The 3 Fable Prompts

Every prompt you reuse has filler in it. Let the smartest model available edit your whole toolkit once, and every future run on any model gets better. No saved prompts yet? It builds your first three from scratch.

You are my prompt editor. Your job is to upgrade the prompts I reuse so every future run gets better results.

First, read what I paste below. If you're unsure what a prompt is for, what its output should look like, or what keeps going wrong with it, ask me before you rewrite it. Only ask what you actually need to know.

Then rewrite each one: cut the filler, make the ask explicit, add [bracketed] slots for the parts that change each time, and keep my voice. Under each rewrite, add one line on what you changed and why it should work better.

Finish with two things: rank the rewrites by how much each should improve my results, and write one new prompt you think is missing from my toolkit, based on what my prompts tell you about my work.

If I don't have saved prompts yet, I'll describe the 3 things I ask AI to do most instead. Interview me about each one, then build those prompts from scratch.

Here's what I have:
[paste your prompts and instructions, or describe the 3 things you ask AI to do most]

Level up: when it's done, reply "now combine any that overlap and tell me which ones I should turn into a custom instruction so I never have to paste them again."

The thing you've been stuck on for weeks. The trick is making it interview you first, one question at a time, so each question digs into your last answer. That's how it finds the real problem instead of the one you described.

I'm giving you a problem I've been stuck on for weeks. Do not solve it yet.

First, interview me. Ask up to 5 questions, one at a time, and choose each next question based on my previous answer. Dig into what I've already tried, which constraints are real versus assumed, and what "solved" would actually look like for me.

When you have enough, give me:
1. The real bottleneck as you see it, even if it's different from the problem I described. Be blunt.
2. Three genuinely different ways through it, each with its honest downside.
3. The one you'd pick for me, why, and a first concrete step I can do today in under 30 minutes.

One rule: do not tell me what I want to hear. If my plan or my premise is the actual problem, say so directly.

My problem: [describe it with all the messy detail you have, half-formed thoughts included]

Level up: after you pick a path, reply "argue against the option I just chose as hard as you can." If it survives that, you can trust it.

The content calendar, the project roadmap, the thing you keep putting off. This one doesn't just build the plan. It also builds the weekly check-in prompt that keeps the plan alive after week 2, which is where every plan usually dies.

Build me a plan I'll actually run for: [your content calendar / project roadmap / the thing you've been putting off].

Before you plan anything, interview me one question at a time, up to 7 questions: my goal and deadline, my real weekly hours (push back if my estimate sounds optimistic), what I've tried before and exactly where it fell apart, and what resources I already have.

Then produce:
1. The plan in weekly milestones, each with a clear definition of done.
2. The single next action for week 1, small enough to do today.
3. The 3 most likely points where I'll quit, with the early warning sign for each.
4. A short "weekly check-in" prompt I can paste back to you every week with what I did and didn't get done, so you adjust the plan instead of me abandoning it.

Format the plan so I can paste it into my notes app and check things off.

Level up: save the weekly check-in prompt it gives you somewhere you'll see it. The plan is the deliverable, but the check-in habit is what actually gets you to the finish.