The Exploration
Before you write a proposal, talk it through. OpenSpec's explore mode is a thinking partner — it asks questions, draws diagrams, challenges assumptions, and helps you see what you actually need to build.
This lesson is different. Instead of code, you'll see a conversation.
How to enter
Run this in your project directory:
$ /opsx:exploreThat's it. You're now in exploration mode. The AI knows the rules: think deeply, visualize freely, ask questions, never write code.
What happened
I wanted to tell people about Spec Before You Ship. Three channels: Buttondown newsletter, my website, and direct messages. I had a draft, but something felt off.
So I entered exploration mode. Here's what happened:
Instead of writing copy, it asked questions. Who? What's the hook? What angles work best?
The a-ha moment
Here's what emerged from that conversation:
What might go in the announcement:
- What it is: Free 7-lesson site teaching OpenSpec workflow
- What you'll learn: How to write proposals, specs, design docs, and tasks
- Why it exists: Because most devs skip planning and pay the "chaos tax"
- Who it's for: Anyone who writes code and wants to ship with less waste
<input type="hidden" name="embed" value="1">.See what happened? I came in with one thing (announcement copy) and left with three things:
- Fixed the sidebar form
- A new lesson idea (The Exploration)
- Clear direction on what that lesson should cover
That's explore mode working. I didn't write code. I didn't create a proposal. And the best next steps emerged naturally.
Another example
Exploration isn't just for announcements. Here's a technical decision I explored:
SQLite. Not even close.
Why: Embedded in the binary, works offline, single file, no server needed.
Unless... is there a sync component?
Five minutes of exploration saved me from a bad architectural choice. The answer was obvious once the right questions were asked.
What exploration is NOT
- Not a requirement — You can skip it. Go straight to proposal if you want.
- Not a commitment — Nothing is saved until you decide to create a proposal.
- Not code — It will never write code. That's the rule.
- Not slow — It can be 5 minutes or 30. You decide.
Why this matters
Most developers skip from "I have an idea" to "Let me write code." Exploration mode sits in between — it's where better decisions happen.
When you enter exploration mode, you're not committing to anything. You're:
- Asking questions — not jumping to solutions
- Visualizing — diagrams, tradeoffs, comparisons
- Grounding — in your actual codebase, not theory
- Leaving when you want — it's optional, not a required step
What comes next
This lesson connects to everything after it:
- The Proposal — When exploration crystallizes into something worth pursuing, that's a proposal
- The Spec — What you learn in exploration shapes what goes in the spec
- The Design Doc — Tradeoffs you surface here become design decisions later
Exploration is where the thinking happens. The rest is just writing it down.