Claude Codeの Agent Skills は設定したほうがいい - じゃあ、おうちで学べる
Claude Codeを使い始めて、様々な発信をしてきました。今回は「Agent Skills」について。これも設定しておくと、Claude Codeがグッと使いやすくなる機能です。 Claude Code の settings.json は設定した方がいい - じゃあ、おうちで学べる Claude Code の CL…
Diving into Fizzy’s Routes: Rails’ resolve and direct
Exploring two underused Rails routing features — direct and resolve — through 37signals’ newly open-sourced Fizzy codebase. Learn how to create custom URL helpers and teach Rails to generate polymorphic URLs for models without their own routes.
Explore the Open-Closed Principle in Ruby through practical examples using inheritance and the Strategy pattern. Learn how to balance strict SOLID adherence with pragmatic design decisions.
[6/4] git worktrees with parallel agents in practice
This is not a comprehensive guide to Git worktrees, but I wanted to share how I'm currently using them to help you work more effectively. I also want to have a snapshot to compare against in a year, when multi-agent workflows will become more critical.
If you want to know
After the hundredth time explaining the same architecture to Claude, I finally admitted the obvious: we needed AI with actual memory. Here’s what we built and why it matters.
Rails Native Composite Primary Keys: A Complete Evolution from Rails 3 to Rails 8
Explore how Rails evolved to support composite primary keys natively in Rails 8, migration strategies from the composite_primary_keys gem, and best practices for modern multi tenant applications.
This year, Rails introduced a new method for ActionMailer called deliver_all_later. This feature simplifies enqueuing multiple emails for delivery through Active Job. Instead of sending each email individually, you can now enqueue a batch of emails, allowing them to be sent asynchronously when their respective jobs run.
任意のWebドキュメントをClaude Agent Skills化するツール site2skill を作りました。PAY.JPのドキュメントを例に、Claude Codeがドキュメントを参照しながら開発する流れを説明します。
GitHub - laiso/site2skillContribute to laiso/site2skill development by creating an account on GitHub.GitHublaiso
LLMが知らないライブラリを使うとき
LLMには知識のカットオフ(学習データの期限)があります。新しいライブラリやマイナーなAPIを使おうとすると、LLMは正確な情報を持っていません。例えばClaude Opus 4.5 の知識は2025年8月のものであり、それ以降にリリースされたライブラリや、学習データに含まれていないドキュメントについては、正確なコードを生成できません。
こういう場面では、Webのドキュメントを要約してチャットに貼り付け、それをもとに実装してもらうという作業を繰り返すことになります。この方法は有効ですが、毎回ドキ
b1rger/carl: carl is a calendar for the commandline. It tries to mimic the various cal(1) implementations out there, but also adds enhanced features like colors and ical support.
carl is a calendar for the commandline. It tries to mimic the various cal(1) implementations out there, but also adds enhanced features like colors and ical support. - b1rger/carl