The Concept · Coda Curious

The Concept

The reasoning and mechanics behind Skill Files.

The Problem Worth Solving

Most people using AI with Coda get mediocre results — not because the AI is bad, but because it has no

clear context⁠

. It doesn’t know your table conventions, your formula patterns, or why you should never use a Select List where a relational table belongs.

Skill files solve this. They carry the accumulated judgement of someone who has made the mistakes, documented the lessons, and knows which Coda decisions survive contact with a real workspace.

What Coda Curious Is

Coda Curious is where Coda expertise becomes teachable. Not tutorials. Not tips. A structured method for encoding what an expert knows into files an AI can follow — so that any maker, at any coding level, can steer an AI to build, maintain, and improve a Coda doc with confidence.

Skill files: Plain-language instruction sets that give AI the architectural rules, formula patterns, and naming conventions of a seasoned Coda expert.

Core Coda competences: Each skill file maps to a real capability — formulas, data architecture, page structure, column design— turning expertise into context an AI can act on.

AI-ready by design: Whether you use Antigravity, Claude, or another AI assistant, the skill file is the bridge between your intent and a result that holds up.

⁠

What You Can Learn Here

🗂 Data architecture: How to model relational tables that stay clean as a workspace grows — and how to encode those decisions into a file an AI can follow.

✍️ Formula logic: The patterns behind thisRow, Filter(), Format(), and the formulas beginners fear. Not memorised — understood, then written into skill files.

📐 Page & UI standards: Naming conventions, callout structure, column lifecycle rules. The decisions that make a Coda doc legible to someone who didn’t build it.

🧠 Steering AI agents: How to write skill files that work — what to include, what to leave out, and how to verify the AI is actually following them.

How the Blogs Feed the Files

Each blog post is a working record of a problem encountered, a pattern tested, and a conclusion reached. Some are conceptual, while others are deeply practical. Over time, certain approaches have been

superseded by more efficient methods⁠

—where a better solution has emerged, I’ve noted the update in the

original post⁠

The blogs feed the skill files, and the skill files lead to higher-quality AI outputs. The goal is a loop of continuous improvement that any maker can enter at any point.

The full archive is on the

Blogs page.⁠

Use the search to find what you need.

⁠