Commit

Skill personalisation We see new AI skills popping up and trending almost every week. We install them, try them out, and add the useful ones to our workflows. Which is fine. But how often do we personalise them? How often do we fork the repositories and fine-tune them to reflect how we actually work? You might have a skill that produces great code but doesn’t follow the single responsibility principle, so you keep pushing back on that. Or you prefer a particular shade of black for dark mode, but the skill keeps defaulting to something else. Individually, these are small corrections. When you’re making the same corrections every time, though, you’re repeatedly explaining preferences that could already be part of the skill. Customising a skill gives you somewhere to put those expectations: your coding style, architectural preferences, design choices, and process. Shorter prompts become more useful because you don’t have to repeat the same instructions, and there’s less back-and-forth over familiar issues. After looking at a number of skills I use and have used, I’ve decided it’s time to start personalising them to fit my own coding style and workflow. A skill can do a great job and still need adjustments to suit the person using it. After all, some developers prefer React class components; others prefer functional components. We shouldn’t expect every other preference to be universal either. Which instruction do you find yourself repeating that should probably live in your skills?