~/blog/series/claude-pro
Claude Pro
Getting the most out of Claude Pro in day-to-day development.
What this series is about
This series is the practical counterweight to the skeptical one. The other series argues about how AI changes the discipline of software work. This series is about what you actually do at the keyboard once you have decided to use Claude Code in earnest.
The arc moves from setup to scaffolding. Getting the basic settings right so the tool stops fighting you. Writing a CLAUDE.md file that gives the agent the context it cannot infer.
Adding hooks so the things you care about happen deterministically, not because the model felt like it. Building skills and plugins so a workflow you ran once becomes a workflow you can hand to a teammate. Stopping the impulse to ask the agent nicely, and giving it rules instead.
These are not theoretical posts. Each one came out of a problem I hit and a pattern that turned out to hold up over weeks of real work. Some of them are short. The ones that are not are that length because the trap they describe is subtle, and the wrong shortcut is a regression you will only notice three weeks later when the bug report lands.
If you have stopped fighting your tools and want to start steering them, this is the series for it.
Where this is going
I keep ending up at the same question. What is the smallest set of practices that hold up across teams, not just for me? Next posts will lean harder in that direction: less "here is a setting I like," more "here is what survives when the second engineer joins."
Superpowers: teaching Claude Code to think before it types
Out of the box, Claude Code starts writing immediately. Superpowers forces it to plan, test, and verify first. On structured workflows, TDD enforcement, and why discipline beats speed.
Claude Code hooks: deterministic control over AI workflows
While claude.md instructions are treated as suggestions, Hooks provide deterministic guarantees. Learn how to use pre- and post-tool hooks to enforce formatting, block dangerous commands, and standardize your team's workflow.
The CLAUDE.md file: give your AI permanent memory
Every Claude Code session starts from zero unless you tell it otherwise. The CLAUDE.md file is how you give it persistent context about your project, your stack, and your preferences.
Stop asking your agent nicely
Your CLAUDE.md is a suggestion. Hooks are a wall. ADRs and custom lint rules are the missing layer between them, the difference between hoping your agent obeys and making it impossible to disobey.
What's new in Claude Code: notes from the London talk
A walkthrough of what Anthropic shipped in Claude Code recently, organised the way they presented it: developer experience and autonomy, with notes on which defaults are actually worth flipping on.
The best number in Opus 4.8 isn't a benchmark
Opus 4.8 beats GPT-5.5 and Gemini on SWE-bench, and that's the least interesting thing about it. The number that actually changes your work is the one Anthropic buried in softer language.
Stale memory is worse than no memory
Persistent memory is sold as a pure win. But a memory that records a temporary fact and never expires keeps steering your agent toward problems that no longer exist.
The agent is just a loop
An agent is a loop around a model with tools. If you use Claude Code you are already inside one, and you can drive it with hooks and slash commands, no SDK required.
Build an MCP server, then ask whether it should exist
A working MCP server is twenty lines of FastMCP. That is exactly the problem. A build tutorial, and the test for whether your server earns a place in the tool list at all.