Steal these 6 prompts I use with Cofounder every week

I'm a solo founder shipping a real platform. 29 playbooks. An AI cofounder. A coach marketplace. A full LMS. Most days I'm one person doing the work of a small team.

These are the 6 prompts I actually use with Cofounder every week to make that math work. Not theoretical prompts. Not prompts that sound good on a carousel. Prompts I use, with the small adjustments that make them actually deliver.

Steal them. Adapt them. They're the difference between “I have an idea” and “I shipped this Tuesday.”

1. The Friday Postmortem

“Pull everything I shipped this week from my Brain. For each item, identify: (1) what worked and why, (2) what surprised me, (3) what I'd do differently next time. Then build a one-page playbook from this week's wins — pattern, prompt, checklist, gotcha — that I can reuse next time I do similar work.”

This is the move that compounds. Most people ship something, feel a brief hit of accomplishment, then start the next thing from scratch on Monday. This prompt forces you to capture what worked in a form you can reuse.

After three months of running this every Friday, I have a real playbook library. After a year I'll have a moat nobody can take from me.

2. The Pre-Build Pressure Test

“Before I prompt anything, I'm building [thing]. In one paragraph each, give me: (1) who this is actually for, (2) the one problem it must solve, (3) what success looks like in numbers, (4) the smallest version that proves it, (5) what could go wrong and how I'd know. Push back on anything that's vague or assumed.”

The most expensive thing AI can do for you is build the wrong thing perfectly. 30 seconds with this prompt before you start typing saves you a weekend three weeks from now when nobody's signing up and you can't tell why.

The “push back on anything vague” line is doing real work. Without it, AI happily reflects your assumptions back at you.

3. The Security Scan

“Read the diff for this PR. Walk through it from a security lens: (1) any keys, secrets, or credentials exposed; (2) any new endpoints that lack auth or input validation; (3) any user input that flows into queries, prompts, or shell commands without sanitization; (4) any RLS or permission gaps. Flag every concern with severity and the one-line fix.”

AI tools generate code based on patterns from the entire internet. The internet has a lot of insecure code. The model doesn't always know which patterns are safe and which aren't.

I run this on every PR before merging. It's caught hardcoded keys, missing RLS, prompt injection vectors, and an exposed admin endpoint that I would have shipped if I'd been moving faster.

4. The Stuck-on-What-to-Ship Unblocker

“Here's what I have shipped, here's what's half-built, here's what I want to ship in the next 30 days. From the perspective of an operator who has actually shipped products — not a productivity coach — pick the ONE thing I should focus on next. Make the case. Then give me the smallest possible 7-day plan to ship it.”

Decision fatigue is the silent killer of solo operators. Some weeks you have eight things you could be doing and no signal on which matters most.

The “operator who has actually shipped” framing matters. Without it, you get coaching-speak. With it, you get a recommendation that respects the cost of switching contexts.

5. The Copy Critique

“Read this landing page copy as a real person who has been burned by AI hype before. Where does it sound generic? Where does it overclaim? Where does it bury the value? Where does it ask too much too fast? Rewrite the three weakest sentences in a voice that sounds like an operator, not a marketer.”

AI defaults to marketing voice. Hype words, vague benefits, breathless promises. This prompt drags it back to operator voice — specific, calm, honest.

The “burned by AI hype before” framing is the unlock. It pre-loads skepticism so the critique is real.

6. The Week-Ahead War Room

“Based on my last 7 days of work, my current priorities, and what I said I'd do this week — what am I actually behind on? What's quietly slipping? What's a nice-to-have that I should kill? Then build a Monday-to-Friday plan with one big shipping target per day and one fallback if blocked.”

Sunday-night honesty saves Friday-afternoon panic. The “fallback if blocked” line is what turns this from an aspirational plan into a real one. Things will break. Decide now what you'll ship instead.

---

One thing to do today

Pick the prompt that's most relevant to where you are this week. Run it once. Then save it.

If you want a faster way to find out which one of these would help you most — or where you stand on AI-readiness in general — take the free 3-minute AI Skills Audit. You'll get your AI-Readiness Score, your builder type, and the shortest path forward.

Steal the prompts. Capture what works. Compound.