I Fill My Substack Notes Queue for a Month in Just 30 Seconds - Here's How

Updated 2026-05-03: I’m rewriting this. The 2025 method has aged, and the workflow below is what I actually run now. I plan to keep updating this post when I find a sharper version of the workflow, not on a schedule, just when something is actually worth changing. Subscribe to catch the next rewrite.

I sit down at the editor. A Note has to go out today. The cursor blinks at me.

I open Claude. I describe the angle. I get back something that has the right idea but the wrong shape. Too long. No hook. I push back.

The next version is the right length but has lost the point. A chunk of the morning is gone and I have 1 Note that still isn’t shipping. Tomorrow I’ll need another one, and I’ll start over.

That loop is the actual problem. Not the blank page. The blank page is downstream of a workflow that resets every time.

I knew Notes mattered. I knew consistency mattered. The thing I couldn’t solve was the cost of writing each one from cold.

If any of that sounds familiar, you know the loop. Your AI did not actually solve the blank page. It just shifted where the friction lives.

The rest of this post is about a Skill I built to close that loop. Set it up once. Drop your context in. After that, filling close to a month of queue takes me about 30 seconds of typing.

Why Notes matter, and why most quit before they pay off

Showing up beats writing better

Most Substacks don’t stall because the writing is bad. They stall because the writer disappears.

People stay subscribed to people who keep showing up in their feed. If you only post the long piece every 2 weeks, you exist for those 2 days a month. The other 28, you are not in anyone’s head.

Notes are how you stay there without being pushy. Small, frequent, low-pressure ways to show up. They keep your name attached to a face the algorithm and your readers can both recognise.

Your Activity page is a portfolio

There is a second thing happening on Notes that nobody warns you about. Your Activity page is a portfolio. When a stranger lands on your profile, your Notes activity is one of the first things they see.

A consistent stream reads as someone who is showing up to their own thing. An empty stretch reads as someone who started something and stopped.

Both are honest signals. You’re sending one of them whether you mean to or not.

Day 30 is where it changes

The first weeks of regular posting feel like nothing happens. You publish a Note. It gets 2 likes. You publish another. It gets 1. You start to wonder if this is for you. That phase is normal. The platform is learning who you are. Your readers haven’t built a reflex yet to look for you. The early flat line is information, not failure.

From what I have watched, somewhere around the first month of regular posting things start to behave differently. Early followers begin sharing. The platform begins matching you to readers who actually like what you write. The feed starts giving you a little more rope.

I want to be honest. This is observation, not promise.

The condition is, you have to actually still be posting at day 30. Most people are not. The reason most people are not is the loop in the previous section.

So the math is clear. The bottleneck is supply. You can’t be consistent if every Note costs 20 minutes of mental friction. The strategy that works on paper requires a queue you can fill on demand. The Skill in the next sections is how I make filling that queue survivable.

Why I built this Skill

Knowing the math didn’t fix it. I read the consistency advice. I agreed with the consistency advice. I still wasn’t writing 2 Notes a day.

The bottleneck wasn’t strategy. It was the act of starting from cold every single time.

Every Note meant repeating the context. Who I am. Who reads me. What kind of Note this should be. What angle I’m hitting. By the time the AI had a workable picture, I had already typed a wall of preamble.

The output came back close but off. I’d send a fix. The next version was off in a different way. By the second Note of the day, I was already exhausted.

That’s not a tool problem. That is me using a generic tool for a job I do every day, with the same shape, against the same audience.

At some point I sat down and wrote the strategist work into a Skill. (Think of a Skill, for now, as a recipe an AI follows when I invoke it. The next two sections unpack what a Skill is in general, then the one I built.)

Now, instead of wrestling a generic AI for each Note, I drop my context in once and the Skill takes over.

But before showing you what the Skill does, here’s the fast version of what a Skill actually is. If you have not used one yet, this is the bridge.

What a Skill actually is

A Skill is a folder of instructions an AI loads when the work in front of it matches. Anthropic introduced the format for Claude. It is now an open standard that Codex, Cursor, Copilot, Gemini CLI and others have adopted.

Plainer: a saved set of instructions you stop having to re-teach every time.

Every time you type a prompt, you re-supply the context, the rules, the format. You teach the model from scratch. A Skill captures that teaching once and the AI applies it again and again. You stop being the person who has to remember the recipe every time.

Notes are repeating work with a stable shape. 50 to 250 words. 8 typical formats. Conversation-driven CTAs. The same set of decisions on every Note. That is the textbook case for a Skill. Define the shape once. Run it forever.

Not every workflow needs a Skill. A one-off task is fine as a prompt. Skills earn their keep on the work you do regularly, where the rules keep slipping out of your head and you keep retyping them. Notes are that work for a Substacker. So this is where I started.

Meet Substack-notes Skill

Substack-notes Skills is the strategist work I kept redoing every Note, written down once.

Inside it: the 8 Note types I actually recognise on Substack, the format rules for each, a writing style that works for Notes, and the anti-fabrication rules I have to keep restating because the AI keeps slipping on them. It works like an assistant that already knows the rules. You don’t re-teach the model what a good Note looks like. It asks you the few questions it needs to write one well.

It does not run from a fixed template. Every time you invoke it, it reads the context you gave it and shapes its questions around the gap.

A few ways I run it:

• “Write me a Note.” Drop nothing else. The Skill asks who it is for, what mood you want the reader to leave with, and proposes a type before drafting.

• “Repurpose this long-form.” Hand it a post you already published. It pulls the strongest line, shapes a teaser, and links back.

• “Write to this audience.” Layer a niche on top of the base reader profile for a single piece. Coaches today, freelancers tomorrow.

• “Write on this topic, with this feeling, ending in this CTA.” Specify the angle as tight as you want. The Skill fills the format around it.

Same Skill across all of them. The shape is fixed. The input is yours.

Less re-typing each Note. Fewer back-and-forth fixes. The rules stopped living in your head and got moved into the Skill.

What it looks like in practice

I typed one line: invoke the Skill, ask for 10 Notes, point at my substack-info folder.

The Skill read its own rules and the context in my folder. Then it stopped.

It didn’t draft yet. It batched 3 decisions back to me, each with a recommendation and a one-line trade-off:

• Goal. Subscriber conversion, reach and engagement, authority, or a balanced pack. The answer drives the type mix.

• Topic. Range across SoloAILab themes, run one belief at 10 angles, or funnel to one specific post.

• Material. Abstract beliefs only, raw moments I’ll feed in next, or skip the story-shaped types entirely.

That set is not a fixed checklist. The Skill reads the context I gave it and asks only the focused questions it needs to write the batch well. A thinner context means more questions. A richer one, fewer.

Once I picked, the Skill drafted. The header on the output told me what it had decided: 4 one-liner, 3 question-prompt, 3 list-format. No story types, because I told it I had no fresh anecdotes to feed in. No contrarian framing, because my audience profile rules it out.

What came back stayed inside what my profiles said. No invented stats. No fake stories. One honest caveat. I still skim every Note before they go in the queue. The Skill writes. I still hit publish.

What the Skill does not do

I want to be clear about what this Skill does not do. It makes Note-writing fast. It does not make me want to write Notes.

It cannot sit down for me. It cannot decide what I have to say this month. It cannot replace the 30 seconds of opening Substack, pasting the Notes in, and clicking publish.

What it does do is remove 3 frictions. The editing-loop tax. The repeat-yourself-to-the-AI tax. The wait-what-type-of-Note-is-this hesitation. 3 frictions gone. The motivation, you still bring.

Less effort, not no effort. If today is a day you don’t even want to open the editor, no tool fixes that. The Skill is for the day you would write if getting started were cheaper. That is most days, in my experience. It is not every day.

Try it once

The Skill is open source. Repo and install instructions are at https://github.com/Solo-AI-Lab/substack-notes-skills.

If you are on Substack and stalled on Notes, the cheapest way to find out whether this approach works for you is to try it once. Drop your context in. Ask for what you need today.

The first try tells you faster than this article does. If the output reads as Notes you would actually publish, keep going. If it doesn’t, your context files are not yet sharp enough, and that is a different and more solvable problem than writing each Note from cold.

---

If this saved you a step, restack it. That is how the next Substacker stuck on the same loop finds it.

And if you want the next version of this post, subscribe. I rewrite it when the workflow shifts, not on a schedule.