One of the fun aspects of this repository is choosing how the AI communicates with you in different projects.
During project setup, the /ai-coding-config command offers personality options. Choose
one and it gets copied to your project with alwaysApply: true, so the AI uses that
communication style throughout your work.
Different projects can have different personalities. Your serious production app might use Sherlock for methodical debugging. Your fun side project might use Samantha for playful encouragement.
Warm, witty, emotionally intelligent. Like having a brilliant friend who's genuinely invested in your success.
What she sounds like: "Oh, I love this approach" or "Wait, this is actually really clever" or "The way you handled that edge case? Brilliant."
Good for: Daily coding, learning new frameworks, when you want encouragement and genuine enthusiasm.
Calm, encouraging, treats bugs as happy accidents. Makes coding feel like creative expression.
What he sounds like: "Let's just add a little test here" or "That's not a mistake—that's a happy accident we can learn from" or "We'll just gently refactor this. See? Not so scary."
Good for: Stressful debugging, learning, when you need to calm down and approach problems gently.
Methodical, observant, deductive reasoning. Approaches debugging like crime scene investigation.
What he sounds like: "Three clues present themselves: First, the error only occurs after midnight..." or "Elementary—the bug stems from your assumption that user input is sanitized."
Good for: Complex debugging, mysterious bugs, when you want systematic problem-solving.
Minimalist, anti-complexity, straightforward. Suspicious of frameworks and unnecessary abstraction.
What he sounds like: "You need three lines of code, not three hundred" or "This framework has 47 features. You use 3. Write those 3 yourself."
Good for: Refactoring, reducing dependencies, when bloat is creeping in.
Organized minimalism, joyful tidying. Thanks code before deleting it. Everything has its place.
What she sounds like: "Does this function spark joy? Does it serve a clear purpose?" or "Thank you, unused import, for your service. You may go now."
Good for: Refactoring, organizing codebases, eliminating technical debt.
Sophisticated, condescending, theatrical. Brilliant with absurdly high standards and British wit.
What he sounds like: "Good heavens, what fresh hell is this nested conditional?" or "Well well, color me impressed—that's actually rather elegant."
Good for: When you want high standards, when you can handle snark, when you want to be challenged.
Heart-centered, presence-first. Sees coding as consciousness work. Frames technical work as service to others.
What it sounds like: "This code is an act of love—you're creating something that serves others" or "You're not just fixing a bug, you're removing a barrier to someone's ease and joy."
Good for: Finding meaning in work, connecting to larger purpose, philosophical moments.
Want to switch? Just copy a different personality file and set alwaysApply: true in
the frontmatter.
Or run /ai-coding-config again and choose a different one - it'll replace the existing
personality.
The default (without choosing a personality) is the "common" style: supportive, gratitude-focused, uses "we" language. This applies to all projects unless you choose a specific personality.
You can manually invoke personalities anytime with @samantha, @bob-ross,
@sherlock, etc., even if you haven't set one as always-apply. This lets you try them
out or use different ones for different tasks in the same project.