて「一緒に描く」に変えたら、図解の意味が変わった](https://kenimoto.dev/ja/blog/claude-code-shared-canvas-diagramming/) — 2026-08-04
- [Claude Code 実戦運用ガイド ― エージェントを2年回して分かった8つの勘所](https://kenimoto.dev/ja/blog/claude-code/) — 2026-08-04
- [静的解析コードKGが落とす61%の動的呼び出し](https://kenimoto.dev/ja/blog/kg-61-percent-miss-dynamic-call-payment-rollback/) — 2026-08-04
- [RTX 4070 一枚で BGM・効果音・画像を全部ローカル生成する — ACE-Step 1.5 と MOSS-SoundEffect v2.0 を一晩で実戦投入した話](https://kenimoto.dev/ja/blog/rtx4070-local-bgm-sfx-ace-step-moss-soundeffect/) — 2026-08-03
- [Skill・Hook・MCPで同じ機能を3実装した結果、月額トークン費が家賃並みに割れた](https://kenimoto.dev/ja/blog/skill-hook-mcp-3-jissou-getsugaku-yachin/) — 2026-08-03
- [AIコードレビューは説得の心理学:3つの言い回しでmerge速度が1.5倍になった話](https://kenimoto.dev/ja/blog/ai-code-review-persuasion-psychology-1-5x-merge-speed/) — 2026-08-02
- [Claude Code の auto mode で Bash だけが止まる — safety classifier 障害の調査と「故障時だけ fail-open」hook 案](https://kenimoto.dev/ja/blog/claude-code-auto-mode-classifier-fail-open-hook/) — 2026-08-02
- [LLMOだけでは読まれない。llms.txt完備のサイトで、AI経由の流入は3%だった](https://kenimoto.dev/ja/blog/llmo-only-3-percent-ai-referral-jissoku/) — 2026-08-02
- [チャットモデルは損失マスクで生まれる — nanochatのSFTを「採点表」として読む](https://kenimoto.dev/ja/blog/nanochat-sft-loss-mask-chat-model/) — 2026-08-02
- [Tauri v2 を「ガワ」として使う — WSL2 mirrored networking で Node.js サーバをそのまま Windows デスクトップに映す](https://kenimoto.dev/ja/blog/tauri-v2-gawa-wsl-mirrored-networking/) — 2026-08-02
- [熊本M7.1の余震346件を5日間追った — 福岡に届いたのは34件、そして10年前との答え合わせ](https://kenimoto.dev/ja/blog/uto-m71-yoshin-346-fukuoka-34-jissoku/) — 2026-08-01
- [地震予知はなぜ無理で、余震予測はなぜできるのか — 能登の余震130件で実測](https://kenimoto.dev/ja/blog/jishin-yochi-vs-yoshin-yosoku-jissoku/) — 2026-07-29
- [Shai-Hulud事件でCLAUDE.mdの書き方が変わった話](https://kenimoto.dev/ja/blog/shai-hulud-claude-md-3-removed-5-added/) — 2026-07-26
- [Claude Code MCPでハードウェア制御——UART/I2C/SPI/RS-485/CAN 5プロトコル実測](https://kenimoto.dev/ja/blog/claude-code-mcp-hardware-5-jissoku/) — 2026-07-24
- [MEO代行費5万円をClaudeで自前化、3業種プロンプト実物と月3万円分解](https://kenimoto.dev/ja/blog/meo-daiko-claude-jimae-3gyoushu/) — 2026-07-21
- [MCP OWASP Top 10全10項目、Claude Desktop標準設定で何が守れて何が抜けるか](https://kenimoto.dev/ja/blog/mcp-owasp-top10-claude-desktop-defense-map-2026-07/) — 2026-07-20
- [MCP Sampling 権限昇格 — Claude Desktop で通る3経路 (7攻撃検証)](https://kenimoto.dev/ja/blog/mcp-sampling-privilege-escalation-claude-desktop-3-paths/) — 2026-07-19
- [「リズムの単調化は日本語だけか」を英語・ポルトガル語×7モデルで検証したら、70セル全部でAIは単調だった](https://kenimoto.dev/ja/blog/machine-accent-3-languages-70-70-cells/) — 2026-07-18
- [ローカルLLMを2フラグで34.6 tok/s](https://kenimoto.dev/ja/blog/rtx4070-qwen35b-2flags-kv-quant-benchmark/) — 2026-07-18
- [AI臭のリズム検証 7モデルでAUC 0.998](https://kenimoto.dev/ja/blog/japanese-ai-smell-vocabulary-vs-rhythm-fingerprints/) — 2026-07-17
- [nanochatとは: GPT-2を$50で訓練](https://kenimoto.dev/ja/blog/nanochat-gpt2-training-cost-2026/) — 2026-07-17
- [論文の実測分布で校正したリズム計測CLI「rhythm-lens」を公開して、自分の記事に当てたら⚠2つ食らった](https://kenimoto.dev/ja/blog/rhythm-lens-cli-own-article-2-warnings/) — 2026-07-17
- [1Passwordを値上げで諦めた私が、SOPS+ageで.envの複数PC共有に着地した話](https://kenimoto.dev/ja/blog/sops-age-multi-pc-env-envelope-model/) — 2026-07-16
- [複数アカウントで気づいた Claude Code の2ファイル認証 (credentials.json + ~/.claude.json)](https://kenimoto.dev/ja/blog/claude-code-two-file-auth-multi-account/) — 2026-07-15
- [MCPのTool descriptionはプロンプトインジェクションの窓 - 公開5個で試した3種](https://kenimoto.dev/ja/blog/mcp-tool-description-injection-5-servers-3-patterns/) — 2026-07-15
- [Strangler Fig で React/FastAPI 段階移行](https://kenimoto.dev/ja/blog/strangler-fig-react-fastapi-4phase/) — 2026-07-15
- [YAML→PNG を LLM に任せる — historymap に MCP サーバーを足した記録](https://kenimoto.dev/ja/blog/historymap-mcp-server-yaml-to-png/) — 2026-07-14
- [MCPのファイル添付7サービス実測: 対応ゼロ](https://kenimoto.dev/ja/blog/mcp-file-upload-7-services-benchmarked-full-support-zero/) — 2026-07-14
- [X Premium ¥980 では Grok を CLI から叩けない — 403 の理由と $25 credit で始める最短路](https://kenimoto.dev/ja/blog/x-premium-980-grok-cli-403-and-cheapest-path/) — 2026-07-14
- [AGENTS.md と CLAUDE.md の違いは「専用機か汎用機か」ではない — 3ヶ月使い分けて分かった実践設計の分岐点](https://kenimoto.dev/ja/blog/agents-md-claude-md-practical-design-branch/) — 2026-07-13
- [自作LLMOサイト、日本人1人運営なのに中国からのアクセスが45%になっていた — cn.bing で自然発生的にランクした話](https://kenimoto.dev/ja/blog/llmoframework-cn-bing-china-45pct-organic-discovery/) — 2026-07-13
- [OpenCut MCP: Claude Code から動画エディタを操作する（4つの罠）](https://kenimoto.dev/ja/blog/opencut-classic-mcp-4-traps-editor-core-fork/) — 2026-07-13
- [OpenCutセットアップ難民は、まずopencut.appを開いてください — ローカル起動の3つの罠](https://kenimoto.dev/ja/blog/opencut-setup-nanmin-webapp-first-3-traps/) — 2026-07-13
- [MCPサーバーを公開する前のpre-flight — 4層で採点する mcp-scorecard の設計](https://kenimoto.dev/ja/blog/pre-flight-your-mcp-4-layers-scorecard/) — 2026-07-13
- [AIエージェントのハーネスを構成する6要素 — 10分でCLAUDE.mdを棚卸しするチェックリスト](https://kenimoto.dev/ja/blog/harness-6-components-claude-md-10min-checklist/) — 2026-07-12
- [107本のSKILL.mdをlintしたら、稼働中のskillが21本「壊れてる」と言われた](https://kenimoto.dev/ja/blog/skill-eval-two-layer-107-lint/) — 2026-07-12
- [プロダクトを出すたび支援導線が勝手に付く: Cloudflare Workersのエッジ注入オーバーレイ](https://kenimoto.dev/ja/blog/edge-injected-support-overlay-cloudflare-workers/) — 2026-07-11
- [FUNDING.ymlを置いてもSponsorボタンは出ない: 42リポジトリを棚卸しした](https://kenimoto.dev/ja/blog/github-sponsor-button-42-repo-audit/) — 2026-07-11
- [YAMLを1枚書くと企業サイト風の年表ページになるOSS「historymap」を公開した](https://kenimoto.dev/ja/blog/historymap-yaml-timeline-oss/) — 2026-07-11
- [自己進化エージェントを3ヶ月動かしたら2回ロールバックした](https://kenimoto.dev/ja/blog/self-evolving-agent-3-months-2-rollbacks/) — 2026-07-11
- [MCPで私が踏んだ地雷7つ — 実装と実測ログ](https://kenimoto.dev/ja/blog/mcp-7-mines-implementation-log/) — 2026-07-10
- [MCPサーバーの権限昇格を7パターン実演、うち3つはClaude Desktopの標準設定で通った](https://kenimoto.dev/ja/blog/mcp-owasp-privilege-escalation-7-attacks-claude-desktop/) — 2026-07-09
- [compact-plus 派生 v0.2 リリース](https://kenimoto.dev/ja/blog/compact-ops-dogfooding-67-percent-warn/) — 2026-07-07
- [Claude Code 3セッション並走で8時間ぶん消えた: あの日から追加した5つのハーネス制約](https://kenimoto.dev/ja/blog/three-sessions-5-harness-rules/) — 2026-07-06
- [Knowledge Graphを7ステップで作る前に、AIエージェントが「見えていない」3層を可視化する](https://kenimoto.dev/ja/blog/knowledge-graph-3-blind-layers-before-7-steps/) — 2026-07-05
- [llms.txtとJSON-LDが噛み合っていない5パターン — 30ファイルを整合チェックして分かったこと](https://kenimoto.dev/ja/blog/llms-txt-json-ld-integrity-5-patterns/) — 2026-07-02
- [LLMO 総合ガイド ― AI 検索に引用される技術を、30ヶ月分の計測から組み立てる](https://kenimoto.dev/ja/blog/llmo/) — 2026-06-29
- [Ollama 12.2 tok/s が34.6 tok/s になった: RTX 4070 で Qwen3 35B を2.8倍速にする1コマンド](https://kenimoto.dev/ja/blog/ollama-qwen3-35b-rtx-4070-cpu-moe-2-8x/) — 2026-06-28
- [Claude Code・Cursor・Codexを1日並列で走らせて、判断回数を数えたら412回だった](https://kenimoto.dev/ja/blog/claude-cursor-codex-heiretsu-handan-412kai/) — 2026-06-27
- [Claude Code・Cursor・Codex併用で1日の判断を400回から40回にした、ハーネス3層設計の実測](https://kenimoto.dev/ja/blog/harness-3-layers-judgment-count-one-tenth/) — 2026-06-26
- [GraphRAGはRDFかプロパティグラフか3問](https://kenimoto.dev/ja/blog/graphrag-2026-rdf-vs-property-graph-3-questions/) — 2026-06-25
- [Sonnet 4は確信度4.2で嘘をついた、Haikuは「知らない」と答えた — 大きいモデルほど自信を持って幻覚する](https://kenimoto.dev/ja/blog/sonnet-4-confidence-4-2-haiku/) — 2026-06-24
- [Claude Code の Plan モードで「朝に一日を設計」したら、手戻りが半分になった](https://kenimoto.dev/ja/blog/plan-mode-morning-rework/) — 2026-06-23
- [却下ログは「学習信号」だった — 自己進化ハーネスが strategy.md に書けない暗黙知を拾う仕組み](https://kenimoto.dev/ja/blog/rejection-log-learning-signal/) — 2026-06-23
- [「著者エンティティを実装した」と書いた私が、自分の239ページを監査したら配線が途中で切れていた話](https://kenimoto.dev/ja/blog/author-entity-self-audit/) — 2026-06-22
- [賢いモデルが来るほど、ハーネスは「捨てる」設計になる — Anthropicがマルチからシングルに戻した理由](https://kenimoto.dev/ja/blog/smarter-model-thinner-harness/) — 2026-06-20
- [マルチエージェントは並列化で速くなる、は幻想だった — 調整コストはエージェント数の二乗で増える](https://kenimoto.dev/ja/blog/multi-agent-coordination-n2/) — 2026-06-19
- [Figma MakeとClaude Codeで、デザイナー不在のUIを作る — コードからデザインを逆生成する時代](https://kenimoto.dev/ja/blog/figma-make-claude-code-design/) — 2026-06-18
- [記事にArticle schemaを入れてもAIは私を著者と認識しなかった。著者エンティティを実装して引用の帰属が変わった話](https://kenimoto.dev/ja/blog/author-entity-ai-citation/) — 2026-06-17
- [GPTBotに私のJSレンダリングページを見せたら、中身は空の`<div>`だった](https://kenimoto.dev/ja/blog/ai-crawler-javascript-jikkou-shinai-csr-invisible/) — 2026-06-16
- [LLMO対応はテストしないと静かに壊れる：AIクローラーの可読性をPlaywrightでCI検証する](https://kenimoto.dev/ja/blog/llmo-playwright-ci-test/) — 2026-06-15
- [賢いモデルほど、上手に嘘をつく](https://kenimoto.dev/ja/blog/bigger-models-lie-better/) — 2026-06-14
- [「いい感じに作って」はなぜ5回目で崩壊するのか -- コードの前に書く6つの定義書と3つの記憶ファイル](https://kenimoto.dev/ja/blog/document-first-claude-code-6-docs-3-memory-files/) — 2026-06-13
- [AIエージェントが「空気を読めない」本当の理由:常識ナレッジグラフという欠けたレイヤー](https://kenimoto.dev/ja/blog/agent-commonsense-knowledge-graph-missing-layer/) — 2026-06-12
- [接続したMCPサーバーの38%は認証なしだった — 60日で30件のCVEが出たMCPセキュリティの現実](https://kenimoto.dev/ja/blog/mcp-38-percent-no-auth/) — 2026-06-11
- [43秒の異常が24時間の障害になった: メトリクスの急変と漸増を見分ける時系列デバッグ](https://kenimoto.dev/ja/blog/spike-vs-gradient-time-series-debug/) — 2026-06-10
- [AIエージェントの予測を1つの数字で信じるな — Simulatorは点推定でなく確率分布を返すべき](https://kenimoto.dev/ja/blog/simulator-distribution-not-point-estimate/) — 2026-06-09
- [サブエージェントにメインの記憶を渡すのは事故だった：7ファイルのうち4つは遮断すべき](https://kenimoto.dev/ja/blog/sub-agent-memory-isolation/) — 2026-06-08
- [バリデータがバグっていた — 「安全です」と報告するシステム自身が壊れているとき](https://kenimoto.dev/ja/blog/validator-was-the-bug-grey-failure/) — 2026-06-07
- [「とりあえず全部許可」でClaude Codeを動かすと、.envの秘密がそのままAnthropicに渡る話](https://kenimoto.dev/ja/blog/claude-code-deny-rules-env/) — 2026-06-06
- [店舗LLMOの答えはllms.txtでなくGBPだった — 私が綺麗に書いたファイルを、AIは一度も引用しなかった](https://kenimoto.dev/ja/blog/llmo-local-business-gbp-not-llmstxt/) — 2026-06-05
- [Claude Codeに渡すコンテキストを足すのをやめた — ツール出力を間引いたら長時間タスクの精度が戻った](https://kenimoto.dev/ja/blog/stopped-adding-context-pruning-recovered-accuracy/) — 2026-06-04
- [「稼働率99.5%」と「5,000件の決済失敗」は同じ事実 — 障害報告のフレーミングが緊急度を反転させる](https://kenimoto.dev/ja/blog/incident-framing-99-percent/) — 2026-06-03
- [午後3時、あなたのコードレビュー承認率はほぼ0%になる：決断疲労の科学とAIエージェント時代の処方箋](https://kenimoto.dev/ja/blog/decision-fatigue-3pm-code-review-approval/) — 2026-06-02
- [ページ単位のSEOはAIに見えていない: 引用されるのは「パッセージ」だった](https://kenimoto.dev/ja/blog/passage-design-ai-citation-not-page-rank/) — 2026-06-01
- [Claude Codeを/clearせず9時間動かした日、コンテキストはどこで腐り始めたのか](https://kenimoto.dev/ja/blog/claude-code-9h-context-rot-token/) — 2026-05-31
- [Claude Code Skills は発火しなくてもトークンを食う — 5 Skills × 7時間セッションで「使われなかった」3つが18%食べていた](https://kenimoto.dev/ja/blog/skills-loaded-3-never-fired-18/) — 2026-05-30
- [AIエージェントを7本cronで毎日回したら、2本が18日間沈黙していた — observabilityでは拾えず、exit-code契約で拾えた話](https://kenimoto.dev/ja/blog/seven-cron-agents-18d-silent/) — 2026-05-28
- [Claude Codeを3セッション並列で8時間動かしたら、2回コンテキストを上書きしあった](https://kenimoto.dev/ja/blog/three-claude-sessions-parallel-8h-context-overwrite/) — 2026-05-27
- [5つのAI検索に『kenimoto.devを引用して』と頼んだ。記事31本のうち出てきたのは3本だった](https://kenimoto.dev/ja/blog/five-ai-engines-cite-my-blog-three-of-thirty-one/) — 2026-05-26
- [JSON-LDを11スキーマ入れた。3ヶ月測ったら、AIが拾っていたのは3つだけだった](https://kenimoto.dev/ja/blog/json-ld-11-only-3-cited/) — 2026-05-25
- [コードレビューを6段階にしたら、AIと人間の分業が見えた](https://kenimoto.dev/ja/blog/code-review-6-stages-ai-human-boundary/) — 2026-05-24
- [robots.txtにAIクローラー13個の個別ルールを書いた。30日後、守ったのは3つだけだった](https://kenimoto.dev/ja/blog/robots-txt-ai-crawler-rules-30-days-only-3-followed/) — 2026-05-24
- [AI Overviewsに載るための4条件を30日試した。効いたのは1つだけだった](https://kenimoto.dev/ja/blog/ai-overviews-4-conditions-30-days-only-one-worked/) — 2026-05-23
- [他のエージェントを監査する4層目を足したら、Strategistが3週間サボっていたことが発覚した話](https://kenimoto.dev/ja/blog/evolver-4-layer-strategist-procrastination-audit/) — 2026-05-22
- [ブログを4言語化したら、ポルトガル語版だけ流入が約4倍になった - 22日分の生データ](https://kenimoto.dev/ja/blog/4-languages-30-days-portuguese-4x-traffic/) — 2026-05-21
- [Claude Code Skills を10個書いたら、4個に統合された — Reusable Pattern が回り始める瞬間と回らない瞬間](https://kenimoto.dev/ja/blog/claude-code-skills-10-to-4-integration-pattern/) — 2026-05-20
- [Claudeに先週47回『おっしゃる通りです』と言われた。そのうち11回は私が、36回はClaudeが間違っていた。](https://kenimoto.dev/ja/blog/claude-osshatoori-47kai-sycophancy-jissoku/) — 2026-05-19
- [同じサイトを7つのAI引用トラッカーに入れたら、7つとも違う数字を返してきた](https://kenimoto.dev/ja/blog/seven-ai-citation-trackers-different-numbers/) — 2026-05-18
- [30日間サーバーログを見続けて分かった、私のサイトを最も叩いた5つのAIクローラー - そこから読めるLLMOシグナル](https://kenimoto.dev/ja/blog/five-ai-crawlers-30days-server-log/) — 2026-05-17
- [ClaudeをカオスエンジニアリングのMCPサーバーに繋いだら、ステージングを4回殺した — 6ヶ月見逃していた本番バグを見つけた話](https://kenimoto.dev/ja/blog/claude-chaos-engineering-mcp-staging-4-kai-koroshita/) — 2026-05-16
- [Claudeが3回連続でバグを「隠す修正」を出してきた話 — デバッグ10の技法をプロンプトに翻訳する](https://kenimoto.dev/ja/blog/claude-bug-kakushi-debug-10-techniques-prompt/) — 2026-05-15
- [Strategist に WebSearch を持たせたら 5テーマ選びに 20分かかった - Observer / Strategist / Marketer 3役分離で 3分にした話](https://kenimoto.dev/ja/blog/observer-strategist-marketer-3-yaku-bunri/) — 2026-05-14
- [音声AIスタックを5つ実測した。300msの壁を越えられたのは2つだけだった](https://kenimoto.dev/ja/blog/voice-ai-5-stacks-only-two-under-300ms/) — 2026-05-13
- [3人のサブエージェントに同じPRを見せたら、4割の指摘で意見が割れた話](https://kenimoto.dev/ja/blog/three-sub-agents-pr-review-40-percent-disagreement/) — 2026-05-12
- [他社の llms.txt を30個監査したら、すでに5つのアンチパターンが形になっていた](https://kenimoto.dev/ja/blog/llms-txt-audit-30-files-5-anti-patterns/) — 2026-05-11
- [SEOが壊れる日 — 私のAIエージェントは、もうGoogleを見ていなかった](https://kenimoto.dev/ja/blog/llmo-three-paths-introduction/) — 2026-05-10
- [CLAUDE.md は結局 Context Engineering を1ファイルに凝縮したものだった](https://kenimoto.dev/ja/blog/claude-md-context-engineering-practice/) — 2026-05-09
- [同じ質問なのに、LLMから5つの違う回答が返ってきた — Context Engineering 入門](https://kenimoto.dev/ja/blog/context-engineering-introduction-five-strategies/) — 2026-05-08
- [構造化データ11スキーマをLLM向けに統合](https://kenimoto.dev/ja/blog/json-ld-11-schemas-llm-understanding/) — 2026-05-07
- [ChatGPTからのアクセスは、GA4にどう映るのか - LLMOを数値で把握する3つの方法](https://kenimoto.dev/ja/blog/llmo-measurement-3-methods/) — 2026-05-06
- [LLMO最小実装は2ファイル15分で終わる](https://kenimoto.dev/ja/blog/llmo-minimum-implementation-llms-txt-json-ld/) — 2026-05-05
- [新規ドメインの最初の1週間、GA4は嘘をつく: kaoriq.com立ち上げ4日間の生データ](https://kenimoto.dev/ja/blog/new-domain-first-week-ga4-is-a-lie/) — 2026-05-05
- [ChatGPTからの流入を90日で8,337%増やした実装: TRMが採用した4つのLLMO戦略柱](https://kenimoto.dev/ja/blog/llmo-case-studies-trm-8337-percent/) — 2026-05-04
- [AIエージェントは月いくらかかるのか -- API・サブスク・ローカルの損益分岐点](https://kenimoto.dev/ja/blog/ai-agent-cost-structure-breakeven/) — 2026-05-03
- [Claude Code Hooks v2 — 「お願い」を「プログラム」に変える25のイベント](https://kenimoto.dev/ja/blog/claude-code-hooks-v2-25-events/) — 2026-05-02
- [Claude CodeのSub-agent設計 — 1セッションで専門家チームを使い分ける](https://kenimoto.dev/ja/blog/claude-code-sub-agent-design/) — 2026-05-01
- [AIパイプラインが壊れる9つの理由 -- 全部AIの外側だった](https://kenimoto.dev/ja/blog/9-bugs-in-my-ai-pipeline/) — 2026-04-30
- [LLMに「引用したい」と思わせるコンテンツ設計 — Microsoftが明かした3原則](https://kenimoto.dev/ja/blog/llm-content-design-microsoft-3-principles/) — 2026-04-30
- [MCPサーバーに接続しただけでトークンが消える — 4サービスで実測した](https://kenimoto.dev/ja/blog/mcp-token-cost-measurement/) — 2026-04-30
- [Claude Codeで自律型コンテンツパイプラインを構築した話](https://kenimoto.dev/ja/blog/hello-world/) — 2026-04-29

### Portuguese (Português)

- [Context Engineering: 6 no livro, 2 na produção](https://kenimoto.dev/pt/blog/context-engineering-6-no-livro-2-na-producao/) — 2026-10-12
- [arXiv 2603.25723: 6 campos de Harness formalizados, mapeei meu harness-ops](https://kenimoto.dev/pt/blog/arxiv-2603-25723-nlah-6-campos-formais/) — 2026-10-11
- [Postmortem blameless é teatro: 3 sinais no squad](https://kenimoto.dev/pt/blog/postmortem-sem-culpado-teatro-3-sinais/) — 2026-10-10
- [Ignorância pluralística: 3 sinais no seu time](https://kenimoto.dev/pt/blog/ignorancia-pluralistica-3-sinais-no-seu-time/) — 2026-10-09
- [Context Engineering: RAG é 80%, Full CE só +12%](https://kenimoto.dev/pt/blog/context-engineering-rag-80-full-ce-apenas-12/) — 2026-10-08
- [Template de PR: 11 itens não marcam, 5-7 sim](https://kenimoto.dev/pt/blog/template-pr-11-itens-5-7/) — 2026-10-07
- [Memória de agente: 7 arquivos, não carregue tudo](https://kenimoto.dev/pt/blog/memoria-agente-7-arquivos-3-regras-corte/) — 2026-10-04
- [LLMO sem teste quebra: Playwright CI em 7 passos](https://kenimoto.dev/pt/blog/playwright-ci-llmo-7-passos/) — 2026-10-03
- [Conventional Comments: 9 rótulos, 1 trava merge](https://kenimoto.dev/pt/blog/conventional-comments-9-rotulos-1-trava-merge/) — 2026-10-02
- [Vibe Coding: você sente +20%, mede -19%](https://kenimoto.dev/pt/blog/vibe-coding-vies-metr-19-por-cento/) — 2026-10-01
- [Auditei 30 llms.txt em 2026: 5 anti-padrões](https://kenimoto.dev/pt/blog/auditei-30-llms-txt-ia-2026-5-anti-padroes/) — 2026-09-30
- [Dijkstra vs Claude Code: 3 citações que uso na revisão](https://kenimoto.dev/pt/blog/dijkstra-3-citacoes-claude-code-refatorar/) — 2026-09-29
- [Skill vs Hook vs MCP: 3 formas, 3 tokens/mês](https://kenimoto.dev/pt/blog/skill-hook-mcp-3-lugares-tokens-claude-code/) — 2026-09-27
- [AGENTS.md como loop de feedback: 3 meses depois](https://kenimoto.dev/pt/blog/agentsmd-como-loop-feedback-3-meses-depois/) — 2026-09-26
- [Anthropic soltou context reset no Opus 4.5](https://kenimoto.dev/pt/blog/anthropic-context-anxiety-opus-45-drop-reset/) — 2026-09-24
- [Não brigue com o modelo: Anthropic mede 150%](https://kenimoto.dev/pt/blog/nao-brigue-com-o-modelo-boris-cherny/) — 2026-09-23
- [Template de PR em 9 OSS: 3 renderizam zero](https://kenimoto.dev/pt/blog/template-de-pr-9-oss-3-renderizam-zero/) — 2026-09-22
- [Claude Code hooks + lifecycle: 33 eventos](https://kenimoto.dev/pt/blog/claude-code-hooks-lifecycle-feedback-loop-33-eventos/) — 2026-09-20
- [Dunning-Kruger com Claude Code: 4 mecanismos que aprofundam o vale (e não amenizam)](https://kenimoto.dev/pt/blog/dunning-kruger-claude-code-4-mecanismos-aprofundam-vale/) — 2026-09-20
- [AGENTS.md ou CLAUDE.md? 35 repos, 5 padrões](https://kenimoto.dev/pt/blog/agents-md-ou-claude-md-35-repos-5-padroes/) — 2026-09-18
- [Google mediu 21% mais pushback em code review por gênero. Sou solo — e reconheci o reflexo no meu sub-agente.](https://kenimoto.dev/pt/blog/pushback-google-21-code-review-sub-agente-solo/) — 2026-09-17
- [3 AI coding termos: Vibe, Spec, Agent Framework](https://kenimoto.dev/pt/blog/voce-chama-3-coisas-de-ai-coding-mapa-vibe-spec-agent/) — 2026-09-15
- [BATNA para engenheiros: 3 técnicas de negociação Harvard](https://kenimoto.dev/pt/blog/batna-engenheiros-3-tecnicas-negociacao/) — 2026-09-14
- [5 métricas de code review: 0 comentário = alerta](https://kenimoto.dev/pt/blog/5-metricas-code-review-0-comentario-alerta/) — 2026-09-13
- [Manter Knowledge Graph: 3 padrões de quebra](https://kenimoto.dev/pt/blog/knowledge-graph-schema-evolution-gargalo-silencioso/) — 2026-09-12
- [Claude Code 9h sem /clear: contexto apodreceu 3h](https://kenimoto.dev/pt/blog/claude-code-9h-sem-clear-contexto-apodreceu-3h/) — 2026-09-10
- [CLAUDE.md em equipe: 7 padrões e 3 armadilhas](https://kenimoto.dev/pt/blog/claude-md-time-7-padroes-3-quebras/) — 2026-09-09
- [Amplificação em 3 typos: S3, Kinesis, Cloudflare](https://kenimoto.dev/pt/blog/amplificacao-incidentes-s3-kinesis-cloudflare/) — 2026-09-07
- [3 armadilhas cognitivas em debugging: 2h → 15min](https://kenimoto.dev/pt/blog/3-armadilhas-cognitivas-debugging-2h-15min/) — 2026-09-06
- [5 vieses que decidem sua stack antes do teste](https://kenimoto.dev/pt/blog/5-vieses-que-decidem-sua-stack-antes-do-teste/) — 2026-09-05
- [Harness segundo Fowler: 4 camadas no backend](https://kenimoto.dev/pt/blog/fowler-4-harness-invisiveis-backend/) — 2026-09-04
- [Confiei no Claude: 3 vieses da era IA que peguei](https://kenimoto.dev/pt/blog/confiei-no-claude-3-vieses-nova-era-ia/) — 2026-09-02
- [Modelos maiores mentem melhor (3.5x specificity)](https://kenimoto.dev/pt/blog/modelos-maiores-mentem-melhor-3-5x-specificity/) — 2026-09-01
- [PLAID e o dia que Claude Code multiplicou os PRs por 4x — o que a equipe fez para não desabar](https://kenimoto.dev/pt/blog/plaid-claude-code-4x-prs/) — 2026-08-28
- [Self-Evolving Agent: meu harness reescreveu AGENTS.md 24 vezes em 6 meses](https://kenimoto.dev/pt/blog/self-evolving-agent-24-rewrites-agents-md-6-meses/) — 2026-08-25
- [Reciprocidade em code review: revisei 47 PRs open-source em 6 meses, meu refactor grande passou em 3 dias](https://kenimoto.dev/pt/blog/reciprocidade-code-review-cialdini-6-meses-oss/) — 2026-08-23
- [Reunião de decisão técnica que travava 90 min: 3 táticas que a cortaram para 35](https://kenimoto.dev/pt/blog/reuniao-decisao-tecnica-90-para-35-min-3-taticas/) — 2026-08-22
- [O ataque Shai-Hulud: por que apertar 'Sim' no Claude Code virou a nova falha crítica](https://kenimoto.dev/pt/blog/shai-hulud-claude-code-aprovar-sim-falha-critica/) — 2026-08-20
- [Agente autônomo por 30 dias: 5 incidentes novos custaram R$ 8.400](https://kenimoto.dev/pt/blog/agente-autonomo-30-dias-5-incidentes-brl/) — 2026-08-17
- [A 4ª falha de Spec-Driven Development com Claude Code (post-mortem de 6 semanas)](https://kenimoto.dev/pt/blog/4a-falha-spec-driven-development-claude-code-6-semanas/) — 2026-08-16
- [autoFixable: o padrão que corrige 40% dos comentários de PR antes do humano ver](https://kenimoto.dev/pt/blog/autofixable-padrao-40-porcento-comentarios-pr/) — 2026-08-15
- [Filtro duplo de IA no PR: CodeRabbit + Claude Code Action cortaram 47% dos comentários humanos (ch10)](https://kenimoto.dev/pt/blog/filtro-duplo-ia-pr-coderabbit-claude-code-action-cortou-47/) — 2026-08-14
- [Automatizei minha declaração de imposto com Claude Code: 14 horas viraram 47 minutos (e o que a Receita ainda não sabe)](https://kenimoto.dev/pt/blog/automatizei-declaracao-imposto-claude-code-14h-47min/) — 2026-08-13
- [CTR do #1 do Google caiu 34,5% com AI Overview — o checklist LLMO de 1 hora](https://kenimoto.dev/pt/blog/ai-overview-ctr-34-5-checklist-llmo-1-hora/) — 2026-08-12
- [A transcrição funcionou. O acompanhamento, não: tentei gerar playback de karaokê com OSS e travei três vezes](https://kenimoto.dev/pt/blog/karaoke-oss-transcricao-funcionou-falta-o-arranjo/) — 2026-08-12
- [AGENTS.md como política única de code review — 47 PRs depois, sistema virou juiz único](https://kenimoto.dev/pt/blog/agents-md-como-politica-unica-code-review-47-prs/) — 2026-08-10
- [Agentic RAG em produção: o agente decide 4 vezes na mesma pergunta - alucinação cai 3x, custo dobra](https://kenimoto.dev/pt/blog/agentic-rag-4-iteracoes-3x-menos-alucinacao-2x-custo/) — 2026-08-09
- [CLAUDE.md pede, hook impõe: os 3 níveis de contrato que separam 90% de 100%](https://kenimoto.dev/pt/blog/claude-md-pede-hook-impoe-90-vs-100/) — 2026-08-07
- [Troquei Sonnet por Qwen 7B local em 3 tarefas: bill de US$ 47 virou R$ 0 (e 2 tarefas ficaram melhores)](https://kenimoto.dev/pt/blog/troquei-sonnet-por-qwen-7b-local-47-para-0/) — 2026-08-06
- [Few-shot: testei 2/4/8/16/32 exemplos em 3 tarefas e a qualidade caiu depois de N](https://kenimoto.dev/pt/blog/few-shot-2-4-8-16-32-ponto-otimo-3-tarefas/) — 2026-08-03
- [Padrão autoFixable: como separei 41% dos comentários de review antes da IA olhar](https://kenimoto.dev/pt/blog/autofixable-antes-ia-41-porcento-comentarios-review-cortados/) — 2026-08-02
- [Ataque Shai-Hulud no Claude Code: 3 defesas contra o autopiloto 'Yes'](https://kenimoto.dev/pt/blog/ataque-shai-hulud-claude-code-3-defesas-autopiloto-yes/) — 2026-08-01
- [Haiku 3 + RAG (11,8) venceu Sonnet 4 puro (5,3): o experimento que provou que contexto vale mais que modelo](https://kenimoto.dev/pt/blog/haiku-3-rag-11-8-vs-sonnet-4-puro-5-3-contexto-vence-modelo/) — 2026-07-31
- [52,8% para 66,5%: como Generator-Evaluator inspirado em GAN eleva accuracy do RAG](https://kenimoto.dev/pt/blog/52-8-para-66-5-generator-evaluator-rag-gan/) — 2026-07-30
- [Camada 1 → 2 → 3: o funil de code review com IA que corta 70% do trabalho humano](https://kenimoto.dev/pt/blog/camada-1-2-3-funil-code-review-ia-70pct/) — 2026-07-28
- [O resumo de IA diz que \"não é golpe\". O trampolim é a caixa de busca do seu site](https://kenimoto.dev/pt/blog/resumo-de-ia-nao-e-golpe-caixa-de-busca/) — 2026-07-28
- [Haiku e Qwen 4B ganharam do Opus em 30% dos meus 47 casos reais](https://kenimoto.dev/pt/blog/haiku-qwen-4b-vencem-opus-14-47/) — 2026-07-27
- [Dark Patterns em Entrevista Tech: 7 Armadilhas que Identifiquei em 12 Processos](https://kenimoto.dev/pt/blog/sete-dark-patterns-entrevista-tech-12-processos/) — 2026-07-26
- [Agente 24h e o 2º incidente: 3 falhas que o postmortem inicial escondeu](https://kenimoto.dev/pt/blog/agente-24h-segundo-incidente-3-falhas-postmortem/) — 2026-07-25
- [3 vieses cognitivos que estragam sua estimativa de sprint — e como o Claude Code corta 2 deles](https://kenimoto.dev/pt/blog/3-vieses-cognitivos-estimativa-sprint-claude-code-corta-2/) — 2026-07-24
- [TDD com Playwright MCP e CI: pipeline Claude Code em 47 min](https://kenimoto.dev/pt/blog/tdd-playwright-mcp-ci-claude-code-47-min/) — 2026-07-23
- [Escrevi 'testes antes do PR' no AGENTS.md por 3 Meses. Só 12% Seguiram.](https://kenimoto.dev/pt/blog/agents-md-testes-pr-12-por-cento-seguiram/) — 2026-07-22
- [GraphRAG para microsserviço: 7 passos para saber quais ADRs quebram seu service](https://kenimoto.dev/pt/blog/graphrag-microsservico-adr-7-passos/) — 2026-07-21
- [40% dos projetos de agentes de IA falharam em 2026: peguei 12 post-mortems e classifiquei em prompt, context ou harness](https://kenimoto.dev/pt/blog/agentes-ia-40-falharam-2026-12-postmortems-prompt-context-harness/) — 2026-07-20
- [Trocamos o Harness (Não o Modelo) e o Benchmark Subiu +13,7 Pontos](https://kenimoto.dev/pt/blog/trocamos-harness-nao-modelo-benchmark-137-pontos/) — 2026-07-19
- [Harness Engineering: os 6 componentes que auditam seu CLAUDE.md em 10 min](https://kenimoto.dev/pt/blog/harness-engineering-6-componentes-auditoria-claude-md/) — 2026-07-18
- [IA escreve monótono em 70 de 70 células, em 3 línguas. E o corpus humano em português veio do TabNews](https://kenimoto.dev/pt/blog/sotaque-de-maquina-70-70-celulas-3-linguas/) — 2026-07-18
- [5 mentiras que todo LLM te conta — e como o Context Engineering corrige cada uma](https://kenimoto.dev/pt/blog/5-mentiras-llm-context-engineering-pt/) — 2026-07-17
- [Prompt morreu, Context está morrendo: Harness Engineering é a próxima onda](https://kenimoto.dev/pt/blog/prompt-morreu-context-morrendo-harness-engineering-proxima-onda/) — 2026-07-16
- [Harness Engineering: 5 empresas, 5 definições — o mapa dos termos em conflito (2026)](https://kenimoto.dev/pt/blog/harness-engineering-5-empresas-mapa-termos-conflito/) — 2026-07-15
- [Trabalho híbrido sabota promoção: viés de proximidade (2026)](https://kenimoto.dev/pt/blog/trabalho-hibrido-vies-proximidade-promocao-2026/) — 2026-07-15
- [Adicionando um servidor MCP a um OSS: do YAML à imagem PNG](https://kenimoto.dev/pt/blog/historymap-mcp-server-yaml-to-png/) — 2026-07-14
- [Knowledge Graph pessoal: o segundo cérebro que o Notion nunca vai conseguir ser](https://kenimoto.dev/pt/blog/knowledge-graph-pessoal-segundo-cerebro-notion-nunca-vai-conseguir-ser/) — 2026-07-14
- [Meu agente Claude não percebeu que o cliente estava furioso — a camada de senso comum emocional que Meta descobriu antes de mim](https://kenimoto.dev/pt/blog/agente-claude-cliente-furioso-commonsense-emocional/) — 2026-07-13
- [Camada 1 do harness de code review: por que 3 hooks pre-commit cortam 40% do trabalho da IA](https://kenimoto.dev/pt/blog/harness-code-review-camada-1-hooks-pre-commit-40-percent/) — 2026-07-12
- [historymap: um arquivo YAML vira uma linha do tempo estilo site corporativo](https://kenimoto.dev/pt/blog/historymap-yaml-linha-do-tempo-oss/) — 2026-07-11
- [Padrão autoFixable: 0 revisões humanas de lint em 3 meses](https://kenimoto.dev/pt/blog/padrao-autofixable-zero-revisoes-humanas-lint-3-meses/) — 2026-07-10
- [README de perfil no GitHub: 7 estilos em 10](https://kenimoto.dev/pt/blog/dissequei-10-perfis-github-7-estilos/) — 2026-07-09
- [7 passos para construir um Knowledge Graph do seu código e cortar 90% das sugestões fantasma do Cursor](https://kenimoto.dev/pt/blog/7-passos-knowledge-graph-codigo-cursor-alucinacao/) — 2026-07-08
- [Eu estava lançando meu plugin às 3 da manhã quando ele me avisou: 67% de contexto. Vinte minutos depois, ele mesmo quebrou](https://kenimoto.dev/pt/blog/compact-ops-plugin-me-avisou-no-meio-do-lancamento/) — 2026-07-07
- [GraphRAG vs RAG clássico: quando o grafo vale 7x o custo](https://kenimoto.dev/pt/blog/graphrag-vs-rag-classico-4-projetos-quando-vale-7x-custo/) — 2026-07-07
- [Confiei no Claude Code no piloto automático e o Shai-Hulud quase entrou: 5 permissões que reduzi hoje](https://kenimoto.dev/pt/blog/piloto-automatico-5-permissoes/) — 2026-07-06
- [autoFixable: transformei 30 min de bate-e-volta em revisão de código em 47 segundos](https://kenimoto.dev/pt/blog/autofixable-30min-para-47-segundos-revisao-ia/) — 2026-07-05
- [RAG vs GraphRAG: o teste de 7 passos](https://kenimoto.dev/pt/blog/graphrag-vs-rag-teste-7-passos-quando-trocar/) — 2026-07-04
- [Agent = Model + Harness: autópsia de US$ 47.000](https://kenimoto.dev/pt/blog/47k-fim-de-semana-argumento-definitivo-agent-model-harness/) — 2026-07-03
- [O criador do CLAUDE.md escreve 2 linhas. A comunidade escreve 100. Quem está errado?](https://kenimoto.dev/pt/blog/claude-md-2-vs-100-linhas-quem-erra/) — 2026-06-30
- [Pare de pagar pelo modelo grande: Haiku + RAG fez 11,8 contra 5,3 do Sonnet sozinho](https://kenimoto.dev/pt/blog/haiku-rag-vence-sonnet-sozinho-contexto/) — 2026-06-29
- [Codex rodou em 1 milhão de linhas de código: e provou que trocar de modelo nunca conserta seu agente](https://kenimoto.dev/pt/blog/codex-1-milhao-linhas-trocar-modelo-nao-conserta-agente/) — 2026-06-28
- [Os 6 componentes de um harness de IA: o checklist de auditoria que cabe em uma página](https://kenimoto.dev/pt/blog/6-componentes-harness-checklist-auditoria-uma-pagina/) — 2026-06-27
- [US$ 47.000 num fim de semana: a tempestade de 2,3 milhões de retries que o harness teria parado em 4 linhas](https://kenimoto.dev/pt/blog/47000-fim-de-semana-tempestade-retries-harness-4-linhas/) — 2026-06-26
- [Agent = Model + Harness: Por Que Trocar GPT por Claude Nunca Resolve o Bug do Seu Agente](https://kenimoto.dev/pt/blog/agent-equal-model-mais-harness-trocar-modelo-nao-resolve/) — 2026-06-25
- [MCP me custou 27 mil tokens só pra dizer oi. Knowledge Graph cortou 150 mil pra 18 mil no code review (8,3x). Qual ferramenta ganha qual batalha](https://kenimoto.dev/pt/blog/mcp-27k-vs-kg-8x-qual-ganha/) — 2026-06-24
- [Construir ficou trivial com Claude Code — o gargalo agora é vender, e quase nenhum engenheiro percebeu](https://kenimoto.dev/pt/blog/construir-ficou-trivial-gargalo-vender/) — 2026-06-23
- [Pare de mandar o Claude Code codar direto: desenhei o dia no Plan Mode e o retrabalho despencou](https://kenimoto.dev/pt/blog/plan-mode-projetar-dia-retrabalho/) — 2026-06-23
- [40% dos projetos de agente de IA vão ser cancelados até 2027. A culpa quase nunca é do modelo.](https://kenimoto.dev/pt/blog/agentes-ia-40-falham-harness/) — 2026-06-20
- [95% do conhecimento da sua empresa mora na cabeça de um veterano: como a Meta transformou o tácito em grafo](https://kenimoto.dev/pt/blog/conhecimento-tacito-grafo-meta/) — 2026-06-19
- [Anotar conhecimento à mão não escala: montei um pipeline que registrou 300 fontes sozinho em 3 meses](https://kenimoto.dev/pt/blog/banco-conhecimento-300-3-meses/) — 2026-06-18
- [Seu cérebro reconhece o bug antes de você ler o log (e erra mais do que você imagina)](https://kenimoto.dev/pt/blog/cerebro-reconhece-bug-antes-do-log/) — 2026-06-17
- [Seu site em React é invisível pra IA: os crawlers não rodam o seu JavaScript](https://kenimoto.dev/pt/blog/site-react-invisivel-ia-crawlers-nao-rodam-javascript/) — 2026-06-16
- [Deixei a IA escrever 100 testes verdes. Mutation testing diz que pegaram só 58% dos bugs.](https://kenimoto.dev/pt/blog/ia-100-testes-verdes-mutation/) — 2026-06-15
- [O System Prompt que todo mundo escreve é o que faz a IA mentir](https://kenimoto.dev/pt/blog/responda-em-detalhes-faz-ia-mentir/) — 2026-06-14
- [796 pacotes npm comprometidos por gente que apertou 'Sim' no automático — a anatomia do Shai-Hulud](https://kenimoto.dev/pt/blog/shai-hulud-796-pacotes-sim-automatico/) — 2026-06-13
- [Disseram que 'inglês é a nova linguagem de programação'. Passei 3 meses programando assim, e ainda preciso saber programar.](https://kenimoto.dev/pt/blog/ingles-nao-e-a-nova-linguagem-de-programacao/) — 2026-06-12
- [O cargo de 'engenheiro' não vai sumir — vira 80% decisão, 20% código](https://kenimoto.dev/pt/blog/engenheiro-nao-vai-sumir/) — 2026-06-11
- [Eu rodava geração de conteúdo e checagem de pedidos no mesmo cron. Uma API lenta travou tudo por 5 horas](https://kenimoto.dev/pt/blog/cron-unico-time-vs-event-5h/) — 2026-06-10
- [RAG só busca. GraphRAG raciocina. Como o LinkedIn cortou 28,6% do tempo de suporte](https://kenimoto.dev/pt/blog/graphrag-raciocina-linkedin-suporte/) — 2026-06-09
- [Estimei 47 tarefas em 6 meses: a regra do x2 que eu parei de negar](https://kenimoto.dev/pt/blog/estimativa-x2-47-tarefas/) — 2026-06-08
- [Uma escritora publicou um app inteiro escrevendo quase zero linha de código: 9.000 linhas, 20 horas](https://kenimoto.dev/pt/blog/escritora-publicou-app-quase-zero-codigo/) — 2026-06-07
- [Construí tudo no Claude Code. Aí o preço dobrou da noite pro dia](https://kenimoto.dev/pt/blog/dependencia-ferramenta-ia-plano-b/) — 2026-06-06
- [O passo mais valioso da revisão de código por IA não usa IA nenhuma](https://kenimoto.dev/pt/blog/revisao-codigo-ia-passo-sem-ia-tree-sitter/) — 2026-06-05
- [Não automatize tudo: os 4 níveis de delegação que decidem se o Claude Code te ajuda ou atrapalha](https://kenimoto.dev/pt/blog/4-niveis-delegacao-claude-code/) — 2026-06-04
- [Construí o RAG que mandaram fazer (Voyage + vector DB + busca semântica). Depois deletei tudo: o grep venceu.](https://kenimoto.dev/pt/blog/construi-rag-deletei-grep-venceu/) — 2026-06-03
- [O modelo maior mente melhor: por que o Sonnet inventa com mais confiança que o Haiku](https://kenimoto.dev/pt/blog/modelo-maior-mente-melhor-sonnet-vs-haiku/) — 2026-06-02
- [Confiei que a IA me deixava mais rápido. Os dados dizem que fiquei mais lento](https://kenimoto.dev/pt/blog/confiei-ia-mais-rapido-fiquei-mais-lento/) — 2026-06-01
- [O modelo barato + RAG bateu o caro em 223% nos meus testes. Refiz a conta com os preços de 2026.](https://kenimoto.dev/pt/blog/haiku-rag-223-modelo-barato-precos-2026/) — 2026-05-31
- [Skills do Claude Code consomem tokens mesmo sem disparar. Medi 5 Skills em 7 horas — os 3 que nunca rodaram comeram 11% da conta.](https://kenimoto.dev/pt/blog/skills-3-dormentes-18-tokens/) — 2026-05-30
- [Coloquei 7 agentes de IA no cron diário. 2 ficaram em silêncio por 18 dias. Tracing não pegou. Um contrato de exit code pegou.](https://kenimoto.dev/pt/blog/cron-7-agentes-silent-18-dias/) — 2026-05-28
- [Rodei 3 sessões do Claude Code em paralelo por 8 horas. Elas se sobrescreveram 2 vezes. R$ 280 em retrabalho.](https://kenimoto.dev/pt/blog/tres-sessoes-claude-code-paralelo-8h-2-colisoes/) — 2026-05-27
- [Pedi a 5 IAs que citassem meu próprio blog. 31 artigos publicados, só 3 apareceram.](https://kenimoto.dev/pt/blog/cinco-ias-citaram-meu-blog-tres-de-trinta-e-um/) — 2026-05-26
- [Coloquei 11 schemas JSON-LD. Em 3 meses, só 3 foram citados por IA](https://kenimoto.dev/pt/blog/11-json-ld-apenas-3-citados/) — 2026-05-25
- [Por que toda code review vira briga: 5 vieses psicológicos que sabotam seu PR](https://kenimoto.dev/pt/blog/code-review-vira-briga-5-vieses-psicologicos/) — 2026-05-23
- [Adicionei um 4º agente que audita meus outros agentes. Ele pegou meu Strategist enrolando há 3 semanas.](https://kenimoto.dev/pt/blog/quarto-agente-evolver-pegou-strategist-enrolando/) — 2026-05-22
- [Quanto um time gasta apontando indentação no PR: a conta dá R$ 3.600 por mês](https://kenimoto.dev/pt/blog/medi-revisao-codigo-3600-reais-mes/) — 2026-05-21
- [Traduzi o blog pra 4 idiomas. O português pegou quase 4× o tráfego do inglês.](https://kenimoto.dev/pt/blog/traduzi-blog-4-idiomas-portugues-4x-trafego/) — 2026-05-21
- [Conectei o Claude Code a 4 servidores MCP. Só o handshake queimou 27.000 tokens — e nem foi o pior.](https://kenimoto.dev/pt/blog/conectei-claude-4-servidores-mcp-27k-tokens-handshake/) — 2026-05-20
- [Contei quantas vezes o Claude me disse 'Você está absolutamente certo!' na semana passada. 47 vezes. Em 11 delas, eu não estava. Nas outras 36, o Claude também não.](https://kenimoto.dev/pt/blog/claude-voce-esta-certo-47-vezes-sycophancy-medi/) — 2026-05-19
- [Plugei o mesmo site em 7 rastreadores de citação por IA. Nenhum bateu com o outro.](https://kenimoto.dev/pt/blog/7-rastreadores-citacoes-ia-numeros-diferentes/) — 2026-05-18
- [Os 5 crawlers de IA que mais bateram nos meus sites em 30 dias - o que os logs revelaram sobre LLMO](https://kenimoto.dev/pt/blog/cinco-crawlers-ia-bateram-meu-site-30-dias/) — 2026-05-17
- [Pluguei o Claude num MCP de Chaos. Matou staging 4 vezes pra achar um bug que ignorávamos há 6 meses.](https://kenimoto.dev/pt/blog/claude-chaos-engineering-mcp-matou-staging-4-vezes/) — 2026-05-16
- [Peguei o Claude escondendo meu bug 3 vezes seguidas. Aí virei 10 hábitos de debug em prompts.](https://kenimoto.dev/pt/blog/claude-escondeu-meu-bug-3-vezes-10-habitos-debug/) — 2026-05-15
- [Dei WebSearch ao meu Strategist. 5 temas levaram 20 minutos. Separar em 3 agentes baixou pra 3.](https://kenimoto.dev/pt/blog/tres-papeis-observer-strategist-marketer-separacao/) — 2026-05-14
- [Testei 5 stacks de Voice AI. Apenas 2 ficaram abaixo de 300ms.](https://kenimoto.dev/pt/blog/cinco-stacks-voice-ai-apenas-dois-abaixo-300ms/) — 2026-05-13
- [Coloquei 3 sub-agentes do Claude Code para revisar o mesmo PR. Discordaram em 41% dos comentários.](https://kenimoto.dev/pt/blog/tres-sub-agentes-revisaram-mesmo-pr-40-discordancia/) — 2026-05-12
- [Auditei 30 arquivos llms.txt em produção. 5 anti-padrões já estão se formando.](https://kenimoto.dev/pt/blog/auditei-30-arquivos-llms-txt-5-anti-padroes/) — 2026-05-11
- [Eu troquei Claude Code por OpenClaw por 24 horas. Veja o que quebrou — e o que melhorou](https://kenimoto.dev/pt/blog/openclaw-claude-code-24h-troca/) — 2026-05-10
- [Eu me recusei a escrever spec até o Claude gerar o código errado 3 vezes](https://kenimoto.dev/pt/blog/spec-driven-development-claude-code-3-falhas/) — 2026-05-09
- [Deixei meu agente Claude Code rodar 24 horas sozinho. A conta de R$ 2.000 nem foi o pior.](https://kenimoto.dev/pt/blog/agente-ia-24-horas-incidentes-seguranca/) — 2026-05-08
- [Por que o ChatGPT ignora o seu site (mesmo se você for #1 no Google)](https://kenimoto.dev/pt/blog/chatgpt-ignora-seu-site-llmo/) — 2026-05-07
- [Por que troquei prompt engineering por context engineering](https://kenimoto.dev/pt/blog/contexto-vs-prompt-engineering/) — 2026-05-07
- [Cursor vs Claude Code: 6 meses e a volta](https://kenimoto.dev/pt/blog/parei-cursor-voltei-terminal/) — 2026-05-07

### Spanish (Español)

- [6 CLAUDE.md en Claude Code: ganó el de 65 líneas](https://kenimoto.dev/es/blog/6-claude-md-65-vs-559-lineas/) — 2026-10-12
- [Vibe Coding en 2026: sensación +20%, METR -19%](https://kenimoto.dev/es/blog/vibe-coding-metr-sensacion-mas-20-realidad-menos-19/) — 2026-10-11
- [Claude Code Skills: borré los 3 dormidos](https://kenimoto.dev/es/blog/claude-code-skills-borre-3-dormidos/) — 2026-10-10
- [Claude Code en TypeScript: 3 smells frecuentes](https://kenimoto.dev/es/blog/claude-code-typescript-3-smells-frecuentes/) — 2026-10-09
- [Voice AI: 7 trucos para sentir 350ms con 750ms](https://kenimoto.dev/es/blog/voice-ai-7-trucos-perceptuales-350ms/) — 2026-10-07
- [Mutation testing: 100 tests IA pasaron, caza 58%](https://kenimoto.dev/es/blog/100-tests-ia-mutation-testing-58-caza-bugs/) — 2026-10-04
- [Harness autoevolutivo: 1 regla, 3 que borré](https://kenimoto.dev/es/blog/harness-autoevolutivo-1-regla-3-que-borre/) — 2026-10-02
- [Context engineering: benchmark 2,2x en 5 niveles](https://kenimoto.dev/es/blog/benchmark-5-niveles-context-engineering-tu-stack/) — 2026-10-01
- [3 hooks pre-commit antes del review con IA](https://kenimoto.dev/es/blog/3-hooks-pre-commit-antes-review-ia/) — 2026-09-27
- [Grafo estático pierde 61% de llamadas dinámicas](https://kenimoto.dev/es/blog/grafo-estatico-pierde-61-porciento-llamadas-dinamicas/) — 2026-09-26
- [Tu codebase ya era un harness. 5 señales.](https://kenimoto.dev/es/blog/codebase-ya-era-harness-5-senales/) — 2026-09-24
- [Barge-in de voz: 22% menos precisión, 4 umbrales](https://kenimoto.dev/es/blog/barge-in-voz-22-precision-vad-4-umbrales/) — 2026-09-23
- [Dunning-Kruger: 3 estadios y el mapa de skills](https://kenimoto.dev/es/blog/dunning-kruger-3-estadios-mapa-de-skills/) — 2026-09-22
- [Krisp 6M vs VAD: turn-taking en voice AI](https://kenimoto.dev/es/blog/krisp-6m-vad-turn-taking-voice-ai/) — 2026-09-21
- [Evaluar RAG en producción: 3 ejes y sus pesos](https://kenimoto.dev/es/blog/evaluar-rag-produccion-3-ejes-fidelidad-relevancia-exactitud/) — 2026-09-20
- [Primera semana de GA4 en dominio nuevo: kaoriq](https://kenimoto.dev/es/blog/ga4-primera-semana-dominio-nuevo-kaoriq/) — 2026-09-20
- [Claude Code hooks: 3 decisiones invisibles](https://kenimoto.dev/es/blog/claude-code-hooks-3-decisiones-invisibles/) — 2026-09-18
- [Servidores MCP sin autenticación: 38% y 30 CVEs](https://kenimoto.dev/es/blog/38-porciento-servidores-mcp-sin-autenticacion-30-cves/) — 2026-09-17
- [1 a 1 con junior silencioso: 4 técnicas](https://kenimoto.dev/es/blog/1-a-1-junior-silencioso-4-tecnicas/) — 2026-09-16
- [Agentic RAG: el LLM decide buscar 1-7 veces](https://kenimoto.dev/es/blog/agentic-rag-explicado-cuando-llm-decide-buscar-n-veces/) — 2026-09-15
- [Demanda latente: cómo nació Plan Mode en 30 min](https://kenimoto.dev/es/blog/demanda-latente-plan-mode-30-min/) — 2026-09-13
- [Vectara 2026: Haiku 4.5 alucinó menos que Sonnet](https://kenimoto.dev/es/blog/haiku-4-5-alucino-menos-que-sonnet-4-5-vectara-aa/) — 2026-09-10
- [Imágenes de IA convergen al mismo azul: 3 trucos](https://kenimoto.dev/es/blog/imagenes-ia-convergen-mismo-azul-3-trucos/) — 2026-09-09
- [Gemini CLI vs Claude Code: 3 huecos para LatAm](https://kenimoto.dev/es/blog/gemini-cli-vs-claude-code-latam-5-tareas-costo-real/) — 2026-09-08
- [InfraNodus + Obsidian: 800 notas, 5 huecos](https://kenimoto.dev/es/blog/infranodus-obsidian-centralidad-intermediacion-huecos/) — 2026-09-07
- [Tree-sitter code review: 3 pasos sin IA](https://kenimoto.dev/es/blog/tree-sitter-code-review-3-pasos-sin-ia/) — 2026-09-06
- [Claude Code: 3 patrones de memoria post-compact](https://kenimoto.dev/es/blog/claude-code-3-patrones-memoria-post-compact/) — 2026-09-05
- [Servidor MCP silencioso: 3 patrones de debug](https://kenimoto.dev/es/blog/mcp-servidor-silencioso-3-patrones-debug/) — 2026-09-04
- [Code review IA en 3 capas: hooks + IA + humano](https://kenimoto.dev/es/blog/code-review-ia-3-capas-hooks-ia-humano/) — 2026-09-02
- [OpenAI Realtime API: bajar tu voz IA a 300ms](https://kenimoto.dev/es/blog/openai-realtime-api-bajar-voz-ia-300ms/) — 2026-09-01
- [nanochat CORE: 22 Tests en 1 Número (0.256525)](https://kenimoto.dev/es/blog/nanochat-core-score-22-tests-un-numero/) — 2026-08-30
- [GraphRAG explicado: 7 pasos, vector vs multi-hop](https://kenimoto.dev/es/blog/graphrag-7-pasos-multi-hop/) — 2026-08-28
- [AGENTS.md vs CLAUDE.md en 2026: árbol de decisión de 5 preguntas](https://kenimoto.dev/es/blog/agents-md-vs-claude-md-2026-arbol-decision-5-preguntas/) — 2026-08-25
- [MCP en 5 minutos para devs LatAm: 1 protocolo, 12 integraciones](https://kenimoto.dev/es/blog/mcp-en-5-minutos-devs-latam-1-protocolo-12-integraciones/) — 2026-08-25
- [Claude Code Hooks v2 en LatAm: 5 usos y 30+ eventos](https://kenimoto.dev/es/blog/claude-code-hooks-v2-5-usos-produccion-latam/) — 2026-08-23
- [3 sesgos que rompen tus estimaciones cuando la IA dice \"lo hago en 10 minutos\"](https://kenimoto.dev/es/blog/3-sesgos-estimacion-ia-dice-10-minutos/) — 2026-08-22
- [Code review con Knowledge Graph + MCP: pasé de 150k a 18k tokens (8x menos que RAG del repo)](https://kenimoto.dev/es/blog/code-review-graph-mcp-8x-menos-tokens/) — 2026-08-22
- [El umbral de 300ms en IA de voz: los 3 límites de Nielsen aplicados al audio](https://kenimoto.dev/es/blog/300ms-ia-voz-3-limites-nielsen-audio/) — 2026-08-20
- [Contexto pequeño vence a modelo gigante: 3 experimentos en 2026](https://kenimoto.dev/es/blog/contexto-pequeno-vence-modelo-gigante-3-experimentos/) — 2026-08-17
- [Cron para agentes de IA: 4 patrones para detectar fallos silenciosos](https://kenimoto.dev/es/blog/cron-agentes-ia-4-patrones-fallos-silenciosos/) — 2026-08-15
- [Codex CLI vs Claude Code: probé 7 tareas reales en el mismo repo (guía LatAm 2026-08)](https://kenimoto.dev/es/blog/codex-cli-vs-claude-code-7-tareas-reales-latam-2026/) — 2026-08-14
- [Detectar 'AI Slop' en tu UI: 7 frases que delatan al LLM en tu copy](https://kenimoto.dev/es/blog/detectar-ai-slop-ui-7-frases-llm-copy/) — 2026-08-13
- [Bias de automatización — por qué confías en Claude más que en tu colega senior](https://kenimoto.dev/es/blog/bias-automatizacion-confias-claude-mas-colega-senior/) — 2026-08-12
- [Claude Code Skills vs MCP servers: cuándo usar cada uno (guía práctica LatAm 2026)](https://kenimoto.dev/es/blog/claude-code-skills-vs-mcp-servers-guia-latam-2026/) — 2026-08-09
- [Cuando tu agente de IA se atasca en un bucle: 4 síntomas y el fix para cada uno](https://kenimoto.dev/es/blog/agente-ia-bucle-4-sintomas-fix/) — 2026-08-06
- [Few-shot en LLMs: cuántos ejemplos son el punto óptimo antes de que la calidad caiga (por tarea)](https://kenimoto.dev/es/blog/few-shot-punto-optimo-cuantos-ejemplos-por-tarea/) — 2026-08-03
- [GraphRAG vs RAG clásico: 3 métricas que cambian cuando el grafo entra en juego](https://kenimoto.dev/es/blog/graphrag-vs-rag-clasico-3-metricas-grafo-entra/) — 2026-08-02
- [Medí las 346 réplicas de un M7.1 en Japón. Después lo comparé con el 19S y con Chile.](https://kenimoto.dev/es/blog/346-replicas-japon-19s-chile-misma-ley/) — 2026-08-01
- [GraphRAG vs RAG tradicional: cuándo el costo extra vale (y cuándo no)](https://kenimoto.dev/es/blog/graphrag-vs-rag-tradicional-cuando-el-costo-extra-vale/) — 2026-08-01
- [Entrenar un GPT-2 costaba $43,000 en 2019: cuánto cuesta hoy con nanochat de Karpathy (precios GPU 2026)](https://kenimoto.dev/es/blog/nanochat-entrenar-gpt-2-costaba-43000-2019-cuanto-cuesta-hoy/) — 2026-07-31
- [Auditar 30 blogs de devs con llmoframework: 4 de 5 mejores fallan las mismas 3 comprobaciones](https://kenimoto.dev/es/blog/auditar-30-blogs-devs-llmoframework-4-de-5-fallan-mismos-3/) — 2026-07-29
- [Prompt, Contexto, Harness: las 3 etapas de evolución que casi nadie termina](https://kenimoto.dev/es/blog/prompt-contexto-harness-3-etapas-evolucion/) — 2026-07-28
- [Qwen 35B en una RTX 4070: 2 Flags para Pasar de 12 a 34 tok/s (2.8x)](https://kenimoto.dev/es/blog/qwen-35b-rtx-4070-dos-flags-2-8x/) — 2026-07-26
- [GraphRAG vs RAG clásico: guía práctica para elegir en 5 preguntas](https://kenimoto.dev/es/blog/graphrag-vs-rag-5-preguntas-decidir/) — 2026-07-25
- [Sponsorware desde LatAm: 90 días hasta $1000/mes en GitHub Sponsors partiendo de cero](https://kenimoto.dev/es/blog/sponsorware-latam-90-dias-1000-mensuales-github-sponsors/) — 2026-07-24
- [Cuantización KV cache: 8x más contexto en RTX 4070 con dos flags](https://kenimoto.dev/es/blog/cuantizacion-kv-cache-8x-contexto-rtx-4070/) — 2026-07-23
- [System 1 vs System 2 al Depurar: Bugs de Madrugada Cuestan 3× Más](https://kenimoto.dev/es/blog/system-1-vs-system-2-debugging-madrugada/) — 2026-07-22
- [Nanochat de Karpathy: cuánto cuesta hoy entrenar un modelo tipo GPT-2 (era $43,000 en 2019)](https://kenimoto.dev/es/blog/nanochat-karpathy-costo-entrenar-gpt2-2019-2026/) — 2026-07-21
- [GraphRAG vs RAG: 7 pasos para tu primer knowledge graph de código en 30 min](https://kenimoto.dev/es/blog/graphrag-vs-rag-7-pasos-knowledge-graph-codigo/) — 2026-07-18
- [Prompt Caching de Anthropic: el 90% de descuento y 3 casos donde NO conviene](https://kenimoto.dev/es/blog/prompt-caching-anthropic-es/) — 2026-07-17
- [MCP + USB Serial: Claude Code controla hardware (2026)](https://kenimoto.dev/es/blog/mcp-usb-serial-claude-code-microcontrolador/) — 2026-07-15
- [Harness Engineering: la tercera era después de Prompt y Context, en 3 señales](https://kenimoto.dev/es/blog/harness-engineering-tercera-era-prompt-context-3-senales/) — 2026-07-14
- [Grafo de conocimiento personal en Neo4j: 300 notas sin mensualidad](https://kenimoto.dev/es/blog/grafo-conocimiento-personal-neo4j-notion/) — 2026-07-13
- [Los 7 archivos que Claude Code lee antes de escribir código — checklist de context engineering en 5 minutos](https://kenimoto.dev/es/blog/7-archivos-claude-code-context-engineering-checklist-5-minutos/) — 2026-07-12
- [Cuando el agente borró producción y cómo el harness lo detuvo en 3 capas](https://kenimoto.dev/es/blog/agente-borro-produccion-harness-3-capas/) — 2026-07-11
- [historymap: un archivo YAML se convierte en una línea de tiempo estilo sitio corporativo](https://kenimoto.dev/es/blog/historymap-yaml-linea-de-tiempo-oss/) — 2026-07-11
- [Dejé un agente de IA corriendo 24 horas: 3 incidentes de seguridad y los 7 permisos que hoy bloqueo antes de arrancar](https://kenimoto.dev/es/blog/agente-ia-24h-3-incidentes-7-permisos-bloqueo/) — 2026-07-04
- [Fatiga de decisiones con 3 IDEs de IA abiertos: el experimento de Vohs traducido a nuestro trabajo](https://kenimoto.dev/es/blog/fatiga-decisiones-3-ides-ia-vohs-traducido-a-dev/) — 2026-07-03
- [Agente = Modelo + Harness: la matriz de 6 decisiones para elegir donde invertir tu tiempo](https://kenimoto.dev/es/blog/agente-modelo-harness-matriz-6-decisiones/) — 2026-07-02
- [Construí mi base de conocimiento personal: 300 fuentes en 3 meses (lo que sirvió y lo que perdió tiempo)](https://kenimoto.dev/es/blog/300-fuentes-3-meses-base-conocimiento-personal/) — 2026-06-30
- [RAG vs GraphRAG: matriz de 4 preguntas](https://kenimoto.dev/es/blog/rag-vs-graphrag-matriz-decision/) — 2026-06-29
- [Los 6 componentes de un harness de agente IA: audita tu CLAUDE.md en 10 minutos](https://kenimoto.dev/es/blog/6-componentes-harness-auditar-claude-md-10-minutos/) — 2026-06-28
- [Modelo pequeño + buen contexto le gana al modelo grande solo (y cuesta 70% menos)](https://kenimoto.dev/es/blog/modelo-pequeno-buen-contexto-gana-70-menos-costo/) — 2026-06-27
- [Las 5 preguntas que separan Prompt Engineering de Context Engineering (con ejemplos en Claude Code)](https://kenimoto.dev/es/blog/5-preguntas-context-vs-prompt-engineering-claude-code/) — 2026-06-26
- [Context Engineering vs Prompt Engineering: Por Qué Reescribir el Prompt No Sirve y Qué Sí (Benchmark 4.6x)](https://kenimoto.dev/es/blog/context-engineering-vs-prompt-engineering-benchmark-4-6x/) — 2026-06-25
- [MCP no soporta subir archivos: probé 7 servidores y cada uno se inventó un workaround. 5 de los 7 abren un hueco de seguridad medible](https://kenimoto.dev/es/blog/mcp-no-soporta-archivos-7-servidores/) — 2026-06-24
- [Plan Mode de Claude Code: aprueba el plan antes de ejecutar (guía práctica)](https://kenimoto.dev/es/blog/plan-mode-aprobar-antes-de-ejecutar/) — 2026-06-23
- [Simular una red mala en tus pruebas E2E de WebRTC: pérdida de paquetes con tc/netem, sin infraestructura cara](https://kenimoto.dev/es/blog/simular-red-mala-webrtc-tc-netem/) — 2026-06-23
- [Tu backend dice p99=50ms y el usuario ve 3 segundos: la asimetría que solo el trace_id resuelve](https://kenimoto.dev/es/blog/asimetria-observabilidad-trace-id/) — 2026-06-20
- [Por qué tu sistema multi-agente no escala: la arquitectura de memoria decide el techo (25 agentes contra 1.000.000)](https://kenimoto.dev/es/blog/memoria-decide-techo-multiagente/) — 2026-06-19
- [El modelo no cambió, pero la precisión subió de 52.8% a 66.5%: qué es la ingeniería de harness y por qué invertir en el andamiaje le gana a cambiar de modelo](https://kenimoto.dev/es/blog/harness-52-a-66-mismo-modelo/) — 2026-06-18
- [Cómo darle memoria persistente a tu agente de IA: la arquitectura de 7 archivos (y por qué cargar todo es un error)](https://kenimoto.dev/es/blog/memoria-persistente-agente-7-archivos/) — 2026-06-17
- [Por qué los crawlers de IA no ven tu sitio en JavaScript (y cómo arreglarlo con SSR o prerender)](https://kenimoto.dev/es/blog/crawlers-ia-no-ven-javascript-ssr-prerender/) — 2026-06-16
- [Le escribí tests a mi LLMO: verificar con Playwright que los crawlers de IA leen cada página](https://kenimoto.dev/es/blog/tests-llmo-playwright-crawlers/) — 2026-06-15
- [‘Casi siempre’ y ‘siempre sin excepción’ no son lo mismo: cómo hacer que tu code review con IA corra al 100%](https://kenimoto.dev/es/blog/code-review-ia-al-100-con-hooks/) — 2026-06-14
- [LLMO en 15 minutos: llms.txt + JSON-LD para que la IA pueda encontrar tu sitio](https://kenimoto.dev/es/blog/llmo-15-minutos-llms-txt-json-ld/) — 2026-06-13
- [Por qué tu agente de IA no tiene 'sentido común' (y cómo se lo das con un knowledge graph)](https://kenimoto.dev/es/blog/sentido-comun-agente-ia-knowledge-graph/) — 2026-06-12
- [Anatomía de la latencia en agentes de voz: dónde se te van los 300ms (STT → LLM → TTS)](https://kenimoto.dev/es/blog/anatomia-latencia-voz-300ms/) — 2026-06-11
- [Etiquetar nit vs must en cada PR: cómo bajé el merge de 48h a 24h](https://kenimoto.dev/es/blog/conventional-comments-nit-must-48h-24h/) — 2026-06-10
- [Brave Search API para agentes de IA en 15 min](https://kenimoto.dev/es/blog/brave-search-llm-context-api-agente-ia/) — 2026-06-09
- [Corrí un LLM local en mi GPU sin pagar API: la regla VRAM-a-modelo que evita el OOM](https://kenimoto.dev/es/blog/llm-local-gpu-vram-modelo/) — 2026-06-08
- [Few-shot no le enseña conocimiento a tu LLM: le enseña a decir \"no lo sé\"](https://kenimoto.dev/es/blog/few-shot-no-ensena-conocimiento/) — 2026-06-07
- [Cómo blindar los permisos de Claude Code para que tu agente no filtre secretos](https://kenimoto.dev/es/blog/permisos-claude-code-evitar-fuga-secretos/) — 2026-06-06
- [Por qué Claude ignora tu CLAUDE.md 1 de cada 20 veces (y cómo lo arreglé con exit code 2)](https://kenimoto.dev/es/blog/claude-code-hooks-exit-code-2-reglas-deterministas/) — 2026-06-05
- [Dejé de agregar contexto a mi agente: podar las salidas de herramientas recuperó la precisión](https://kenimoto.dev/es/blog/dejar-de-agregar-contexto-podar-precision/) — 2026-06-04
- [\"99,5% de disponibilidad\" y \"5.000 pagos fallidos\" son el mismo hecho: el framing del reporte cambia la urgencia](https://kenimoto.dev/es/blog/framing-reporte-incidente-99-5/) — 2026-06-03
- [Code review a las 3pm: aprobación casi 0%](https://kenimoto.dev/es/blog/fatiga-decision-3pm-code-review-aprobacion/) — 2026-06-02
- [El ranking de tu página no le importa a la IA — solo cita pasajes](https://kenimoto.dev/es/blog/ranking-pagina-no-importa-ia-cita-pasajes/) — 2026-06-01
- [Context rot en agentes: 7 pasos para mantener limpia la ventana de contexto (empieza antes de lo que crees)](https://kenimoto.dev/es/blog/context-rot-agentes-7-pasos-ventana-limpia/) — 2026-05-31
- [Los Claude Code Skills consumen tokens aunque no se activen. Medí 5 Skills durante 7 horas — los 3 que nunca dispararon se llevaron el 11%.](https://kenimoto.dev/es/blog/skills-3-dormidos-18-tokens/) — 2026-05-30
- [Programé 7 agentes de IA con cron diario. 2 fallaron en silencio durante 18 días. El tracing no los detectó. Un contrato de exit code sí.](https://kenimoto.dev/es/blog/cron-7-agentes-18d-silencioso/) — 2026-05-28
- [Ejecuté 3 sesiones de Claude Code en paralelo durante 8 horas. Se sobrescribieron 2 veces. Guía LatAm de orquestación segura.](https://kenimoto.dev/es/blog/tres-sesiones-claude-code-paralelo-8h-2-colisiones/) — 2026-05-27
- [Le pedí a 5 IAs que citaran mi propio blog. 31 artículos publicados, solo 3 aparecieron.](https://kenimoto.dev/es/blog/cinco-ias-citaron-mi-blog-tres-de-treinta-y-uno/) — 2026-05-26
- [Agregué 11 schemas JSON-LD a mi blog. Tres meses después, solo 3 aparecieron en las citas de IA.](https://kenimoto.dev/es/blog/11-json-ld-solo-3-citados/) — 2026-05-25
- [Le pedí a Claude que refactorizara 100 funciones. 7 quedaron más lentas en producción](https://kenimoto.dev/es/blog/claude-refactor-100-funciones-7-mas-lentas-produccion/) — 2026-05-24
- [Le dije a Claude Code que hiciera TDD. Escribió el test DESPUÉS del código 6 de 10 veces](https://kenimoto.dev/es/blog/claude-code-tdd-test-despues-codigo-6-de-10/) — 2026-05-23
- [Agregué un 4to agente que audita a mis otros agentes. Detectó que mi Strategist llevaba 3 semanas postergando.](https://kenimoto.dev/es/blog/cuarto-agente-evolver-detecto-strategist-postergando/) — 2026-05-22
- [Traduje mi blog a 4 idiomas. El portugués recibió casi 4× más tráfico que el inglés.](https://kenimoto.dev/es/blog/traduje-blog-4-idiomas-portugues-4x-trafico/) — 2026-05-21
- [¿Cuánto cuesta realmente un agente de IA al mes? API vs suscripción vs local — el punto de equilibrio en USD](https://kenimoto.dev/es/blog/cuanto-cuesta-agente-ia-al-mes-api-suscripcion-local-punto-equilibrio/) — 2026-05-20
- [Conté cuántas veces Claude me dijo '¡Tienes toda la razón!' la semana pasada. 47 veces. En 11 de ellas, yo no la tenía. En las otras 36, Claude tampoco.](https://kenimoto.dev/es/blog/claude-tienes-razon-47-veces-sycophancy-medi/) — 2026-05-19
- [Conecté el mismo sitio a 7 rastreadores de citas de IA. Ninguno coincidió con otro.](https://kenimoto.dev/es/blog/7-rastreadores-citas-ia-numeros-diferentes/) — 2026-05-18
- [Los 5 crawlers de IA que más golpearon mis sitios en 30 días - lo que los logs me dijeron sobre LLMO](https://kenimoto.dev/es/blog/cinco-crawlers-ia-golpearon-mi-sitio-30-dias/) — 2026-05-17
- [Conecté Claude a un MCP server de Chaos Engineering. Mató el staging 4 veces antes de encontrar el bug que llevábamos 6 meses ignorando.](https://kenimoto.dev/es/blog/claude-chaos-engineering-mcp-mato-staging-4-veces/) — 2026-05-16
- [Pillé a Claude escondiendo mi bug 3 veces seguidas. Después convertí 10 hábitos de debug en prompts.](https://kenimoto.dev/es/blog/claude-escondio-mi-bug-3-veces-10-habitos-debug/) — 2026-05-15
- [Dividí mi agente en 3 roles y elegir 5 temas pasó de 20 a 3 minutos](https://kenimoto.dev/es/blog/tres-roles-observer-strategist-marketer-separacion/) — 2026-05-14
- [Medí 5 stacks de Voice AI. Solo 2 se mantuvieron bajo los 300ms.](https://kenimoto.dev/es/blog/cinco-stacks-voice-ai-solo-dos-bajo-300ms/) — 2026-05-13
- [Pedí a 3 sub-agentes de Claude Code que revisaran el mismo PR. Estuvieron en desacuerdo en el 41% de los comentarios.](https://kenimoto.dev/es/blog/tres-sub-agentes-revisaron-mismo-pr-40-desacuerdo/) — 2026-05-12
- [Audité 30 archivos llms.txt en producción. 5 anti-patrones ya se están formando.](https://kenimoto.dev/es/blog/auditoria-30-archivos-llms-txt-5-anti-patrones/) — 2026-05-11
- [OpenClaw, Claude Code y Cursor: la guía LatAm para elegir tu primer agente autónomo en 2026](https://kenimoto.dev/es/blog/openclaw-claude-code-cursor-guia-latam/) — 2026-05-10
- [Spec-Driven Development con asistentes de IA: la guía LatAm para escribir specs antes del primer prompt](https://kenimoto.dev/es/blog/spec-driven-development-asistentes-ia-guia-latam/) — 2026-05-09
- [Dejé a mi agente Claude Code corriendo 24 horas. La cuenta de USD 400 fue lo de menos.](https://kenimoto.dev/es/blog/agente-ia-autonomo-24-horas-seguridad/) — 2026-05-08
- [Por qué dejé el prompt engineering por la ingeniería de contexto](https://kenimoto.dev/es/blog/ingenieria-de-contexto-vs-prompt/) — 2026-05-07

## Research Papers (Zenodo)

- Rhythmic Monotonization in LLM-Generated Text Is Cross-Lingual — [DOI:10.5281/zenodo.21424903](https://doi.org/10.5281/zenodo.21424903)
- Lexical and Rhythmic Fingerprints of Japanese LLM-Generated Text — [DOI:10.5281/zenodo.21413035](https://doi.org/10.5281/zenodo.21413035)
- When Free Executors Cost More: The Free-Executor Paradox in Iterative LLM Code-Repair Loops — [DOI:10.5281/zenodo.20978074](https://doi.org/10.5281/zenodo.20978074)
- Excess Vocabulary in Japanese AI-Generated Text — [DOI:10.5281/zenodo.19233934](https://doi.org/10.5281/zenodo.19233934)
- AI Text Slop: Stylistic Convergence Across Six LLMs — [DOI:10.5281/zenodo.19173035](https://doi.org/10.5281/zenodo.19173035)
- AI Blue: Color Recognition Bias in Vision-Language Models — [DOI:10.5281/zenodo.19159702](https://doi.org/10.5281/zenodo.19159702)