Creator of the LLMO Framework. I design harnesses that automate dev, ops, and marketing with Claude Code — then publish what actually happened, with numbers.
400K+ article views (Qiita + Zenn) · 40+ books (Kindle + Zenn) · 4 research papers
Building AI-search-ready content? Start with the LLMO Framework.
If this work has saved you time, you can support its continued maintenance through GitHub Sponsors or Ko-fi.
| LLMO Framework | The open methodology for LLM Optimization (AEO + GEO) — getting your content found and cited by AI search |
| llmo-checker | Lighthouse-style CLI that scores how AI-retrievable a URL is: llms.txt, JSON-LD, robots policy, semantic structure → 0–100 LLMO Score |
A focused set of tools I use daily to keep Claude Code sessions productive across long work. Follow one, you'll want the others.
| compact-ops | Preserve session state across /compact — no lost context, no manual re-briefing |
| claude-shift | Switch between multiple Claude Code accounts safely; observe 5-hour / weekly usage across them |
| hook-chain-lens | Diagnose your Claude Code hooks: see the merged fire order across user / project / plugin scopes, catch dead matchers and missing script paths |
More: open-llmo (research initiative) · voice-clone (3-sec voice cloning on Qwen3-TTS) · speech-habit-lens (speech analyzer) · domain-pre-flight (domain checks) · persona-hub (persona SDK)
New products ship weekly: kenimoto.dev → Products
- When Free Executors Cost More: The Free-Executor Paradox in Iterative LLM Code-Repair Loops — DOI:10.5281/zenodo.20978074 · code
- Excess Vocabulary in Japanese AI-Generated Text — DOI:10.5281/zenodo.19233934 · code
- AI Text Slop: Stylistic Convergence Across Six LLMs — DOI:10.5281/zenodo.19173035 · code
- AI Blue: Color Recognition Bias in Vision-Language Models — DOI:10.5281/zenodo.19159702 · code
 Harness Engineering |
 Context Engineering |
 Claude Code |
 LLMO |
 Knowledge Graph |
 300ms Threshold |
Full catalog (EN / JA / PT / ES): kenimoto.dev → Publications
|
Kimi K3.1 Video QA Loop: 3 Rounds, 29 Minutes — 2026-11-02 php2shell Mitigation: 7 Steps Before Patching — 2026-10-22 Kimi K3 vs Claude: Video Review, Same Findings — 2026-10-14 Codex Hosted Search: 24/24 Only in Codex — 2026-10-14 Claude Code Max vs API vs Qwen: 4h Breakeven — 2026-09-09 More on kenimoto.dev |
OSS 303 PRを測ったら、pydanticの半数は他者レビューなしでマージ — 2026-11-04 BATNAを3つ持たないエンジニアの技術交渉は会議室で既に負けている — 2026-10-22 ChatGPT流入+8,337%の施策を分解したら、llms.txtじゃなかった — 2026-10-15 4つのLLMのウェブ検索を比べたら、Codexのキャッシュが全問正解 — 2026-10-14 CLAUDE.mdがあるとAGENTS.mdは黙って無視される【11通り実測】 — 2026-09-29 More on zenn.dev/kenimo49 |
パスワードを257文字打つだけでroot権限まで。*sshole、なんで？ — 2026-11-06 GraphRAG自作はJEMHopQA 120問でベクトル検索73.3%に負けた — 2026-11-02 MCP description の正規表現lintは、本物PoCを1本素通しした — 2026-11-02 作る・見る・聞くのマルチAIエージェントで動画検品、修正AIがズルを覚えた — 2026-11-02 PHP脆弱性php2shell対応、7.4本番を止めずに守った夜と実際の攻撃 — 2026-10-22 More on qiita.com/kenimo49 |
If my tools or research helped you, you can sponsor my open-source work.
This README rebuilds itself daily via GitHub Actions, pulling from kenimoto.dev/ai/publications.md (the AI-readable feed), Zenn, and Qiita. Last refreshed: 2026-11-08 06:04 JST