v4.2 shipped — 142 engineered prompts

Not generated.Engineered.

142 production-grade prompts built on the same methodologies used at Google and Meta — versioned, benchmarked, and engineered for output you can ship. No templates. No fluff. Just context that works.

7-day guarantee
Ex-Google ML
142 engineered prompts
Lifetime updates

2,847 operators shipping

Launch pricingv4.2.0

$17$49—65%

One-time. Lifetime updates.

142 engineered prompts across 9 domains
Versioned & benchmarked outputs
Chain-of-thought methodology baked in
Quarterly updates — lifetime access
Notion + Markdown + JSON formats
7-day no-questions-back guarantee

Secure checkout via Stripe

7-day refund

Used by data scientists at Stripe

2,847 operators shipping

Rated 4.9 / 5 on Gumroad

Built on methodologies from ex-Google ML

Zero fluff. Engineered context.

New prompts added quarterly

Used by data scientists at Stripe

2,847 operators shipping

Rated 4.9 / 5 on Gumroad

Built on methodologies from ex-Google ML

Zero fluff. Engineered context.

New prompts added quarterly

What you get

Here's exactly what's inside.

142 engineered prompts

Not templates. Every prompt has a role, context, constraints, and output spec.

9 real-world domains

Growth, product, sales, analytics, ops, writing, research, engineering, finance.

Benchmarked outputs

Every prompt has a measurable pass/fail. If it doesn't beat the baseline, it doesn't ship.

Notion + Markdown + JSON

Three formats included. Use the one that fits your workflow.

Lifetime updates

Every major GPT, Claude, or Gemini release triggers a new version.

Works with all major models

GPT-5, Claude Opus 4.7, Gemini 3.1 Pro. Model-specific variants included.

Instant download after purchase. No account needed.

Inside the pack

Each prompt is a small program.

Role, objective, methodology, guardrails, output contract. Read any prompt and you know what it will return — before you run it.

market-positioning.v3

1# Role & Context — load this first
2ROLE: "Senior positioning strategist (25y, B2B SaaS)"
3OBJECTIVE: extract {{unique_insight}} from {{product_dump}}
4
5# Methodology — April Dunford, adapted
6STEP_1 → enumerate alternatives (inc. 'do nothing')
7STEP_2 → cluster unique attributes → benefits
8STEP_3 → target who values them # <= ICP
9
10OUTPUT: markdown table + 3 narrative positions

cohort-analysis.v2

1# Expected input: CSV with user_id, event, ts
2ROLE: "Staff data scientist, retention-focused"
3
4COMPUTE(weekly_cohorts) →
5  • retention matrix (W0..W12)
6  • identify breakage week # statistical, not eyeballed
7  • segment by acquisition channel
8
9GUARDRAIL: reject if n < 200 per cohort
10OUTPUT: Markdown + annotated recommendations

feature-prd.v4

1# Replaces ad-hoc PRD writing. Amazon 6-pager style.
2ROLE: "Principal PM — thinks in pressure tests"
3
4# 6-section skeleton — fill each before moving on
5SECTION(1) — Problem // who, pain, frequency, cost
6SECTION(2) — Context & constraints
7SECTION(3) — Proposed solution (+ 2 alt)
8… (4) Metrics  (5) Risks  (6) Rollout
9
10OUTPUT: narrative doc, no bullets in body

Who it's for

Use it for the work you already do.

Pick your domain. Open the prompt. Get structured output.

Growth & Analytics

Retention diagnosis, funnel breakdowns, cohort analysis.

cohort-analysis.v2

Product & PRDs

Amazon 6-pager specs, feature briefs, rollout plans.

feature-prd.v4

Cold Outreach

Sequences built on proven frameworks, not guesswork.

outreach-sequence.v3

Pricing & Positioning

April Dunford methodology. ICP clarity in minutes.

market-positioning.v3

AI Automation

Multi-step chain-of-thought workflows that actually ship.

agent-orchestration.v1

Content & SEO

Clusters, editorial briefs, topical authority maps.

topical-map.v2

Same question. Two answers.

The difference is context.

Prompt: "How do I improve retention?"

Generic ChatGPT

Here are some strategies to improve user retention:
1.Send email reminders to inactive users
2.Create a loyalty program
3.Offer personalized experiences
4.Improve onboarding
5.Collect feedback and iterate

vagueno datano actionno metric

no context · no method · 14 sec

PromptPack · retention-diagnosis.v2

§1

Primary lever — Week-2 activation

Your D7 → D14 drop is 42% vs 28% benchmark. Root cause: users who skip step-3 onboarding churn 3.1× faster.

§2

Action (this sprint)

Gate feature discovery behind step-3 completion. Expected lift: +11pp D14 retention (p<0.05, n=2,400).

§3

Measurement

Instrument step_3_completed event. Read cohort retention at D14, not D7. Guard against novelty bias — wait 3 cohorts.

data-groundedmeasurabletime-boxedp-value

context loaded · methodology applied · 38 sec

Setup

From purchase to production in 2 minutes.

Buy

Stripe checkout. Instant download. No account, no onboarding.

Open

One-click Notion duplicate. Or drop the Markdown files into your editor.

Ship

Copy a prompt. Paste in ChatGPT, Claude, or Gemini. Get structured output.

Why it works

Three reasons this isn't another prompt pack.

Methodology

Built on how experts actually think.

Every prompt encodes a real methodology — April Dunford on positioning, Amazon 6-pagers for specs, cohort-first analytics. No generic 'act as an expert' prefixes. The reasoning is loaded in.

frameworks: 17
roles: ex-Google, ex-Meta

Versioning

Prompts that evolve, not rot.

Semantic versioning, changelogs, and a benchmark harness — every prompt has a measurable pass/fail. When a model drifts, we ship v+1. You get the diff and an explanation.

latest: v4.2.0
cadence: quarterly

Coverage

142 prompts across 9 domains.

Growth, analytics, product, engineering, ops, research, writing, design, and finance. If you ship work that depends on output quality, there's a pack for it.

prompts: 142
formats: MD · JSON · Notion

The difference

Most prompt packs look like this.

Other prompt packs

PromptPack

Vague 1-sentence templates

Role + Context + Constraints + Output spec

No versioning or changelog

Semantic versioning — v4.2.0, full diff on update

Same prompt for every model

Model-specific variants for GPT-5, Claude Opus 4.7, Gemini 3.1 Pro

No way to know if it works

Every prompt has a pass/fail benchmark

Abandoned after launch

Updated at every major model release

Who built this

Engineered by a team that shipped models you've used.

ex-Google · ML Engineerex-Meta · Data Scientist

Most prompt 'packs' on the internet are screenshots of ideas. We shipped production ML and data tools at FAANG scale — so we build prompts the way we'd build a feature: with a spec, a benchmark, and a rollback plan. If it doesn't beat the baseline, it doesn't ship.

Years in ML / data: 0
Prompts tested in prod: 0.0k+
Models evaluated: 0
Operators served: 0

Operators talking

Less prompting. More shipping.

4.9 / 5 · 320 reviews

“Replaced three of my go-to prompt templates in a week. The positioning pack alone saved me a two-day research cycle on our last launch.”

Maya Ortiz

Head of Growth, early-stage SaaS

−2 days per launch cycle

“I've seen a lot of 'prompt engineering' nonsense. This is the first one that reads like something an actual staff-level DS would write. The cohort-analysis prompt is unfairly good.”

Dan Hwang

Staff Data Scientist, fintech

+11pp forecast accuracy

“I stopped tweaking prompts and started shipping specs. The PRD-v4 prompt is now part of my team's working agreement.”

Priya Mehta

Group PM, B2B platform

3× PRD throughput

Try it for 7 days.

If it doesn't change how you prompt — get every dollar back. No questions asked.

No questions asked

Instant refund via Stripe

Keep the files

Joins 2,847 operators already inside.

Questions

Asked and answered.

A downloadable pack of 142 engineered prompts across 9 domains, delivered as Markdown + JSON + a Notion template. Every prompt is versioned, documented, and benchmarked. Lifetime updates included.

Stop prompting blind.

142 engineered prompts. Lifetime updates. $17, one time. Ship the work — skip the trial-and-error.

7-day refundInstant download · Stripe checkout

2,847 operators already in