Observability Is a Testing Strategy

Oct 29, 2026

4 min read quality , testing , observability , monitoring , sre , shift-right

A test is an assertion about how a system should behave. An alert is exactly the same thing - it just runs in production, against real traffic, forever.

Code Review Is a Social Technology

Aug 16, 2026

15 min read code-review , process , culture , teams , collaboration , ai

Every team I know does code review. Very few have ever asked themselves what they are actually doing when they do it.

How Much Is a Bug Worth?

Jul 29, 2026

5 min read quality , economics , process , testing , risk

A few evenings ago I was out for dinner with a friend who works as a developer. At some point the conversation drifted to work: how do you actually bring some quality into a company that doesn’t seem to want any? Which arguments land, and which ones fall flat.

minottobot: encoding a decade of QA experience into a Claude skill

May 1, 2026

3 min read qa , testing , ai , claude , skill , developer-experience

Back in 2016, my colleagues called me “minottobot” - out of both affection and frustration. Every pull request came back with meticulous comments: syntax, style, naming, indentation. I was a human linter. Relentless. Insufferable.

Years later, I moved deeper into quality engineering. The line-by-line work became irrelevant - it was treating symptoms, not disease. The real questions were: Why are tests flaky? How often can we actually deploy? Do developers understand what they’re maintaining?

But then AI got good.

I found myself asking: what if I could automate that old version of me? Not just syntax checking - anyone’s linter does that. But the reasoning. The strategy. The coaching. What if I could take a decade of QA knowledge, convert it into a decision-making process, and hand it off to Claude?

That’s where minottobot came from. It’s a Claude skill that audits an engineering team’s quality practices and proposes concrete, prioritized improvements.

Are you sure? Did you try it?

Apr 24, 2026

7 min read manual , testing , javascript , qa , code-coverage

There’s a question that echoes through every legacy codebase, every late-night deploy, every “it works on my machine” moment: did you actually try it? Not in a dismissive way - in the most honest, pragmatic sense. Did someone, with their own hands, open a browser and verify that the thing works the way it’s supposed to?