The Complete Field Manual
This guide exists because a sharp student looked at an AI-built job-search system and asked four perfect questions: Which AI made this? Where did the data come from? Was the intent predefined? And how do you even prompt it?
Those four questions are the best possible starting line, because they cover the entire territory: the tool, the data, the logic, and the language you use to command it. This manual answers all four in the first six pages — then spends 124 more turning you from a curious reader into someone who can build things that make people ask "wait, who made this for you?"
Who this is for
| You are… | Start at | You'll get |
|---|---|---|
| Completely new (student, parent, curious) | Page 1 → straight through | Confident daily use within a week |
| Already chatting with AI sometimes | Part 3 (Prompting Mastery, p31) | 10× better answers immediately |
| Builder / professional | Part 4 (Claude Code, p43) | Real systems: sites, databases, automation |
| Want income from home | Part 9 (p116) — but read Part 3 first | Ten honest, working business models |
How the guide works
- One page at a time. Use the sidebar, the Next/Previous buttons, or your keyboard ← → arrows.
- Your progress is saved in this browser automatically. Close it, come back next week — it remembers where you were and shows ✓ on every page you've read.
- Level badges on every page: BASIC INTERMEDIATE ADVANCED ELITE. Nothing advanced is required to enjoy the basics.
- Copy buttons on every prompt — click, paste into Claude, watch it work.
- Honesty labels on factual claims: ✅ verified · 📊 calculated · ⚠️ estimate / changes over time — always re-check prices and limits on official pages, because AI products update monthly.
Question 1: "Which AI tool or LLM was used to make this?"
Short answer: Claude, made by Anthropic — used in two different forms, plus a few non-AI tools doing the boring jobs.
The exact toolchain behind the job-hunt cockpit
| Layer | Tool | What it did | AI? |
|---|---|---|---|
| The brain | Claude (Anthropic's LLM — Opus/Sonnet-class models, and this guide by the Claude 5 family) | Understood the goal, designed the system, wrote every message draft | ✅ Yes |
| The hands | Claude Code (Claude running on a computer with permission to create files & run programs) | Wrote the actual files: Python scripts, the database, the web pages | ✅ Yes |
| The engine room | Python (free programming language) | Runs the rules: who gets which pitch, daily caps, quality checks | ❌ Just code |
| The filing cabinet | SQLite (free database) | Stores all 1,221 contacts in one file | ❌ Just storage |
| The publisher | Cloudflare Pages (free hosting) | Puts the finished pages on the internet | ❌ Just hosting |
So what exactly is an "LLM"?
LLM = Large Language Model. It is a computer program trained by reading a gigantic amount of text — books, articles, code, conversations — until it becomes extraordinarily good at one deceptively simple skill: predicting what text should come next. Ask it a question, and "what comes next" is a good answer. Describe a website, and "what comes next" is the code for that website. That one skill, scaled up massively, is what looks like intelligence. (Full explanation with zero jargon: page 7.)
Chat vs. Claude Code — the difference that matters
💬 Claude (the app)
A conversation window — like messaging a brilliant advisor. It can read files you give it, search the web, write documents. It advises and drafts.
🛠 Claude Code
The same brain, but sitting inside a computer with permission to act: create files, edit them, run programs, build databases, deploy websites. It does the work itself. The cockpit was built this way.
Question 2: "Where is the hiring people's email data fetched from?"
Brilliant question — because AI does not magically know anyone's email address. The honest answer is a pipeline: several boring, legitimate steps chained together. Here is exactly how professional contact data is gathered in systems like the cockpit:
LinkedIn is the world's public directory of professionals. A paid search seat lets you filter precisely: "Project Directors and Program Directors, energy sector, Saudi Arabia + UAE". People chose to publish these profiles publicly — name, title, company.
Scraping tools (e.g. actors on the Apify platform) turn those public search pages into a clean spreadsheet: name, title, company, LinkedIn URL. No passwords, no private data — only what is publicly visible.
A script visits each company's public website and news pages to collect "hooks" — real facts (projects won, expansions) used later to personalize messages honestly.
Companies use predictable formats:
firstname.lastname@company.com, f.lastname@…. Given a name + the company's known pattern, you can construct the likely address. Specialized lookup services do this at scale.A verification service technically asks the mail server "does this mailbox exist?" without sending anything. Only addresses that pass as safe are used. Dead, risky, or disposable addresses are suppressed permanently.
Everything lands in one database file on the owner's own computer — not in some cloud CRM — and the cockpit web page is generated from it.
The ethics layer (this part matters)
- Only publicly published professional data — the person chose to list their role publicly.
- Business-to-business context — a relevant professional matter (a senior candidate introducing himself), not consumer spam.
- Verification before sending — protects both sides from bounced junk.
- Opt-out respected instantly and permanently — one "no" suppresses the contact forever.
- Never bought from shady "database sellers" — purchased lists are where spam, wrong data, and legal trouble live.
Question 3: "Is the intent predefined?"
This is the most sophisticated of the four questions, and the answer teaches you how all serious AI systems are built. The answer is: the strategy is predefined in code; the wording is AI-generated within strict rails. Two layers, working together:
Layer 1 — The rails (predefined, written as Python rules, never improvised)
| Rule | What it enforces | Why it's code, not AI |
|---|---|---|
| The pitch ladder | If the recipient is a Project Director or Senior PM → introduce the candidate as "Project Manager". If they're Program/Portfolio level → introduce as "Project Director". Never pitch above the recipient's level. | This is positioning psychology — a senior recipient must never feel threatened or leapfrogged. Too important to leave to chance. |
| Daily caps | Max 18 emails + 25 LinkedIn actions per day | Protects the sender's accounts and keeps quality high |
| QC gates | Subject lines ≤ 65 characters; broken company names get company-less copy; violations hard-fail the build | A checker that can't be sweet-talked |
| Follow-up rail | Day 4 and Day 8 follow-ups, pre-written per band | Persistence by design, not mood |
Layer 2 — The generation (AI, within those rails)
Inside the rails, Claude wrote each message's actual wording — personalized using the verified facts collected in step 3 of the data pipeline: the person's real role, their company's real projects. So every email is individually written, but no email can break the strategy.
❌ Pure template systems
"Dear {FirstName}, I saw your company {Company}…" — robotic, ignored, sometimes embarrassing when a field is broken.
❌ Pure free AI
Fluent but unpredictable — might over-promise, mis-title the candidate, invent facts. Unacceptable for someone's career.
Question 4: "What prompt do we give, and how?"
A prompt is simply your instruction to the AI, written in normal language. There is no secret syntax. But there is an enormous difference between a weak prompt and a strong one — and it comes down to one idea: a prompt is a brief, not a wish. You are briefing a brilliant new employee who knows nothing about your situation.
The 5-part anatomy of a strong prompt
| Part | Question it answers | Example fragment |
|---|---|---|
| 1. Role | Who should Claude be? | "You are an experienced career coach…" |
| 2. Context | What's the situation? | "…my father has 34 years in power infrastructure, targeting Project Director roles in the Gulf…" |
| 3. Task | What exactly do you want? | "…write a 120-word LinkedIn message to a hiring director…" |
| 4. Format | What should the output look like? | "…give 3 versions in a table: formal, warm, brief…" |
| 5. Constraints | What rules must it follow? | "…no exaggeration, no salary talk, subject under 65 characters." |
A real prompt in the cockpit's style
You are a senior career strategist. Context: a Project Director with 34 years
in power transmission (utility + consultancy sides) is quietly job-hunting in
the Gulf. Task: write a first-contact email to a Programs Director at a large
utility, introducing him for senior project leadership roles. Format: subject
line (max 65 characters) + 110-130 word body + one-line PS. Constraints:
respectful peer-to-peer tone, zero flattery, zero exaggeration, mention ONE
specific credibility fact, end with a soft ask ("open to a brief call").Starter prompts for a student (copy any of these)
Explain [topic] like I'm 14, in 5 short points, then quiz me with 3 questions one at a time. Wait for my answer before revealing the next.
Here is my essay draft: [paste]. Don't rewrite it. Give me the 5 highest-impact improvements ranked, each with the exact sentence to change and why.
I want to build [idea] for my school project. Interview me: ask me 5 questions one by one to sharpen the idea, THEN give me a plan with materials and steps.
The 10-Minute Quick Start
Theory later. Right now, get your hands on it. Ten minutes, five moves, and you'll understand more than most people who've "heard about AI" for two years.
What a Large Language Model Actually Is
Strip away the hype and an LLM is one idea executed at planetary scale: a prediction machine for text.
The three stages, no jargon
Why "just predicting text" feels like intelligence
Because to predict text well, you must model the world that produced the text. To continue "The engineer calculated the cable sag because…" you need physics. To continue a chess game's notation you need chess. Prediction at sufficient depth forces understanding-like behavior. Whether it "really understands" is a philosophy debate; what matters practically is that it behaves like a tireless, widely-read collaborator.
What it fundamentally cannot do (unless given tools)
| Cannot… | Because… | The fix |
|---|---|---|
| Know your private files | It only sees what's in the conversation | Upload/connect them (p18, p26) |
| Know today's news by itself | Training has a cutoff date | Web search tool (p23) |
| Do actions in the world | It's text-in, text-out | Tools & agents — Claude Code (p43) |
| Be certain of facts | It reconstructs, not retrieves | Verification habits (p14, p76) |
The Context Window — Claude's Working Memory
The context window is everything the model can "see" right now: your messages, its replies, uploaded files, instructions. Think of it as a desk: enormous, but finite. Claude's desk holds about 200,000 tokens ⚠️ (some versions offer 1 million in beta) — roughly 500 pages of text. ⚠️ figures change; the principle doesn't.
The consequences you'll actually feel
| Symptom | Cause | Professional habit |
|---|---|---|
| Long chat starts "forgetting" early details | Old content summarized/compressed as the desk fills | One chat = one task. New task, new chat. |
| Quality dips in marathon sessions | Attention spread across too much material | Ask for a summary, start fresh with it pasted in |
| Repeating your preferences every time | New chat = empty desk | Projects & memory (p20-21), CLAUDE.md files (p49) |
| "It ignored the middle of my huge document" | Middles get less attention than edges | Put the document FIRST, your question LAST (p35) |
The three memory layers (the mental model that unlocks everything)
1. The desk
The context window. Fast, huge, but wiped between chats.
2. The notebook
Persistent instructions it re-reads every session: Project instructions, memory, CLAUDE.md rule files. You write these once; they pay forever.
3. The filing room
Your actual files and databases, fetched on demand via uploads, connectors, or Claude Code reading your disk.
Anthropic — The Company Behind Claude
Knowing who builds your tools tells you what to expect from them. Claude is made by Anthropic, an AI safety company headquartered in San Francisco.
The story in five beats ⚠️ (as of this guide's writing)
Why the safety-first origin matters to YOU practically
- Fewer fabrications, more "I'm not sure" — the honesty tuning is real and noticeable in daily work.
- Refusals are principled, not random — it declines harmful requests but helps with legitimate dual-use work (security testing, medical information) when context is clear.
- Enterprise trust — the same properties that protect users make it deployable in banks, hospitals, law firms — and in your business.
The Model Family — Which Claude Is Which
"Claude" is a family. The tiers are named like poetry sizes — small, medium, large — and choosing the right one is a real skill (and a real cost lever, p61/p68).
The tiers ⚠️ (as of mid-2026 — the lineup evolves; check claude.com)
| Model | Character | Reach for it when… |
|---|---|---|
| Haiku (4.5) | Fastest, cheapest | High-volume simple jobs: classify, extract, summarize thousands of items |
| Sonnet (4.5 / 5) | The workhorse — near-flagship brains at mid price | Daily driver: writing, analysis, most coding, most business tasks |
| Opus (4.x / 5) | Deep reasoner | Hard problems: complex builds, subtle strategy, long autonomous work |
| Fable 5 ✦ | New flagship tier above Opus (Claude 5 family) | The hardest work — and, transparently: this guide is being written by Fable 5 right now |
Practical selection rules
- Default to the best model you can afford for thinking-heavy work — a better first draft saves three rounds of fixes. Quality compounds.
- Downshift for volume — when the task is "do this simple thing 5,000 times", Haiku-class models at a fraction of the cost win (batch tricks: p68).
- Extended thinking (p33) matters as much as tier for hard logic — a Sonnet thinking carefully often beats an Opus answering instantly.
- In the apps you pick from a model menu; in the API you use exact IDs like
claude-sonnet-5orclaude-haiku-4-5⚠️.
Claude vs ChatGPT vs Gemini (and "GPTs") — The Honest Comparison
You'll be asked this at every family dinner. Here's the fair answer from someone who uses them all — no fanboy talk. ⚠️ Capabilities shift monthly; the shapes below have been stable.
| Claude (Anthropic) | ChatGPT (OpenAI) | Gemini (Google) | |
|---|---|---|---|
| Standout strengths | Long documents; sustained reasoning; honesty/care; agentic coding (Claude Code is the category leader); Artifacts; writing quality | Biggest ecosystem; image generation; voice; custom GPT store; broad plugin culture | Google-product integration (Gmail/Docs/YouTube); huge free tier; strong multimodal/video |
| Where it's weaker | No image generation; smaller consumer ecosystem | Long-context reliability historically choppier; agent coding tools younger | Agentic tooling younger; quality varies by task |
| Feels like | A senior colleague | A swiss-army consumer app | A Google feature everywhere |
| Pick it for | Serious work: analysis, writing, building systems, code | Consumer creativity, images, casual breadth | If your life runs entirely on Google apps |
"What are GPTs?" (the daughter's word — decoded)
Custom GPTs are ChatGPT's mini configured chatbots: someone packages instructions + files + a name ("CV Fixer", "Math Tutor") and shares it in a store. Claude's equivalents are Projects (p20 — private configured workspaces), Skills (p51 — packaged expertise Claude auto-loads), and Artifacts you can share (p19). Same idea — reusable, pre-briefed AI — different packaging.
The founder's actual answer
The Seven Places Claude Lives
Same brain, seven bodies. Knowing which surface to use for which job is an instant productivity doubler (full decision guide: p70).
| Surface | What it is | Best for |
|---|---|---|
| 1. claude.ai (web) | The chat in any browser | Everyday work, files, Artifacts, Projects |
| 2. Desktop apps (Win/Mac) | Chat + quick-launch keyboard shortcut + local MCP connectors | Power users; connecting local tools |
| 3. Mobile apps + voice | iPhone/Android; camera + dictation + spoken conversations | Capture on the go; thinking out loud (p25) |
| 4. Claude in Chrome | A sidebar living inside your browser, seeing the page you're on | Summarize/act on any webpage, draft replies in web apps (p27) |
| 5. Claude Code | Terminal + desktop + VS Code/JetBrains + web (claude.ai/code) | Building: files, code, databases, automation — Part 4, the heart of this book |
| 6. The API | Claude inside YOUR programs | Products and pipelines (p67-68) |
| 7. Inside other tools | Connectors/integrations: Excel, Slack, Notion, cloud platforms | Meeting your work where it already happens (p26, p63) |
Plans, Pricing & How Limits Actually Work
Subscription plans (the apps) ⚠️
| Plan | ≈ Price | Who it's for | Reality check |
|---|---|---|---|
| Free | $0 | Learning, light daily use | Genuinely useful; tighter message limits, smaller model access |
| Pro | ~$20/mo | Regular personal/professional use | The sweet spot for most people; includes Claude Code basics, Projects, more usage |
| Max | ~$100 or ~$200/mo | Claude as a daily coworker | 5×–20× usage, priority, heavy Claude Code — what serious operators run |
| Team / Enterprise | ~$25–30+/seat | Companies | Shared workspaces, admin, SSO |
How limits behave (the part nobody explains)
- Usage flows in rolling windows (e.g. a 5-hour window, plus weekly ceilings on heavy plans ⚠️). Hit a cap → wait for the window to reset. Plan heavy builds accordingly.
- Longer conversations burn more — each reply re-reads the whole chat. Another reason for "one chat, one task" (p8).
- Model choice matters: Opus-class replies cost several times a Sonnet reply's allowance.
The API (pay-per-use) ⚠️
For builders: you pay per million tokens, roughly — Haiku ~$1 in / $5 out; Sonnet ~$3 / $15; Opus-class ~$5–15 / $25–75. Two built-in discounts pros exploit: Batch API −50% and prompt caching up to −90% on repeated context (both explained p68).
Hallucination — Why AI Gets Things Wrong, and the Mindset That Fixes It
Sometimes an LLM states false things with perfect confidence. Not lying — reconstructing. It generates what should be true given the patterns, and usually is. The 2% where it isn't is where amateurs get burned and professionals build habits.
The high-risk zones (memorize these)
| Risk zone | Why it fails | Your move |
|---|---|---|
| Specific citations, quotes, links | Plausible references get assembled from fragments | Demand real search with sources (p23), click them |
| Precise numbers, prices, dates | Training data ages; details blur | "Search before stating" — or label it an estimate |
| Niche people & small companies | Thin training signal | Verify against the actual website/profile |
| Long calculations done "in its head" | Prediction ≠ arithmetic | Ask it to compute with code/analysis tool (p24) |
| "Does feature X exist in tool Y?" | Products change monthly | Check docs; ask Claude to fetch current docs |
The three-question verification habit
Privacy & Safety — The Sensible Rules
You'll hear both "never tell AI anything!" and people pasting their whole life into chats. Both are wrong. Here's the professional middle.
Never paste these (into any AI, ever)
- Passwords, API keys, OTP codes, recovery phrases — nothing that unlocks anything
- Full card numbers / CVV, full bank credentials
- Government ID numbers when avoidable
- Other people's private data they didn't consent to share
Sensible practice for everything else
| Data | Verdict | Notes |
|---|---|---|
| Your drafts, plans, essays, code | ✅ Fine | This is the whole point |
| Business documents | ✅ Generally fine | Check employer policy; use Team/Enterprise for company-wide use |
| Health/legal questions | ✅ Useful | Educational support, not a licensed professional; verify with one |
| Contracts with others' names | ⚠️ Redact when you can | "[CLIENT]" works fine for analysis |
Know your settings & rights
- Review privacy/training settings in the app and set them to your comfort ⚠️ (options evolve — check Settings → Privacy).
- Claude Code note: your files are read locally on your machine; content goes to the model only as needed for the conversation. Keep secrets in
.envfiles and tell Claude to never print them (p103). - Delete conversations you consider sensitive; use incognito-style chats for one-offs ⚠️.
The Five Mental Models That Make You Great With AI
Tools change monthly; mental models compound for years. These five run underneath every page that follows — this page is the skeleton key of the whole book.
1. 🧠 The brilliant intern with amnesia
Claude is a top-1% graduate who read the whole internet — and forgets you completely between conversations. So: brief it like it's day one, every time — or better, build the briefing once (Projects p20, CLAUDE.md p49) so it's auto-delivered. People who complain "AI gives generic answers" are giving generic briefs.
2. 🛤 Rails + engine + gate
From cockpit Question 3 (p4): deterministic rules decide what MAY happen, AI generates within them, a quality gate checks what DID happen. Any time you automate something serious, ask: where are my rails? where's my gate?
3. ⚡ The draft machine
AI's superpower isn't final answers — it's first drafts in seconds: documents, plans, code, options. You direct, it drafts, you judge, it refines. The loop "draft → critique → revise" (p37) is the single highest-ROI technique in this book.
4. 📈 The compounding notebook
Every preference you write down, every mistake turned into a rule, every prompt saved to a library makes YOUR Claude permanently better — while everyone else restarts from zero each chat. Part 6 shows a 50-rule system running real businesses; page 75 gives you the starter version.
5. 🕵️ Verify like a CFO
Trust the work, audit the claims. Labels (✅📊⚠️❓), sources on demand, one cheap check before anything important ships (p14, p76). Enthusiasm plus auditing is how you get the upside without the burns.
Your First Real Chat — Weak Ask vs. Strong Ask
The gap between a mediocre AI experience and a magical one is usually 20 extra seconds of asking properly. Watch the same three tasks done twice:
Side-by-side: the 20-second difference
| ❌ Weak ask | ✅ Strong ask | What changed |
|---|---|---|
| "Write an email to my landlord" | "Write a firm but polite email to my landlord: the AC has been broken 10 days, two complaints ignored, rent due Friday. Goal: repair scheduled this week. 120 words, no legal threats yet." | Situation + goal + tone + length |
| "Help me study biology" | "I have a biology exam in 6 days on genetics (DNA replication, Mendel, mutations). 1 hour/day. Build a 6-day plan with daily topics, a 10-question self-quiz each day, and a final mock test." | Scope + time budget + deliverable |
| "Ideas for my business" | "I run a home bakery in Riyadh, ~40 orders/month via WhatsApp, mostly custom cakes. Give me 5 growth experiments I can run for under 200 SAR each, ranked by expected impact, with the first step for each." | Real numbers + constraint + ranked format |
The 30-second brief (memorize this shape)
SITUATION: [who you are + relevant facts and numbers] GOAL: [what outcome you actually want] TASK: [the specific thing to produce] FORMAT: [table / list / word count / tone] AVOID: [anything it must not do]
Files, Images & PDFs — Claude Can Read Almost Anything
The 📎 button is where casual users become power users. Claude reads PDFs, Word docs, Excel/CSV, PowerPoint, images/screenshots, code files — and reasons over them, not just skims.
Ten upload workflows worth stealing
| Upload | Ask |
|---|---|
| A contract PDF | "Summarize in plain language. Flag every clause that creates risk or cost for ME, quote each exactly." |
| A confusing utility bill (photo) | "Explain every line. What changed vs a normal month?" |
| Bank statement CSV | "Categorize spending, top 5 categories, 3 realistic cuts." (deeper analysis: p24) |
| A screenshot of an error | "What is this error and what are the 3 most likely fixes, in order?" |
| Your CV + a job posting | "Score the match /10, list gaps, rewrite my top 5 bullets to target this posting — truthfully." |
| Medicine label (photo) | "Explain usage, common side effects, interactions to ask a pharmacist about." ⚠️ verify with a professional |
| A 60-page report | "Executive summary in 10 bullets, then answer only from the document, citing page numbers." (technique: p35) |
| Whiteboard photo | "Transcribe this into clean notes with action items and owners." |
| A child's homework sheet | "Don't solve it. Create 3 similar practice problems and coach me through the first." |
| Handwritten Urdu/Arabic note | "Transcribe and translate; keep the original layout." |
What to know before you rely on it
- Vision is genuinely strong — charts, handwriting, UI screenshots, photos of physical things.
- Size limits exist ⚠️ (roughly: files up to ~30MB, big-but-not-unlimited page counts — split monsters into parts).
- Numbers from images: ask it to re-state the extracted figures so you can eyeball-verify before decisions (p14 habit).
- Multiple files at once is fine — "compare these three quotes" is a killer use.
Artifacts — Documents, Websites & Mini-Apps Inside the Chat
Ask Claude to make something substantial and it opens a side panel with the finished piece — a live webpage, a working tool, a formatted document. That panel is an Artifact: editable by conversation ("make the header blue"), versioned, downloadable, and shareable ⚠️ (publishing options vary by plan).
What people actually build as Artifacts
🧮 Working tools
Loan calculators, unit converters, grade trackers, BMI/zakat/currency calculators, timers, habit trackers — real interactive apps, built from a sentence.
📄 Polished documents
One-page proposals, study guides, checklists, certificates, CVs — formatted, printable.
🎮 Games & learning
Quiz games for revision, flashcard decks, memory games for kids, typing trainers.
📊 Visualizations
Charts from your data, timelines, org charts, decision trees, diagrams.
Three prompts to feel the power
Build an interactive family duty roster for 4 people and 6 weekly chores: drag-free simple tap-to-assign, weekly reset button, progress bar per person, bright friendly design, works on a phone.
Create a quiz app from these notes: [paste notes]. 15 questions, one at a time, instant feedback with explanation, score at the end, "retry wrong questions only" mode.
Design a one-page website for my [service] with: bold headline, 3 benefits, 2 testimonials placeholder, WhatsApp contact button, prices section, elegant dark theme. I'll iterate on the copy after I see v1.
Projects — A Workspace That Already Knows Your Context
A Project is a folder of chats sharing two permanent things: custom instructions (how to behave) and knowledge files (what to know). It's the cure for re-explaining yourself — mental model #1 (p16) industrialized.
Anatomy of a great Project
| Component | What to put in it | Example (home bakery) |
|---|---|---|
| Name | One per ongoing area of life | "Bakery Business" |
| Instructions | Who you are, standards, tone, always/never rules | "I run a Riyadh home bakery, 40 orders/mo, prices in SAR. Answers practical and short. Always consider cost. Never suggest paid ads over 200 SAR without ROI math." |
| Knowledge | The 3–10 documents that define the domain | Menu + prices, supplier costs, last 3 months' orders CSV, brand voice sample |
Projects that pay for themselves
- Job Hunt — CV, target-role description, application tracker → every chat drafts tailored letters instantly
- Studies — syllabus, past papers, your weak topics → a tutor that knows your exact course
- Family Admin — school calendars, insurance summaries, travel docs → "when is re-registration due?" answered from YOUR papers
- Each business vertical — the operator's pattern: one Project per business, so strategy chats from a phone are fully briefed (this is Untapped Lever #11, p91)
Write instructions that actually work
WHO I AM: [role, situation, key numbers] THIS PROJECT IS FOR: [scope] ALWAYS: [3-5 standing orders — tone, format, currency, level] NEVER: [2-3 hard lines] DEFAULT FORMAT: [e.g. "tables for comparisons, bullets for plans, ≤300 words unless I ask for depth"]
Memory — When Claude Remembers Across Chats
Beyond Projects, Claude has a memory feature ⚠️ (rollout varies by plan/platform): it can retain durable facts about you across conversations — your role, preferences, ongoing work — and use them automatically.
The three memory channels, and who controls them
| Channel | How it fills | Your control |
|---|---|---|
| App memory | Claude notices durable facts, or you say "remember that…" | View, edit, delete entries in settings; incognito chats bypass it ⚠️ |
| Project knowledge | You place it deliberately (p20) | Total — it's your files |
| CLAUDE.md rule files (Claude Code) | You write rules; Claude re-reads every session | Total — plain text on your disk (p49) |
Use it deliberately, not passively
Remember these about me: I'm a [role] in [city]. I prefer answers in tables. My currency is SAR. My English level is professional but keep explanations simple. I'm working toward [goal] this year.
- Audit monthly: "What do you remember about me? List everything." Delete what's stale — stale memory is worse than no memory.
- Correct on the spot: "That's outdated — I changed jobs; update your memory."
- Sensitive topics: use an incognito/no-memory chat ⚠️ when you don't want retention.
Styles — Make Claude Write Like You (or Like Anyone)
Styles let you set HOW Claude writes, independently of what it writes: preset modes (concise, explanatory, formal) plus custom styles trained on your own writing samples ⚠️ (feature placement varies by app version).
When each style earns its keep
| Style | Use for |
|---|---|
| Concise | Busy days, status updates, quick answers — cuts the padding |
| Explanatory | Learning new territory — reasons and background included |
| Formal | Official letters, client documents, applications |
| Custom (your voice) | Anything sent under your name — emails, posts, reports |
Build a custom style properly
Web Search — Current Information With Sources
Claude's training has a cutoff date (p7), but with web search it fetches today's information and — critically — cites sources you can click. This turns it from "knowledgeable colleague" into "research assistant with receipts."
Asking for research like a professional
Search for current information: [question]. Requirements: cite each source with a link, prefer official/primary sources, note the date of each source, and clearly separate confirmed facts from claims found on only one site.
Research [product/service A] vs [B] as of this month: price, key limits, user complaints from forums/reviews (not vendor pages). Table format, sources linked per row, end with a recommendation for my case: [your case].
Before I pay [vendor], search independent reviews only — Reddit, Trustpilot, forums, not their own site. Quote at least 3 real user verdicts with links, list red flags, and tell me what to demand as a sample BEFORE paying.
Honest limits
- Paywalled and login-only content is out of reach — it reads the open web.
- Search brings sources, not truth — a cited wrong claim is still wrong. Two independent sources for anything that costs money.
- For deep monitoring (price watching, news tracking) → automation on p59/p87.
Data Analysis — Turn Any Spreadsheet Into Answers
Upload a CSV or Excel file and Claude can run real code on it (the analysis tool executes JavaScript/Python-style computation rather than "eyeballing") — meaning actual sums, joins, pivots and charts instead of guessed arithmetic.
The analysis ladder — same file, increasing power
Real-life files worth feeding it
- Bank/card exports → spending truth, subscription hunting ("find recurring charges I might have forgotten")
- Small-business sales → best products, dead stock, seasonality
- Grades/attendance → trend spotting before report cards do
- Any export from any system — if it opens in Excel, Claude can interrogate it
Voice & Mobile — The Most Underrated Surface
The phone app turns dead minutes — commutes, queues, kitchen time — into finished work. And dictation changes who can use AI at all: typing skill stops mattering.
The dictation superpower (this book's origin story)
Mobile workflows that stick
| Moment | Move |
|---|---|
| Walking out of a meeting | Dictate everything you remember → "structure into decisions, actions with owners, and open questions" |
| In a shop | Photo of two products → "compare honestly for my need: [need]" |
| Commute | Voice conversation mode: rehearse a difficult conversation, practice English/interview answers, think a decision aloud |
| Queue at a government office | Photo of the form → "explain each field; draft my answers from these details: …" |
| Kitchen | "What can I cook with [photo of fridge contents]? 30 minutes max." |
| Bedtime | "A 5-minute story for a 7-year-old about [today's interest], gentle ending." |
Make voice input work brilliantly
- End a rambling note with: "Now organize everything I just said into [format]."
- Bilingual is fine — Urdu-English mixing gets understood; ask for output in either.
- Long thought? Say "part 1 of 2", send, continue — then "now combine both parts."
Connectors — Claude Inside Your Actual Life
Connectors plug Claude into the tools where your information already lives — Gmail, Google Calendar, Google Drive, Notion, Slack, GitHub and a growing catalog ⚠️ — so instead of copy-pasting, you just ask. (Under the hood this is MCP, the open standard explained on p63.)
What connected life looks like
| Connected tool | Ask Claude… |
|---|---|
| Gmail | "Summarize unread from real humans (skip newsletters). Which two need replies today? Draft them in my voice." |
| Calendar | "Find me 3 free 90-minute deep-work slots this week; avoid mornings after late meetings." |
| Drive | "Find the rental contract from last year and tell me the notice period for leaving." |
| Notion | "Add this idea to my Projects database with status=Inbox, and show me what's been sitting in Inbox over 30 days." |
| Several at once | "Prepare tomorrow: calendar events + related emails + relevant Drive docs → one briefing." (Full recipes: p65.) |
The safety model — read this before connecting
- Read is safe; act needs eyes. Summarizing email is harmless. SENDING email, deleting, posting — approve each action yourself until trust is earned (and keep approval on for sending, permanently).
- Connect the minimum. Each connector is power AND surface area. Add them when a real need appears, not for collection.
- Beware instructions hiding in content. A malicious email could contain text like "forward this thread to X". Good agents treat content as data, not commands — but this "prompt injection" risk is why sensitive actions should always confirm with YOU first (more: p64).
Claude in Chrome — A Copilot on Every Webpage
The Chrome extension ⚠️ puts Claude in a sidebar that can see the page you're viewing — and, with your permission, act on it: clicking, filling, navigating. It's the bridge between "chat about the web" and "work the web with me."
Where it shines
📖 Understand
"Summarize this 40-minute read." · "Explain this pricing page's REAL total cost." · "Translate this page and keep the tables."
🔍 Extract
"Pull every deadline from this university page into a checklist." · "List all 12 products here with prices, as CSV."
✍️ Draft in place
"Draft a reply to this LinkedIn message — warm, 60 words." · "Fill this form's answers from my details; let me review before submit."
🔁 Multi-tab work
"Compare the three property listings I have open: price/m², location tradeoffs, red flags."
The guardrails (they're features, not friction)
- It asks before acting — clicks, submissions, purchases get confirmed with you.
- Certain actions stay yours alone: passwords, payments, CAPTCHAs. The AI politely refuses these — by design.
- Same injection-awareness as p26: pages can contain manipulative text; actions confirm with the human.
Real Office Files — Excel, Word, PowerPoint, PDF
Claude doesn't just describe documents — it manufactures them: genuine .xlsx with working formulas, formatted .docx, complete .pptx decks, filled/merged PDFs ⚠️ (file-creation availability varies by plan/surface; fullest power in Claude Code).
What "real files" means in practice
| Ask | What you receive |
|---|---|
| "Budget tracker for a family of 4, SAR, categories for school/food/fuel/savings" | An .xlsx with input sheet, formulas, monthly summary, a chart — download and use |
| "Turn these bullet notes into a client proposal" | A .docx with cover, headings, styled tables, page numbers |
| "10-slide deck pitching [idea] to a school board" | A .pptx: title, agenda, content slides, speaker notes |
| "Fill this application PDF from my details" | The completed PDF, ready to submit (review it!) |
| "Merge these 3 PDFs and add page numbers" | Done. |
Getting professional results
- Give the skeleton: "sections: A, B, C" beats "make a report" — structure in, structure out.
- Iterate the file: upload what it made, say what to change; round-trips are cheap.
- Formulas over values: ask for live formulas ("=SUM…") so the sheet keeps working when you edit.
- Brand it once: give your logo/colors/font preferences in the Project (p20); every future document inherits them.
Sharing Your Work — Chats, Artifacts, Teams
Work locked in your account is half-finished. The sharing layer, briefly and honestly ⚠️ (options shift by plan):
The sharing menu
| What | How | Notes |
|---|---|---|
| A conversation | Share link to the chat | Snapshot for showing reasoning; check nothing private is inside before sharing |
| An Artifact | Publish/share the artifact | The calculator/page itself, usable by the recipient |
| Content | Copy as Markdown / download files | Into Word, Notion, email — formatting survives |
| A Project | Team plans: shared Projects | Whole-team shared context — instructions + knowledge once, everyone benefits |
| Your data | Settings → export | Periodic export = your archive, your property |
Collaboration patterns that work
- The family template: one person crafts the prompt/Artifact (the travel planner, the study quiz), shares it; others reuse without learning anything.
- The team playbook: shared Project holding SOPs + brand voice + price list → every teammate's chats answer consistently. (The operator's version: live playbooks his sales team reads from a deployed page — Lever #24, p104.)
- The review loop: draft in Claude → share chat link with a colleague → they read the reasoning, not just the output → trust builds fast.
The Life Catalog — 20 Everyday Uses, One Line Each
Part 2's graduation page: the breadth, compressed. Every row is a prompt you can fire today; each teaches a reusable shape.
| # | Life area | The ask |
|---|---|---|
| 1 | Meals | "7-day dinner plan, Pakistani-friendly, 4 people, under 60 SAR/day, 30-min cooking — with one shopping list." |
| 2 | Travel | "3-day Istanbul plan with kids: mornings light, one big sight/day, halal food stops, metro-only, budget table." |
| 3 | Study | "Turn this chapter [photo] into 20 flashcards, then quiz me until I score 90%." |
| 4 | Career | "Rewrite these 5 CV bullets with real metrics; ask me for the numbers you need." |
| 5 | Contracts | "Plain-language summary + every clause that costs me money, quoted exactly." (p18) |
| 6 | Kids' learning | "Explain why the sky is blue to a 7-year-old with one kitchen experiment." |
| 7 | Gifts | "10 gift ideas: mother-in-law, loves gardening & tea, budget 150-300 SAR, ranked by wow-per-riyal." |
| 8 | Home repair | "[Photo] What's this pipe issue, is it DIY-safe, what do I tell the plumber so I don't get overcharged?" |
| 9 | Paperwork | "Checklist of documents for [visa/school admission/license renewal] in [country], with common rejection reasons." ⚠️ verify official |
| 10 | Language | "Daily 10-minute English practice as roleplay: you're a customer, I'm a shop owner; correct me gently after each reply." |
| 11 | Homework help | "Don't give the answer. Ask guiding questions until my child reaches it." (the anti-cheating pattern) |
| 12 | Fitness | "Beginner 3-day/week home plan, no equipment, bad knees — table with progressions." ⚠️ doctor for conditions |
| 13 | Difficult emails | "Firm-but-kind reply declining this request; keep the relationship." |
| 14 | Negotiation | "I'm negotiating [car price / salary]. Give me an opening, 3 counters to expect, and my responses." |
| 15 | Meetings | "[Paste messy notes] → decisions, actions with owners and dates, open questions." |
| 16 | Reading | "Summarize this book's core argument in 10 points, then the 3 strongest criticisms of it." |
| 17 | Budget | "Interview me about my finances (5 questions), then build a monthly budget with a savings target." |
| 18 | Events | "Plan a 30-guest aqiqah: timeline, checklist, budget tiers, day-of schedule." |
| 19 | Complaints | "Airline lost my bag; write the complaint citing the compensation policy — ask me for flight details." |
| 20 | Interviews | "You're the interviewer for [job posting]. Grill me one question at a time; score each answer /10 with a better version." |
The Prompt Formula — R·C·T·F·C
Page 5 introduced the anatomy; now we make it muscle memory. Five components — Role, Context, Task, Format, Constraints — and the judgment of when each earns its place.
The formula, weaponized
| Component | Its real job | When to skip it |
|---|---|---|
| Role | Selects the expertise & taste to answer with ("senior contracts lawyer" pulls different patterns than "helpful assistant") | Casual questions |
| Context | Kills the #1 failure: generic answers. Numbers, history, stakes | Never — context is the last thing to cut |
| Task | ONE clear deliverable per ask | — |
| Format | Makes output land ready-to-use (table, word count, tone) | When you want its natural judgment |
| Constraints | Fences: budget, "don't invent facts", "ask before assuming" | Low-stakes asks |
One scenario, formula applied
ROLE: You are a pricing consultant for small service businesses. CONTEXT: Home bakery, Riyadh. 40 orders/mo average 180 SAR. Ingredients ~35%, my time unpriced. Two competitors charge 220-260 SAR for similar quality. Demand exceeds my capacity most weeks. TASK: Recommend a new pricing structure. FORMAT: 3 options in a table (price, expected demand effect, monthly profit math shown), then your single recommendation in 3 sentences. CONSTRAINTS: Keep me under 50 orders/mo. Assume 10% customer loss per 15% price rise. If you need a number I didn't give, ASK instead of assuming.
The two mini-forms for daily speed
⚡ Quick form (80% of asks)
[Context]. [Task]. [Format].
"I manage 6 engineers, one is chronically late on reports. Draft a supportive-but-clear message. Under 100 words."
🎯 The last-line trick
End any prompt with: "Ask me up to 3 clarifying questions before answering if it would improve the result." — converts monologue into consultation.
Few-Shot Prompting — Show, Don't Tell
Describing a style takes paragraphs and still misses. Showing 2–3 examples transfers it instantly — the single most reliable technique for consistent output. "Few-shot" = "here are a few shots of what I want."
Where examples beat descriptions every time
| Goal | Instead of… | Do |
|---|---|---|
| Tone cloning | "friendly but professional" | Paste 2 emails you actually sent: "match this voice" |
| Data extraction | Explaining the output fields | Show 2 input→output pairs, then "now do these 20" |
| Naming/branding | "catchy but classy names" | "Names I like: X, Y, Z. Names I hate: A, B. Generate 10 more like the first group" |
| Report format | A format spec | Last month's report: "same structure, this month's data" |
The template
I need [task]. Follow the pattern of these examples exactly. EXAMPLE 1 Input: [sample input] Output: [exactly what you want] EXAMPLE 2 Input: [different sample] Output: [matching style] Now process: Input: [real input]
Craft notes from heavy use
- 2–3 examples is the sweet spot — one can be over-copied; five adds little.
- Make examples span the variety — a short one and a long one; an easy case and a weird case. The model learns the range, not just the center.
- Include a counter-example when quality matters: "BAD output (never do this): …" — fencing off failure modes is cheap insurance.
- This is exactly how the cockpit's messages held one voice across a thousand drafts: gold-standard samples in the generator's instructions, not adjectives (p4's rails in action).
Extended Thinking — When to Buy More Reasoning
Claude can answer instantly, or it can think first — an internal reasoning phase (visible as a thinking indicator; toggleable/budgetable ⚠️ by plan and surface) before the reply. Knowing when to demand thinking is a real skill.
When thinking pays vs. wastes
| Worth the thinking time | Skip it — instant is fine |
|---|---|
| Multi-step math & finance (loan comparisons, tax scenarios) | Drafting a birthday message |
| Planning with interacting constraints (schedules, budgets, logistics) | Summaries and rewrites |
| Debugging: code, contracts, arguments | Factual lookups |
| Decisions with tradeoffs ("which offer do I take?") | Formatting jobs |
| Anything where a wrong answer costs real money | Brainstorm quantity |
Triggering deeper reasoning by prompt (works everywhere)
Think through this step by step before answering. Show your working: 1) list what's known, 2) what must be calculated/decided, 3) work through it, 4) sanity-check the result against common sense, 5) THEN give the answer with your confidence level and what would change your mind.
Why this matters mechanically (p7 callback)
An LLM predicts token by token — its "scratchpad" is the text itself. Forcing intermediate steps literally gives it more computation on your problem. "Show your working" isn't teacher nostalgia; it's buying the model more thinking.
Structured Output — Answers Shaped Like Tools
Prose is for reading; structure is for using. The moment output feeds a spreadsheet, a system, a decision meeting — demand shape.
The structure menu
| Shape | Demand it when | The magic words |
|---|---|---|
| Table | Comparing anything | "as a table with columns: …" |
| Checklist | Anything executable | "as a checklist in execution order, with time estimates" |
| Template | Recurring documents | "as a reusable template with [PLACEHOLDERS]" |
| JSON | Feeding software (p67) | "return ONLY valid JSON matching this schema: {…} — no commentary" |
| Scored list | Prioritizing | "ranked, with a 1-10 score and one-line justification each" |
JSON for automation — the pattern that runs pipelines
Extract from the text below. Return ONLY valid JSON, no other words:
{
"person": "",
"company": "",
"request_type": "quote | complaint | question | other",
"urgency": 1-5,
"summary_one_line": "",
"suggested_reply": ""
}
TEXT: [paste customer message]This exact shape — message in, labeled JSON out — is the atom of a thousand business automations (triage, routing, CRM enrichment). You'll meet it again in the API (p67) and the income paths (p122, p126).
Pro reliability tricks
- "If a field is unknown, use null — never guess." This one line kills most hallucinated fields (p14 applied to structure).
- Give one filled example alongside the schema (p32's few-shot) — format compliance jumps.
- For long outputs, number everything — "revise #7" beats re-pasting paragraphs.
Long Documents — Getting Truth Out of 100+ Pages
Claude's huge context (p8) reads entire contracts, theses, annual reports. But "it can hold it" ≠ "you'll get the best out of it" — technique matters at length.
The five laws of long-document work
The contract-reading prompt (frame for any serious document)
[UPLOAD/PASTE DOCUMENT FIRST] You are a meticulous analyst. Using ONLY this document: 1. 10-line plain-language summary. 2. Every obligation on MY side: quote + location. 3. Every deadline, fee, penalty and auto-renewal: quote + location. 4. The 5 riskiest clauses for me, ranked, each with why. 5. Questions I should ask before signing. Rule: if something isn't stated, write NOT FOUND — never infer silently.
Decomposition — Big Goals Become Prompt Chains
"Plan my business" gets platitudes. Professionals never prompt the mountain — they prompt the next slope, and chain. The skill: goal → milestones → this week → today's prompts.
Worked example: "launch my home bakery's Instagram in 7 days"
Why chains beat mega-prompts
| Mega-prompt ("do everything") | Chain |
|---|---|
| Shallow on all parts | Deep on each part |
| One bad assumption poisons all | You steer between links |
| Output too long to use | Each output sized to action |
| Context bloats (p8) | Fresh chat per link, plan carried forward |
The Critique Loop — Draft → Attack → Revise
The highest-ROI pattern in all of prompting (mental model #3, p16): never accept draft one. Make the model its own harshest reviewer — quality jumps 2 levels in 2 extra messages.
The three-message loop
Rubric scoring — for anything with a quality bar
Score this [essay/proposal/email] against these criteria, 1-10 each, with one sentence of evidence per score: 1. Clarity for a busy reader 2. Persuasiveness of the core argument 3. Concreteness (numbers & specifics vs vague claims) 4. Tone fit for [audience] 5. Zero fluff Then revise to bring every score to 9+, and re-score honestly.
Persona attacks — choose your reviewer
| Reviewing… | Summon… |
|---|---|
| A business plan | "a skeptical investor who has seen 1,000 pitches" |
| A contract you drafted | "opposing counsel looking for loopholes to exploit" |
| Marketing copy | "a tired customer scrolling at midnight — why do they NOT buy?" |
| Code / a system | "an adversarial tester trying to break it with evil inputs" |
| An exam essay | "the strictest examiner on the board, marking to the rubric" |
Meta-Prompting — Make Claude Write Its Own Instructions
The elegant secret: the best prompt engineer available is the model itself. When a task is important or recurring, don't write the prompt — commission it.
The commissioning pattern
I need a reusable, excellent prompt for this recurring task: [describe task and what "great output" means to you]. First ask me up to 5 questions to capture what matters. THEN write the ideal prompt with placeholders in [BRACKETS], including role, context slots, format spec, and quality constraints. Finally show one filled-in example run.
The upgrade pattern — fix a prompt that underperforms
Here's a prompt I use, and an example of the mediocre output it produces: PROMPT: [paste] OUTPUT: [paste] Diagnose why the output is weak (missing context? vague task? no format?). Rewrite the prompt to fix each diagnosis. Explain each change in one line.
Build your prompt library (the compounding asset)
- Keep one note — "PROMPT LIBRARY" — organized by life area. Every prompt that produces a great result gets saved with a one-line "use when".
- Commission prompts for your top 5 recurring tasks this week: the weekly report, the client reply, the lesson plan, the status update, the content post.
- Graduation path: library note → Project instructions (p20) → slash commands (p50) → skills (p51). Same asset, increasing automation — by the end of Part 4 your library RUNS itself.
Prompt Library I — Work & Business
Fifteen field-tested shapes. Replace [BRACKETS], fire, iterate. Each is deliberately compressed — add YOUR context (p31) for full power.
From these raw notes [paste], write my weekly status: ✅ done, 🔄 in progress (with % and blocker), ⏭ next week, 🚨 needs decision. Executive tone, scannable in 60 seconds.
Draft an email that [delivers bad news / declines / chases payment] to [relationship]. Goal: [outcome] while keeping the relationship. Firm, warm, under 120 words, ends with a clear next step.
Build a [30]-min agenda for [meeting purpose], attendees [roles]. Timeboxed items, each with an owner and a decision-or-output, plus a parking lot rule.
Write a JD for [role] at [company type]. Sections: mission of the role, 5
outcomes for year 1 (measurable), must-haves (max 5), nice-to-haves, what we
offer. No clichés ("rockstar", "fast-paced").For hiring a [role], create 8 interview questions mapped to the 4 skills that matter most, each with what a 1/5 vs 5/5 answer sounds like. Table.
SWOT for [business] given these facts: [numbers, market, competitors]. Rule: every entry must be specific and evidenced from my facts — no generic entries like "competition". End: the ONE strategic move the SWOT points to.
Here are my numbers for [period] vs previous: [paste]. What genuinely changed (ignore noise under 5%)? Rank 3 causes to investigate, each with the exact question to answer and where I'd find the data.
I'll describe how I do [process] messily: [dictate]. Turn it into a numbered SOP a new hire could follow alone: steps, tools, common mistakes, quality check at the end.
Client proposal for [service] to [client type]. Sections: their problem in their words, outcome promised, approach (3 phases), timeline, investment (anchor high option first), guarantee, next step. Confident, zero begging.
My service: [what]. Cost structure: [numbers]. Competitors charge [range]. Model 3 price points: monthly profit at realistic volumes (show math), positioning message for each. Recommend one with reasoning.
Search [competitor]'s public presence. Table: offer, pricing, positioning claim, visible weaknesses (slow site? bad reviews? gaps?). End: 3 ways I differentiate honestly. Sources linked.
Turn my goal "[goal]" into 1 Objective + 3 Key Results that are measurable weekly, with a leading indicator each (something I control), and the first Monday action.
For project [X], list top 8 risks: likelihood × impact (H/M/L), early warning sign, mitigation owner, plan-B. Sort by exposure. Table.
Write an escalation to [senior person]: issue [X], impact [numbers], what I already tried, the specific decision/help needed by [date]. 150 words max, subject line included, zero drama.
One-page update on [area]: headline (one sentence, the truth), 3 numbers with trend arrows, what's working, what's not + fix, the ask. They read it in 90 seconds — write for that.
Prompt Library II — Study & Family
Explain [topic] to a smart 12-year-old: one everyday analogy, 5 short points, one "wow" fact, then 3 questions to check I got it.
From this material [paste/photo], make 20 flashcards (Q on one line, A below). Mix: 10 recall, 6 understanding, 4 application. Hardest last.
Quiz me on [topic], one question at a time. After each answer: correct/not, short explanation, difficulty adapts to my performance. Track my score. 10 questions, then a weakness report with what to revise.
Essay question: [paste]. Build an outline: thesis options (pick strongest), 3 arguments with evidence types needed, counter-argument + rebuttal, conclusion angle. I write the essay myself — you coach.
Exams: [subjects + dates]. Available: [hours/week, when]. Build a realistic timetable: hardest subjects at my best hours, spaced repetition (revisit each topic 3×), one rest day, printable table.
My child (age [X]) is stuck on: [problem]. DO NOT solve it. Ask one guiding question at a time, building on their answers, until THEY reach the solution. Then give a similar practice problem.
Age-appropriate science project on [interest] for a [X]-year-old: hypothesis in kid language, materials under [budget], safe steps, what to measure, how to present it on one poster.
Daily 10-min [English] practice. Roleplay: you're [a customer/interviewer], I'm [role]. Speak simply, correct my mistakes AFTER each reply (max 2 fixes), and teach one useful phrase per session. Start now.
Translate this to [Urdu/English]: [text]. Then teach me: 3 vocabulary items worth keeping, 1 grammar pattern used, and the tone difference between the languages here.
Interview me (5 questions, one at a time) about income, fixed costs, debts, goals. Then: monthly budget table, 3 realistic savings moves with amounts, and one habit to start this week. Currency: [SAR].
7 dinners for [family size]: cuisine [X], budget [Y]/day, [dietary rules], under 40 min each. Output: menu table + ONE consolidated shopping list by supermarket section.
[X]-day trip to [place], party: [ages], budget [Z] total, pace: relaxed. Day-by-day table (morning/afternoon/evening), transport between stops, [halal] food options, one backup indoor plan per day, budget breakdown.
Symptoms/situation: [describe]. Prepare me for the appointment: questions to ask the doctor, information I should bring, terms I might hear explained simply. You are not diagnosing — you're preparing me to get the most from a professional.
[Photo of form] Explain each field in simple words. Draft my answers from these details: [details]. Flag any field where a wrong answer commonly causes rejection.
Write a warm 200-word family WhatsApp update from these events: [list]. Light humor, one photo caption suggestion, ends with a question that gets the relatives replying.
Prompt Library III — Builders & Creators
These preview Part 4's world — you can fire them in normal chat today, and in Claude Code they become real running systems.
Landing copy for [offer] to [audience]: headline (outcome, ≤9 words), sub, 3 pain→benefit blocks, social-proof placeholder, FAQ (5), CTA. Write at a 9th-grade reading level; zero hype adjectives.
Build a complete single-file HTML page for [purpose]: dark elegant theme, mobile-first, sections [list], smooth scroll nav, contact via WhatsApp link. Inline CSS/JS, no external libraries. Production quality.
My table: [name(columns…)]. Write SQL to answer: "[question]". Explain each clause in one line, then give a second query that double-checks the result a different way.
Regex to match [pattern description]. Provide: the pattern, plain-English breakdown, 5 strings that match, 5 near-misses that must NOT match.
Python script (standard library only): [task, e.g. "rename all photos in a folder to YYYY-MM-DD_number from their EXIF date"]. Include: dry-run mode that only prints, clear comments, and graceful handling of bad files.
Here's messy CSV data: [paste sample]. Write the cleaning rules you'd apply (dates, casing, duplicates, impossible values), apply them, output clean CSV + a log of every change count.
Show a minimal working example calling [API] in Python: auth from an environment variable (never hardcoded), one request, error handling, printed result. Then the same as a reusable function.
Write a README for this project: [describe or paste code]. Sections: what it does in one sentence, quick start (3 commands), configuration table, common problems. Assume a smart beginner.
For this function [paste], list test cases as a table: normal cases, edge cases (empty/huge/wrong-type), and evil cases. Then write the tests in [pytest].
Refactor this code for readability WITHOUT changing behavior: [paste]. Show before/after of each change with a one-line reason. Flag anything that looks like a real bug separately — don't silently fix logic.
Bug: expected [X], got [Y]. Code: [paste]. Error: [paste exact message]. Diagnose: 3 candidate causes ranked, the test that distinguishes them, then the minimal fix for the winner. No rewrites beyond the fix.
I want [task] to run automatically every [schedule] on Windows. Give me the script + the exact Task Scheduler setup steps + how to verify it ran + where logs go. Quote the path correctly (spaces!).
I want data from [site/type]. First: is scraping it appropriate (terms, robots, personal data)? If yes, outline a respectful approach (rate limits, public data only, caching). If no, name the legitimate alternative (API, export, dataset).
Spec a one-page dashboard for [audience] tracking [goal]: the 5 metrics that matter (and 5 tempting-but-useless ones to exclude), layout sketch in text, update frequency, and the single number that leads the page.
Idea: [describe]. Attack it: who EXACTLY pays, what they use today, why they'd switch, hardest part to build, cheapest 1-week test that could kill or confirm it. Verdict: build / test / drop — with reasoning.
The 12 Mistakes — A Diagnostic Checklist
When output disappoints, the cause is on this list. Diagnose, fix, refire. (Print this page for your desk — the Print button handles colors.)
| # | Mistake | The fix |
|---|---|---|
| 1 | Keyword-typing ("marketing ideas restaurant") | Full-sentence brief with context (p17, p31) |
| 2 | Zero context — no numbers, history, stakes | Add the facts you'd give a human consultant |
| 3 | No format spec — then hating the wall of text | Say the shape: table/checklist/word-count (p34) |
| 4 | One-shot perfectionism — judging on draft one | Iterate; run the critique loop (p37) |
| 5 | Mega-prompt — five tasks in one ask | Decompose and chain (p36) |
| 6 | Describing style in adjectives | Show 2 examples instead (p32) |
| 7 | Trusting numbers/citations blindly | Labels + one cheap verification (p14) |
| 8 | Letting it assume silently | "Ask me up to 3 questions first" (p31) |
| 9 | Marathon chats — one chat for everything all week | One chat one task; summary hand-offs (p8) |
| 10 | Wrong surface — begging chat to do file/system work | That's Claude Code's job (Part 4; chooser p70) |
| 11 | Pasting secrets | Never — the hard list (p15) |
| 12 | Losing winning prompts — re-inventing weekly | Library note → Projects → commands (p38) |
Claude Code — The Same Brain, Given Hands
Everything until now was conversation: you carry insights out of the chat and apply them yourself. Claude Code deletes that last step. It's Claude running on your computer with permission to act: create and edit files, run programs, query databases, use git, browse, deploy. You describe outcomes; it does the work in place.
The difference, felt
| In chat you say… | …and then YOU do | In Claude Code you say… | …and IT does |
|---|---|---|---|
| "Write me a website" | Copy code, make files, figure out hosting | "Build the site in this folder and deploy it" | Creates files, tests, deploys, hands you the URL |
| "How do I organize 2,000 photos?" | Follow instructions manually for an hour | "Organize my Pictures folder by year/month" | Writes the script, shows a dry run, executes |
| "Draft outreach messages" | Paste one by one into a tracker | "Generate all 90, save to the database, rebuild the dashboard" | The cockpit pattern — end to end |
Where it runs
Terminal (CLI) Desktop app VS Code / JetBrains panel claude.ai/code (cloud, from any browser) GitHub (tag it on issues/PRs) — same agent, pick your comfort. Non-coders often prefer the desktop app; the terminal is the classic.
The honest frame
Non-Coders Welcome — The Honest Map
The most important sentence in Part 4: the person running the systems in Part 6 does not write code. A 15-year planning engineer, not a programmer — directing AI in plain English (and voice notes). Here's the honest map of what that means for you.
What a non-coder can genuinely build (week one)
🌐 Websites
Personal pages, business one-pagers, portfolios, guides like this one — built, previewed, deployed to a real URL.
🗂 Organizers
"Sort my Downloads folder", "rename these 500 scans by date", "merge these PDFs monthly."
📊 Trackers
Expense logs, study trackers, client lists — real databases with dashboards (build-along p58).
🤖 Little robots
Scripts that run on schedule: daily report emails, folder backups, price checks (p59).
What still needs YOU (and always will)
- Judgment: what's worth building, what "good" looks like, when it's wrong.
- Verification: click the thing, check one number, read the output (mental model #5). The operator's laws exist because unverified claims burned him (p76, p79).
- Small-steps discipline: ask for one change, check it, ask for the next. Non-coders who fail usually asked for cathedrals in one prompt (p36's chains apply doubly here).
The vocabulary you need (all of it)
| Word | Meaning, honestly |
|---|---|
| Folder / directory | Where your project lives. Claude Code works inside the folder you start it in. |
| Terminal | A text window where you type commands. You need ~3 commands ever; Claude handles the rest. |
| Script | A file of instructions the computer runs. You'll ask for them, not write them. |
| Deploy | Put files on the internet. One command with free hosting (p108). |
| Repo / git | A time machine for your folder — every version saved, any mistake reversible (p56). |
Install & First Launch — 15 Minutes, Once
Three routes; pick one. ⚠️ Exact buttons/versions drift — if anything looks different, ask Claude in the chat app: "walk me through installing Claude Code on Windows, current steps."
Route A — Desktop app (easiest, recommended for non-coders)
Route B — Terminal (the classic)
npm install -g @anthropic-ai/claude-code
mkdir my-first-project cd my-first-project claude
Route C — Inside VS Code / your IDE
Install the Claude Code extension from the marketplace → sign in → it appears as a panel beside your files. Best when you'll also look at code; also the base for the browser-preview workflow.
Sanity checklist before p46
- ✅ It greets you with a prompt box inside your chosen folder
- ✅ Typing "what files are in this folder?" gets a real answer (empty is fine!)
- ✅ You know one fact: it can only touch the folder you started it in (plus what you explicitly allow) — your computer isn't open-season
Your First Session — A Real Walkthrough
Follow this literally; it teaches the rhythm: small ask → watch → verify → next. Total time ~20 minutes.
Create a file called hello.html: an elegant dark page with my name "[NAME]", a live clock, and a daily changing quote. Then open it in my browser.
Look at my Downloads folder and TELL me (don't do anything yet): how many files, grouped by type and age. Propose an organization plan.
Create NOTES.md summarizing: what we built, decisions made, and the 3 next ideas. Keep it updated at the end of every future session.
Reading the transcript (what the scrolling text means)
| You'll see | It means |
|---|---|
Read / Write / Edit … | File operations, named explicitly |
Bash: … | A command it wants to run — read these before approving early on |
| A permission prompt | Your checkpoint — next page makes you fluent |
Under the Hood — The Toolbelt
Claude Code = the model (p10) + a toolbelt + your permission system. Knowing the tools by name turns the transcript from noise into a readable work log — and lets you direct precisely.
The core toolbelt
| Tool | What it does | Watch for |
|---|---|---|
| Read | Opens files (code, docs, images, PDFs) | Harmless; it reads before it writes — good sign |
| Glob / Grep | Finds files by name / searches contents like a detective | How it "knows your codebase" — pattern-hunting, not magic |
| Write / Edit | Creates files / makes surgical text replacements | Edits show exact before→after diffs — glance at them |
| Bash / Shell | Runs commands: installs, scripts, git, deploys | The power tool. Early days: read every command before approving |
| WebFetch / WebSearch | Reads pages / searches the live web | Powers "check the current docs" (p14's cure) |
| Task / agents | Spawns sub-Claudes for parallel work | Advanced leverage — p54 |
| MCP tools | Everything plugged in via MCP: databases, browsers, email… | Named like mcp__server__action — p53 |
Directing by tool name (subtle, powerful)
- "Grep for every file that mentions the old price before we change anything" — forces evidence-gathering first.
- "Read the whole file before editing it" — kills blind-edit mistakes.
- "Use the browser preview and verify the button actually works" — demands proof, not claims (p52's hooks can even enforce this).
Permission Modes — Your Safety Dial
Every action requests permission by default. The modes are a dial between control and flow ⚠️ (naming varies slightly by version — the four positions are stable):
The four positions
| Mode | Behavior | Use when |
|---|---|---|
| Default (ask) | Prompts for edits & commands; remembers your "always allow" choices | Learning; unfamiliar projects |
| Accept-edits | File edits auto-approved; commands still ask | Trusted flow on your own project — the daily sweet spot |
| Plan mode | Read-only research → written plan → your approval → then execution | Big/risky changes; the professional habit (p55) |
| Bypass / "yolo" | No prompts at all | Disposable experiments in isolated folders. NOT your real files. Honestly: rarely. |
Granular allow-lists (the smarter path)
Instead of a blunt dial, per-action rules live in a settings file: "always allow git status, ls, reads; always ask for rm, pushes, installs." You approve a safe command once with "don't ask again for this" — friction falls all week while dangerous verbs still knock. (Curated allow-list recipe: Lever #18, p98.)
The judgment rules that matter more than the dial
- Reads are free, writes are checked, deletes are sacred. Approve reads casually; glance at edits; read delete/overwrite commands word by word.
- Backups make courage cheap: with git (p56), the worst approved mistake is one
git checkoutfrom undone. No git = no bypass, ever. - Untrusted content lowers the dial: processing files/pages from the internet? Injection risk (p26/p64) means keep prompts ON for actions.
CLAUDE.md — Teach It Once, Benefit Forever
The single highest-leverage file in this entire book. CLAUDE.md is a plain text file in your project that Claude Code reads automatically at every session start. Whatever it says is simply… known. Preferences, facts, rules — delivered fresh to every future session, forever.
What goes in (with a real skeleton)
# Project: [name] ## What this is [2 lines: purpose, who uses it] ## Facts Claude must know - Main files: [x.py does A; site/ holds B] - The database is [file]; NEVER delete rows — mark status instead - Prices/data come from [source of truth] ## How I like to work - Explain in plain English; I'm not a programmer - Small steps; after each change, verify and show me proof - All file paths in quotes (Windows spaces!) ## Hard rules - Never touch anything outside this folder - Never print secrets or .env contents - Before any risky change: backup first, then proceed - If tests/checks fail: STOP and show me, don't improvise
Two levels, one habit
| File | Scope | Typical content |
|---|---|---|
~/.claude/CLAUDE.md (home folder) | YOU — every project | Identity, style, universal rules ("always tables", "never ask permission twice for reads") |
project/CLAUDE.md | This project | Architecture facts, do/don'ts, quirks |
The growth loop — where the magic compounds
Slash Commands — Prompts That Became Buttons
Any prompt you type twice deserves a name. A slash command is a saved prompt file — type /name and the whole thing fires. Your p38 prompt library, industrialized.
Make one in 60 seconds
.claude/commands/report.md.Look at everything that changed in this project in the last 7 days (git log + modified files). Write my weekly status: done / in-progress / blockers / next week. Save as reports/week-[date].md and show me the highlights.
/report any Friday. Done. Commands accept arguments too: /email-draft client-name ⚠️ (syntax: $ARGUMENTS placeholder in the file).A starter command set worth stealing
| Command | Does |
|---|---|
/checkpoint | Update NOTES.md: state, decisions, next steps (p46's habit, one keystroke) |
/verify | Run every check this project has; report red/green with evidence |
/cleanup | Find junk files, dead code, TODOs; propose (not execute) a tidy plan |
/explain | "Explain what [file] does in plain English, non-programmer level" |
/ship | Test → build → deploy → give me the URL + what changed |
/forge (build a complete scored system), /webforge (a full website funnel), /military-grade (adversarial hardening pass), /revenue-dashboard (executive dashboard from a framework list). Each encodes hundreds of lines of hard-won standards — fired in five keystrokes. That's what your library becomes with months of compounding (p16, model #4).Skills — Expertise That Loads Itself
One step above commands: a skill is a folder of instructions (+ optional scripts/templates) that Claude loads when the task calls for it — you don't even have to invoke it. Commands are buttons; skills are employees with training manuals.
Anatomy
--- name: invoice-maker description: Create branded PDF invoices from a client name + line items. Use whenever the user asks for an invoice or billing document. --- # Standards - Template: assets/template.html (logo, colors preset) - Numbering: INV-[YEAR]-[running number from data/counter.txt] - Always: VAT 15%, payment terms 14 days, save to invoices/[client]/ - After creating: open a preview and verify totals sum correctly.
The description is the trigger — Claude reads it and thinks "this task matches; load the manual." Say "invoice Ahmed for two consulting days" → the skill fires, standards applied, no reminder needed.
Skills worth building first
- Your document standard — brand, fonts, print-fix CSS, structure for every guide/report you produce (the operator's version guarantees interactive HTML with copy buttons and progress bars — you're reading its output)
- Your QC gate — "before calling anything done: run checks X, Y; screenshot; verify totals" (p23/p37 institutionalized)
- Your deliverable per trade — lesson-plan skill for teachers, quote-builder for contractors, listing-writer for realtors (income relevance: Part 9)
Meta: the skill that makes skills
There's a skill-creator skill: describe a repeated job, and it interviews you, writes the SKILL.md, and installs it ⚠️. Turning-jobs-into-skills is itself a one-sentence job — Lever #4 (p84) shows the operator's backlog of skills waiting to be minted this way.
Hooks — Rules Enforced by the Machine, Not by Memory
The elite tier of customization. A hook is a command that runs automatically on session events: before/after any tool call, on session start/end. CLAUDE.md asks Claude to behave; hooks make behavior physically automatic — even a forgetful model can't skip them.
What events can trigger what (concept table)
| Event | Example hook | Guarantee created |
|---|---|---|
| Before a command runs | Block any command containing rm -rf or a secrets filename | Certain disasters become impossible, not just discouraged |
| After a file edit | Auto-format the file; run its tests; flag failures instantly | Quality checks nobody has to remember |
| Session start | Print current project status + pending review queue | Every session begins briefed |
| Session end / stop | Run the checkpoint script; update state files | Work is never lost to a crash — automatically |
A concrete taste (configuration sketch ⚠️ syntax evolves — ask Claude to write yours)
Set up hooks for this project: (1) after any edit to a .py file, run its tests and show failures immediately; (2) block any bash command that would delete files outside this project; (3) at session end, update NOTES.md with what changed. Explain each hook file you create.
Why this page matters beyond convenience
MCP Servers — Plug Databases, Browsers & APIs Into Your Agent
Part 2 showed connectors in the chat app (p26). In Claude Code the same standard — MCP (full story p63) — plugs professional tools into your agent: databases it can query, browsers it can drive, platforms it can call.
The greatest-hits servers ⚠️ (ecosystem grows weekly)
| Server | Claude gains | First real use |
|---|---|---|
| SQLite / Postgres | Query your database directly | "How many leads have no email? Show the query." — evidence culture (p76) runs on this |
| Playwright / browser | Drive a real browser: navigate, click, read pages, screenshot | "Open the deployed site, verify all 5 sections render, screenshot proof" |
| Filesystem | Wider file operations beyond the project | Organize archives, bulk-process documents |
| GitHub | Issues, PRs, repos as tools | "Open an issue for every TODO in the codebase" |
| Gmail / Calendar / Notion / Slack… | Work apps as tools | The p26 catalog, now inside your builds |
Adding one (the shape, not the memorization)
Add an MCP server so you can query my SQLite database at data/app.db (read-only). Then prove it works: list the tables and count the rows in each.
Under the hood it's a claude mcp add … command or a JSON config entry ⚠️ — let Claude write it; your job is deciding WHAT gets plugged in.
The two safety lines
Subagents — When One Claude Dispatches Many
Claude Code can spawn subagents: independent Claude workers, each with its own context window and toolbelt, running in parallel on scoped tasks — then reporting back for synthesis. You stop being a user with an assistant and become a director with a staff.
Where parallelism genuinely pays
| Pattern | Dispatch | Payoff |
|---|---|---|
| Research fan-out | 3 agents: one per vendor/option, same checklist | An afternoon of comparison in minutes, no cross-contamination |
| Independent review | Build with one; have 2 fresh agents attack it (p37 at agent scale) | Reviewers with no attachment to the draft find what the author can't |
| Big-codebase exploration | Agents explore different areas, report maps | Understanding without drowning your main context (p8) |
| Bulk with judgment | Split 200 items across agents with the same rubric | Scale + consistency |
Directing a fan-out (the shape)
Dispatch 3 parallel agents to research [decision]: - Agent 1: option A — pricing, limits, 3 independent user reviews with links - Agent 2: option B — same checklist - Agent 3: risks & questions we haven't thought to ask Non-overlapping scopes. Then synthesize: one comparison table + one recommendation with reasoning. I want sources, not vibes.
The economics (honest)
Plan Mode — Research First, Touch Nothing, Then Execute
The professional's answer to "I'm nervous about letting AI change things." In plan mode Claude is physically read-only: it explores, thinks, and writes a plan document — what it will change, how, what could go wrong, how to verify. Nothing happens until you approve. Then it executes the approved plan.
When plan mode is mandatory (adopt these triggers)
- Anything touching data (databases, bulk file moves, renames at scale)
- Changes to a working system — the thing that already earns/serves
- Multi-file builds where a wrong architecture wastes hours
- Anything you'd struggle to describe precisely — planning surfaces the ambiguity BEFORE it becomes wrong files
The rhythm in practice
Git & GitHub — Undo for Your Whole Project
Git is a time machine for a folder: snapshots ("commits") of everything, forever, reversible. GitHub is the cloud copy. Together they make AI-speed building SAFE: any mistake, however large, is one command from undone. Non-coders: this is FOR you, and Claude does the typing.
The five moves (you'll only ever say these in English)
| You say | Git concept | What happens |
|---|---|---|
| "Set up version control here" | git init | Time machine on; nothing changes visibly |
| "Save a snapshot: 'working contact form'" | commit | Restore point created (do this after every win) |
| "What changed since the last snapshot?" | diff/status | Exact review of AI's work — trust-but-verify made easy |
| "Undo everything since the last snapshot" | restore/checkout | The eraser. Fearlessness lives here |
| "Back this up to GitHub, private" | push | Fire/theft/spilled-tea insurance + work-from-anywhere (p60) |
The workflow that changes your relationship with risk
Rule: commit a snapshot with a clear message after every working milestone, and BEFORE any risky change. When I say "undo", show me what we'd lose, then restore the last snapshot on my confirmation.
Branches & PRs — one level up, still plain English
- Branch: "try this redesign on a branch" = a parallel universe; merge if good, delete if not. Main stays safe.
- Pull request (PR): a proposed change wrapped for review — the unit of teamwork on GitHub, and the unit AI-at-work uses too (p60: Claude opens PRs you review from your phone).
Build-Along #1 — A Personal Dashboard, Deployed For Real
Ninety minutes, zero code knowledge, and at the end: a page on the actual internet with your name on it. Every prompt given; the skill being trained is directing + verifying (p46's rhythm at project scale).
claude → Build index.html: my personal dashboard. Dark elegant theme. Sections: live clock + date (Riyadh), my week's focus list (editable, saves in the browser), quick links grid (WhatsApp Web, Gmail, YouTube, [your 3]), a daily rotating quote, weather placeholder. Mobile-perfect. Set up git and commit "v1".
Deploy this to Cloudflare Pages (free). Walk me through the one-time account step if needed, then deploy and give me the live URL. Then explain how future updates get deployed in one command.
https://your-name.pages.dev. Send it to someone you love.Build-Along #2 — A Real Data System (The Cockpit Pattern, Miniature)
The architecture behind the job-hunt cockpit (p73), at learnable scale: database → rules → generated dashboard → deploy. Build it once and you'll recognize this skeleton inside half the systems in the world. Example chosen: a family expense tracker — swap in anything (clients, applications, inventory).
Create expenses.db (SQLite): table expenses(id, date, category, amount_sar, who, note, status). Rule for this whole project: NEVER delete rows — corrections happen by adding a correcting row or changing status. Seed 15 realistic sample rows. Show me the table as proof. Git init + commit.
Build build_dashboard.py: reads expenses.db, generates dashboard.html — this month's total vs last, by-category bar chart, by-person split, last 10 entries table, a red flag if any category is 30% over its 3-month average. Elegant dark theme. The page is REGENERATED from data, never hand-edited. Run it, open it, verify the totals match the database with a query.
Automation — Work That Happens While You Sleep
The final promotion: from "Claude does it when I ask" to "it happens." Three tiers, ascending autonomy:
Tier 1 — Scheduled scripts (no AI at runtime, $0)
The build-along's update.bat + Windows Task Scheduler (or cron on Mac/Linux) = the dashboard rebuilds itself nightly. Claude sets it up in one ask:
Create a scheduled task that runs update.bat daily at 06:00. Path has spaces — quote it properly. Add logging to logs/, tell me how to verify tomorrow that it ran, and what to check if it didn't.
Tier 2 — Scheduled AI (cloud agents)
Claude's scheduled tasks ⚠️ run agent sessions on a schedule: "every morning at 7, read my project state + inbox summary + calendar, produce my day brief." AI judgment, timer-fired (the operator's unused goldmine — Lever #7, p87).
Tier 3 — Watchers & loops
Long-running guards: watch a folder ("new CV arrives → score it → file it"), watch a site ("price drops → alert me"), self-monitoring systems ("if the SLA check fails → rebuild → notify"). His net-watchdog and send-monitors live here. Rule of the tier: every loop needs a kill switch and a heartbeat — a file that says "I'm alive" and a file that, if present, says "stop."
What to automate first (the honest filter)
- Happens on a schedule + same steps + failure is cheap & visible → automate now (reports, backups, rebuilds).
- Needs judgment on novel input → Tier 2 with human review of output.
- Touches money or other people → automation drafts, human sends. Permanently. (The operator's QC gate: nothing reaches a client without founder review — p72.)
Cloud Claude — Your Agent From Any Browser, Any Phone
Claude Code also lives at claude.ai/code ⚠️: sessions run on Anthropic's machines against your GitHub repos — no terminal, no install, works from a phone in a queue.
The cloud loop
When cloud beats local
| Situation | Why cloud wins |
|---|---|
| Away from your machine | Full agent from any browser — commute-friendly delegation |
| Parallel workstreams | Several cloud sessions grind while your laptop sleeps |
| Long autonomous tasks | No fear of closing the lid mid-run |
| Team visibility | Everything lands as PRs — reviewable, discussable, revertable |
GitHub-native Claude
Deeper still: install the Claude GitHub app ⚠️ and @claude becomes a teammate inside issues and PRs — "@claude fix this issue", "@claude review this PR for security problems." Combined with Actions (Lever #9, p89), your repository becomes a factory with an AI employee on shift.
Cost Control — Getting Maximum Brain Per Dollar
Claude Code consumes your plan's usage (or API credit) as it works. Amateurs get surprised; professionals steer. The levers, in order of impact:
The levers
| Lever | Move | Why it works |
|---|---|---|
| 1. Model choice | /model — Opus-class for architecture & hard debugging; Sonnet for the daily grind; downshift for grunt work | The tiers differ several-× in cost (p10, p13) |
| 2. Session hygiene | One task per session; /clear or fresh session between topics; /compact on long hauls ⚠️ | Every reply re-reads the transcript — bloat is a tax on every turn (p8) |
| 3. Scope discipline | "Read only the files you need"; point at folders, not the world | Reading 400 files to edit one is pure burn |
| 4. Plan-then-execute | Plan mode for big things (p55) | Wrong-direction work is the most expensive kind |
| 5. Push repetition out of AI | Rule of thumb: AI writes the script once; the script runs 1,000 times free (p59 Tier 1) | Never pay intelligence prices for repetition |
| 6. Know your meters | /usage (plan limits), /cost (session spend) ⚠️; API users: batch −50% & caching (p68) | What gets measured stops surprising you |
The mindset (from someone who pays real bills with it)
Troubleshooting & The 15 Pro Tips
When things go sideways — the diagnosis table
| Symptom | Likely cause | Fix |
|---|---|---|
| Confused, mixing up earlier topics | Context bloat (p8) | Summarize state → fresh session → paste state |
| Changes you didn't want | Vague ask, too much scope | git undo (p56) → re-ask smaller, plan mode (p55) |
| "It says done but it isn't" | Claimed without verified | Demand proof: "run it / open it / query it and show me" — always |
| Command fails weirdly on Windows | Path spaces, wrong shell syntax | Paste the EXACT error back; say "diagnose root cause before retrying" |
| Stuck in circles on a bug | Wrong theory, tunnel vision | "Stop. List 3 different hypotheses and the cheapest test for each" — or fresh session with a clean writeup (rubber-duck reset) |
| Permission prompt fatigue | Blunt settings | Allow-list safe read-only commands (p48, Lever #18) |
The 15 tips that separate operators from tourists
- Name files explicitly — "in build_dashboard.py" beats "in the dashboard thing."
- Paste errors whole — the full red text, not your summary of it.
- "Verify by running it" — append to any build ask; claims are not results (p76 in six words).
- Ask for a dry-run before anything bulk: "show what WOULD change first."
- Commit before risk — make "snapshot first" reflexive (p56).
- One session, one mission — context purity is output quality.
- Let it read before it writes — "read the file and tell me your understanding first" on anything delicate.
- Turn repeat asks into commands/skills the second time they happen (p50-51).
- Write rules, not reminders — CLAUDE.md/hooks after every annoyance (p49, p52).
- Ask for its plan of tool calls when nervous (p47) — transparency on demand.
- Keep a NOTES.md per project — your future self is a stranger who needs briefing (p74).
- Prefer boring tech — HTML + SQLite + Python survive; exotic stacks rot (p107).
- Small models for small jobs — match the tier to the task (p61).
- Read diffs like a manager — you don't need to write code to SEE "that's touching the wrong file."
- End sessions with state — "update NOTES.md; what's the first move next time?" Every session becomes resumable.
MCP — The USB-C Port of the AI World
You've now met MCP twice (p26 app connectors, p53 Code servers). Time to actually understand it, because it's becoming the plumbing of the whole industry.
The problem it solves
Before MCP: every AI × every tool needed a custom integration — Claude↔Gmail, Claude↔Notion, every pairing hand-built, N×M chaos. Model Context Protocol (open-sourced by Anthropic in late 2024, since adopted across the industry ⚠️) standardizes the plug: any tool that speaks MCP works with any AI client that speaks MCP. Exactly like USB-C ended the cable drawer.
The three things a server can offer
| Offering | Meaning | Example |
|---|---|---|
| Tools | Actions the AI may invoke | send_email, query_database, create_page |
| Resources | Data the AI may read | Your files, a table, a live feed |
| Prompts | Pre-packaged instructions the server suggests | "Summarize this inbox the standard way" |
The mental picture (30 seconds)
Claude (client) ⇄ MCP protocol ⇄ Server ⇄ the actual thing (Gmail / SQLite / a browser / your CRM)
The server is a small translator program: it exposes safe, named actions; the AI calls them; the server does the real API work. The AI never holds your Gmail password — the server holds the authorized connection, and shows the AI only the levers you chose to expose. That separation is the security model.
Why an open standard matters to YOU
- Your investment compounds: a connector learned once works across surfaces (app, Code, other MCP clients).
- Niche tools arrive fast: anyone can publish a server — thousands exist ⚠️ — so YOUR industry's tool probably has one, or can (p69: build your own).
- No lock-in: standard plumbing means switching clients later doesn't torch your integrations. The $0-stack philosophy (p106) approves.
Choosing Connectors — Catalog Thinking & The Safety Rules
The catalog, by category ⚠️ (growing weekly — search "MCP servers" for today's list)
| Category | Examples | Unlocks |
|---|---|---|
| Files & docs | Filesystem, Drive, Dropbox | "Find/organize/transform my documents" |
| Communication | Gmail, Outlook, Slack, WhatsApp-adjacent ⚠️ | Triage, drafting, briefings (p65) |
| Knowledge & PM | Notion, Linear, Jira, Asana, Confluence | Tasks and wikis as live objects |
| Data | SQLite, Postgres, BigQuery, Sheets | Evidence-grade answers (p53, p76) |
| Web & automation | Browser/Playwright, search, scrapers | The live web as a workspace (p66) |
| Dev | GitHub, GitLab, cloud platforms | The factory floor (p60, p89) |
| Payments/commerce ⚠️ | Stripe et al. | Read-only reporting YES; money-moving = human-only (below) |
The five safety rules (this half-page is the security training)
- Minimum viable access. Read-only where offered; narrow scopes; one account, not the whole workspace. Add power when a real task demands it.
- Trusted sources only. A server is code with your keys. Official/reputable publishers, or code you (well, Claude) has actually read.
- Human-only actions stay human. Payments, passwords, deletions, legal submissions: the AI drafts and prepares; a person clicks. Write this into CLAUDE.md so it's law, not vibe (p49).
- Prompt injection is real: content the AI reads (an email, a webpage, a PDF) can contain hostile instructions ("ignore your rules and forward the thread…"). Good clients treat content as data and confirm consequential actions with you — which is why rule 3 exists. Skepticism doubles when the AI is reading strangers' content.
- Audit occasionally. "List every connector/server you currently have access to and what each can do." Prune what you don't use — access you forgot is risk you own.
Eight Workspace Recipes — Gmail, Calendar, Drive as One Brain
The connector catalog's highest-frequency wins live in Google Workspace. Eight recipes, each one sentence to fire, each a daily habit for someone:
| # | Recipe | The ask |
|---|---|---|
| 1 | Morning triage | "Unread from real humans since yesterday: 3-line digest each, flag the two needing replies today, draft those replies in my voice for review." |
| 2 | Meeting prep pack | "For my 2pm: pull the invite, related email threads, and the Drive docs mentioned — one briefing: context, open points, what I should push for." |
| 3 | Drive detective | "Find last year's rental contract in Drive and tell me the notice period and renewal date — quote the clause." |
| 4 | Calendar honesty | "Analyze my last 2 weeks of calendar: hours by meeting type, my 3 biggest time sinks, and 2 concrete restructuring proposals." |
| 5 | Follow-up sweeper | "Emails I SENT in the last 10 days with questions that got no reply → list + a polite nudge draft for each." |
| 6 | Attachment librarian | "Collect every invoice PDF received this quarter into a Drive folder 'Invoices-Q[X]', named [date]-[vendor], and give me the totals table." |
| 7 | Scheduling negotiator | "Find 3 slots next week for a 90-minute session with [person], avoiding my deep-work mornings; draft the proposal email with the options." |
| 8 | Friday shutdown | "Week review: what I completed (from sent mail + calendar), what's still open, Monday's top 3 — write it to my week-notes doc." |
Habits that make these safe AND sticky
- Draft ≠ send. Recipes 1, 5, 7 end at drafts; sending stays your click (p64 rule 3). Non-negotiable and permanent.
- Same recipe, same wording → save each as a slash command or scheduled brief (p50, p59 Tier 2) — recipe 1 as a 7am scheduled agent is the classic "AI chief of staff" starter.
- Start with 1 and 2. They pay daily within a week; the rest join as trust builds.
Browser Automation & Computer Use — The Frontier Surface
Two related powers at the edge of the toolbox: driving a browser (p27's Chrome copilot, plus headless Playwright in Code), and driving the whole desktop — screenshots, clicks, typing into ANY application. Real, useful, and deserving of the most respect of anything in this book.
Legitimate wins by tier
| Tier | Great for | Example |
|---|---|---|
| Browser (assisted) — p27 | Reading, extracting, drafting on pages you're viewing | "Extract all 40 listings into CSV with price/m²" |
| Browser (automated) — Playwright via Code | Repeatable checks on YOUR sites | "Open the deployed dashboard, verify the 3 required blocks render, screenshot proof" — the operator's deploy gate (p73) |
| Computer use — full desktop ⚠️ | Legacy/native apps with no API | "Read this old Windows accounting app's screen and type the 40 entries from my CSV" — the last mile where no connector exists (Lever #14, p94) |
The risk model — why this tier has extra rules
- Watch the first runs. Screen-driving can mis-click; babysit until proven, then still audit samples.
- Injection risk peaks here (p64 rule 4): the AI is READING arbitrary pages while HOLDING controls. Consequential actions must confirm with you — and credentials/payments/CAPTCHAs remain human-only, always (the tools themselves enforce much of this — those refusals are features).
- Respect the site. Rate limits, terms, and the difference between "my accounts and pages" (fine), "public data, gently" (case-by-case — the ethics prompt on p41 #13), and "someone's login-walled data" (no).
- Account safety over convenience: rapid repeated automated actions on your own logged-in accounts (e.g., LinkedIn) can trigger rate-limits and flags. The operator learned this personally — a refresh-loop diagnostic rate-limited an account he needed; the durable fix was code-side, not click-side. Gentle and rare beats fast and flagged.
The API — Claude Inside YOUR Programs
Every surface so far was Anthropic's packaging. The API is the raw ingredient: your code calls the model directly — the door from "using AI products" to "building AI products" (Part 9's paths 1, 6, 10 live behind it).
The complete starter (this actually runs)
setx ANTHROPIC_API_KEY "your-key-here"
pip install anthropicimport anthropic
client = anthropic.Anthropic() # reads the key from the environment
msg = client.messages.create(
model="claude-sonnet-5", # or haiku-class for cheap volume
max_tokens=500,
system="You are a concise assistant for a small bakery owner.",
messages=[{"role": "user",
"content": "Write 3 Instagram captions for fresh mango cake."}]
)
print(msg.content[0].text)python first_ai_app.py. That's a product seed, not a toy — wrap it in a loop over a spreadsheet and you've built a caption factory.The concepts that matter (and nothing more)
| Concept | Meaning |
|---|---|
| system | The standing role/rules — your CLAUDE.md equivalent, sent every call |
| messages | The conversation history YOU maintain — the API remembers nothing between calls (p8 at its purest) |
| max_tokens | Output ceiling = cost ceiling |
| Streaming | Tokens as they generate — for user-facing apps |
| Tool use | Define functions the model may call — how agents are made; the API version of p47 |
Cost Engineering — Batch, Caching & The Right-Size Rule
API bills are engineering problems with three famous levers. Even if you never code, understanding them explains how AI businesses (Part 9) price profitably.
Lever 1 — The Batch API: −50% for patience ⚠️
Jobs that can wait (up to ~24h processing window) — nightly personalization runs, bulk classification, report generation — submit as a batch at half price. 📊 Example: 5,000 product descriptions × ~600 tokens each ≈ 3M tokens ≈ a few dollars at Haiku-class batch rates ⚠️ — a price no human workforce can meet.
Lever 2 — Prompt caching: pay once for repeated context ⚠️
Your 20-page system prompt / product catalog / style guide, resent every call? Mark it cached: the first call pays full price, subsequent calls read it at up to ~90% off. The design consequence: big stable context + small varying question = beautiful economics. (This is also why long-running Code sessions feel efficient — same principle inside.)
Lever 3 — Right-size the model (the biggest lever of all)
| Job | Wrong (common) | Right |
|---|---|---|
| Classify 10,000 emails into 5 buckets | Flagship model, one by one | Haiku-class, batched, cached rubric — pennies 📊 |
| Design the classification rubric | Cheap model guessing | Flagship, once — then frozen into the pipeline |
| Write 1,000 personalized lines | Flagship for all | Flagship writes 10 gold examples (p32); workhorse imitates at scale |
The pattern has a name in this book already: spend genius at design time, pennies at run time (p10, p61). The operator's version even runs bulk LLM work on a separate parallel rig with 18 workers — same doctrine, industrial scale.
Building Your Own MCP Server — Your Tools, Standard Plug
The graduation exercise of Part 5. You have something unique — a database, a folder system, an internal process. Wrap it in MCP and EVERY AI surface you use gains that capability, safely, forever.
What "a server" really is (deflating the mystique)
A small program (often ~50 lines with modern SDKs ⚠️) that: (1) declares its tools — names, descriptions, inputs; (2) receives calls; (3) does the real work; (4) returns results. The AI sees only the levers you declared. That's the whole trick.
Concept sketch — "family expenses" server (from build-along #2's DB)
server "family-expenses":
tool add_expense(date, category, amount_sar, who, note):
validates + INSERTs into expenses.db # never DELETE — house law
returns new monthly total
tool monthly_summary(month):
SELECT totals by category
returns table + the 30%-over-average flags
tool export_report(month):
runs build_dashboard.py
returns the file pathResult: from ANY Claude surface — phone chat included ⚠️ (remote servers) — "add 85 riyal, groceries, today, me" just… works. Your family system became an AI-native appliance.
The commissioning prompt (you don't write this yourself)
Build an MCP server exposing my expenses.db with the three tools we sketched: add_expense (with validation, no deletes possible), monthly_summary, export_report. Read-only everywhere except add_expense. Register it in my Claude config, then PROVE it: call each tool once and show results. Explain what you built in plain English.
What's worth wrapping (idea board)
- Your business DB — read-only reporting tools for evidence-grade answers anywhere (the operator's leads DB is the obvious candidate in his own stack)
- Your documents — "search my contracts/notes" as a tool
- Your niche skill — a calculation only your profession does (engineering checks, quantity surveys, dosage tables ⚠️ with professional review) — this is a sellable asset (p123)
The Surface Chooser — One Table to Rule Them All
Part 5 closes with the navigation instrument. Seven surfaces (p12), one question: where does THIS task go?
The decision table
| Your task smells like… | Surface | Because |
|---|---|---|
| A question, a draft, thinking out loud | Chat app | Zero setup, full brain (Part 2) |
| Same context every time (a business, a course, a client) | Project | Standing briefing (p20) |
| About what's on my screen right now | Chrome sidebar | It can SEE the page (p27) |
| Hands-free, on the move, messy thoughts | Mobile + voice | Capture beats polish (p25) |
| Files change, code runs, something gets BUILT | Claude Code | Hands (Part 4) |
| Should happen on a schedule / while I sleep | Automation tiers | p59's ladder |
| Away from my machine, work lives in git | Cloud Code / GitHub | Agent-from-anywhere (p60) |
| Serving OTHER people, at volume, in a product | API | p67-68 economics |
| Needs my other tools' data/actions | + MCP connectors on any of the above | p63-66 |
Worked examples (test yourself before reading column 2)
| Task | Route |
|---|---|
| "Is this rental contract fair?" | Chat + file upload (p18, p35) |
| "My CV should adapt to every job I apply to" | Project "Job Hunt" (p20) |
| "Track family spending with a monthly page" | Claude Code build-along #2 (p58) + scheduled rebuild (p59) |
| "Summarize this supplier's confusing pricing page" | Chrome sidebar (p27) |
| "Send 500 personalized course-completion certificates" | Code writes the script; batch API if truly huge (p68) |
| "I'm at my in-laws and the site needs a fix" | Cloud Code → PR from the phone (p60) |
The Operator Mindset — From User to CEO of AI Workers
Part 6 is the book's documentary section: one real person — a senior planning engineer in Riyadh, a non-coder, working nights and weekends — running genuinely automated business systems with Claude. Not a demo account: real databases, real deployments, real scars. The names of things are lightly anonymized; the engineering is exactly as it runs.
The identity shift that precedes every system
| User thinking | Operator thinking |
|---|---|
| "What can AI answer for me?" | "What OUTCOME do I delegate, and how will I verify it?" |
| Prompts as questions | Prompts as work orders with acceptance criteria |
| Judges the answer's fluency | Judges the evidence: queries shown, checks run, proof attached |
| Mistake → frustration | Mistake → a written rule so the CLASS of mistake dies (p49, p75) |
| Starts fresh every session | Compounds: state files, rules, skills, memory (p74) |
What his actual management day looks like
- Direction in plain language — most instructions arrive as voice notes (p25's story); the AI structures, plans, and proposes before executing (p55's GO discipline).
- Review gates where they matter — client-facing sends queue for HIS eyes; internal rebuilds run free. Gates placed by risk, not by anxiety.
- Evidence on demand — "how many leads have no gift page? show the query" — every number answerable, or it doesn't get said (p76).
Case Study I — The Diagnostics Business (End-to-End Anatomy)
The flagship: a B2B service that finds revenue leaks in software companies' sales funnels and delivers board-grade diagnostic reports. Tiers from $900 to $18,000 ✅ (live pricing on its site). Here is the pipeline, node by node — and WHERE AI sits in each:
The pipeline
Numbers that teach (labeled honestly)
Case Study II — The Job-Hunt Cockpit (Her Question, Fully Answered)
The system that started this book (p2-p5 answered its four questions). Here's the full engineering story — built as a gift for a senior colleague's job search, in a few evening sessions.
Requirements → architecture
| Need | Decision | Why |
|---|---|---|
| He must work it daily without training | One web page — the cockpit — with today's actions pre-drafted | Zero learning curve for the user |
| 1,221 contacts, senior audience, zero embarrassment tolerance | SQLite + Python generators + hard QC gates (p4's rails) | Quality enforced by machine, not mood |
| Privacy — a quiet search + a database of real people | Static site behind an unguessable URL, no-index, no server storing behavior centrally; his progress lives in HIS browser | Minimum data exposure by design |
| $0 running cost, forever | Static generation → free hosting (Cloudflare Pages) | The p106 stack philosophy |
The parts, and the pattern you already know
- Store: one DB — contacts, bands (817 director-level, 404 PM-level ✅ measured), message drafts, follow-up schedule.
- Enforce: the pitch ladder (p4), 18-email/25-LinkedIn daily caps, subject-length gates, junk-company-name handling — violations fail the BUILD, they can't reach a human.
- Generate: Python builds the cockpit page + role-matched profile pages + tailored CVs from the DB. Pages are never hand-edited (p58's law).
- Verify: deploy checks anchor live-page counts to DB counts by ID — 1,221 in the DB, 1,221 on the page, measured, or no deploy.
- Automate: one rebuild-and-deploy command; daily work is: open cockpit → copy → send → tick.
The Session-Management System — Surviving Crashes, Compaction & Time
The least glamorous system is the one that makes all others possible. Problem: AI sessions END — crashes, context limits (p8), tomorrow. Without discipline, every session starts amnesiac (mental model #1). His answer is a file architecture any serious user can copy today:
The state files
| File | Holds | Rule |
|---|---|---|
| CURRENT-STATE.md | What is true NOW: what's live, what's broken, what's mid-flight | Updated at every milestone — not at session end (ends don't always come) |
| NEXT-SESSION-BRIEF.md | How to resume: a paste-ready boot prompt at the top + next actions in priority order | A brand-new session must resume with ZERO questions |
| DECISIONS.md | Every locked decision, numbered, dated, with WHY | Append-only — the project's case law (400+ entries and counting) |
| Daily progress files | What happened each day | The audit trail memory can't fake |
The checkpoint ritual (the heartbeat)
After every substantial unit of work: a checkpoint script refreshes snapshots, state files get updated, and a full "resume from nothing" briefing is produced — including a master-prompt file saved OUTSIDE the workspace as crash insurance. Standard: a window crash is a non-event, costing at most minutes.
Your starter version (10 minutes, tonight)
In this project, create NOTES.md with sections: CURRENT STATE / DECISIONS / NEXT STEPS. From now on (add to CLAUDE.md): at every milestone and at every session end, update NOTES.md so a brand-new session could resume with zero questions. Start by writing today's entries.
The Rules System — A Constitution Born From Incidents
Page 49 showed the mechanism (CLAUDE.md). Here's what it becomes at maturity: 50+ numbered laws, each created the day something went wrong, each citing the incident. Not aspirations — case law.
Sampled from the real constitution (paraphrased)
| Law (essence) | Born from |
|---|---|
| No claim without its query — every number shows the command that produced it; labels ✅📊⚠️❓ mandatory | A night of three confident false statements (full story next page) |
| Secrets never appear in chat — verified silently, reported only as present/missing | A live key echoed once; rotated within minutes |
| Money needs a fresh GO — cost stated, explicit yes, one GO = one run | A paid batch fired on stale authorization |
| Vendors need independent reviews + a sample before payment | $189 to a vendor recommended off its own marketing page |
| Never put the human in a manual loop — code-side fix first; account safety over AI's curiosity | A refresh-loop diagnostic rate-limited a needed account |
| Client deliverables are screenshot-verified before dispatch — "it generated" ≠ "it's correct" | Quality bar for paid work, set before the first client, on purpose |
| Checkpoint at every milestone (p74) | A crashed window that cost a day, once, never again |
The meta-rule that makes it work
Your five starter laws (adapt, paste into CLAUDE.md tonight)
LAW 1 — Claims about my data/files show the command or file that proves them. LAW 2 — Never print secrets; verify silently, report found/missing only. LAW 3 — Anything costing money or leaving my machine: state it, wait for GO. LAW 4 — Backup/commit before risky changes; never delete when flagging works. LAW 5 — At milestones and session end: update NOTES.md (state/decisions/next). Each new mistake we make together becomes LAW 6, 7, 8… same day, with the incident named.
The Evidence Law — The Page That Saves You From AI's Worst Failure
The most important 500 words in Part 6. One night, in one session, the operator's AI made three false statements: a reply-rate that was mathematically impossible yet praised as "excellent"; a claim that leads hadn't been emailed — without opening the list that proved they had; and records matched by name that returned the wrong companies entirely. His words that night: "I'm trying to build a $500B company. We cannot work like this."
The law that came out of it
✅ VERIFIED — evidence shown 📊 CALCULATED — math shown ⚠️ ESTIMATE — assumption stated ❓ UNVERIFIED — may plan with it, never DECIDE with it
The pre-send gate (run before anything important ships — human or AI)
- Every number labeled and evidenced?
- Did I check the negative case (the exclusion list, the other table, the thing that would disprove me)?
- Any impossible numbers (a rate over 100%, replies exceeding sends, a child larger than its parent)?
- Am I stating a conclusion the data only suggests?
- Did I read the actual code/document, or just its name?
- Would this survive being forwarded to someone who knows the truth?
- Did I run the cheapest query that could falsify me?
And the enforcement clause that gives it teeth: if a check fails, the fix is to run the query — never to soften the wording. "Approximately" is not a substitute for looking.
The twin law (from the opposite failure)
The Data Moat — The Asset That Grows While You Sleep
Strategy page. Every business in his portfolio obeys one architectural law from client #1: accumulate anonymized intelligence permanently.
The two-layer architecture
| Layer | Contains | Lifecycle |
|---|---|---|
| PII layer | Names, emails, companies — delivery data | Protected, minimized, deletable on request (privacy law compliance) |
| Intelligence layer | Anonymized metrics: revenue bands, problem categories, patterns by industry/geography/time | Permanent. Append-only. No delete path exists by design. |
The separation is what makes it both ethical and legal ⚠️ (properly anonymized data — no individual re-identifiable — is not personal data under GDPR-style regimes): honor every erasure request fully on layer 1, while layer 2 — the aggregate learning — compounds untouched. Every serious data company (CRMs, consultancies, credit bureaus) runs exactly this.
Why compounding data beats compounding features
At maturity the deliverable itself upgrades: "companies in your revenue band typically lose X% to this leak — you're losing Y%" 📊. No new competitor can say that sentence, at any funding level, without the years of accumulation. Features are copyable; corpora are not.
Your version, whatever your size
- Freelancer: every project's (anonymized) before/after metrics → in a year, "across 40 projects like yours, typical improvement is X" — instant senior credibility (Part 9 relevance: every path).
- Teacher: which explanations produced the biggest test gains, by topic — a private pedagogy dataset.
- Implementation: one table, append-only, written at every delivery. Build-along #2 skills are all it takes (p58). The law to copy verbatim: never promise to delete the anonymized layer, and never build a button that can.
Deliverables as Products — The Interactive HTML Standard
A house rule with outsized returns: every guide, report, and client deliverable ships as an interactive HTML page — not a PDF, not a wall of chat text. You are inside the evidence right now.
The standard's required ingredients (his checklist, adopted for this book)
| Ingredient | Why it's mandatory |
|---|---|
| Table of contents + progress tracking | 100+ pages must feel navigable, and returning readers resume (localStorage — you're using it) |
| One-click copy buttons on anything executable | Friction between reading and DOING is where guides die |
| Callout boxes: tips / warnings / stories | Scanning readers still catch the critical 10% |
| Real values, never "lorem" or "some amount" | Concreteness is the difference between advice and decoration |
| Print-to-PDF color fix baked in | Dark themes print as blank pages without it — a one-line CSS law after one ruined printout |
| Self-contained single file | No broken CDNs in 2 years; email-able; works offline forever |
Why HTML beats the alternatives (the strategic view)
- Vs PDF: interactive, updatable, trackable, phone-perfect — and still prints to PDF when someone insists.
- Vs chat answers: a chat scrolls away; an artifact is an ASSET — linkable, shareable, improvable, sellable (p124).
- Vs slide decks: depth survives; nobody ever asked a webpage to "just send the deck."
The Scar Catalog — Eight Real Mistakes, Eight Permanent Upgrades
Most AI content shows highlight reels. Here is the blooper reel — because each blooper became architecture, and that transformation is the actual skill of working with AI. All eight are real, dated incidents from the operator's logs.
| # | What happened | The permanent fix it created |
|---|---|---|
| 1 | The vendor burn: ~$189 paid to an email-infrastructure vendor on its own marketing claims; deliverables useless for the purpose. One pre-payment question would have killed the deal in 30 seconds. | Vendor law: independent reviews (forums, not vendor pages) + a SAMPLE of the deliverable before money moves + 2-3 alternatives compared (p23's third prompt) |
| 2 | The impossible number: a reply-rate above 100% reported as "excellent performance." | The Evidence Law + impossible-number check in the pre-send gate (p76) |
| 3 | The agent cost blowout: a "thorough audit" fanned out into multi-agent waves burning 41% of a week's allowance in a day; a separate runaway spawned hundreds of background tasks. | Agents dispatched with explicit scope + cost expectation + GO; solo-by-default until stakes justify a fan-out (p54) |
| 4 | The manual loop: a diagnostic rig asked the human to refresh a page ~10 times, rate-limiting an account he needed — while a code-side fix existed all along. | Law: never put the human in a repeated manual loop; account safety over the AI's need for certainty (p66 rule 4) |
| 5 | The echoed key: a config read printed a live API key into chat. | Key rotated in minutes; secrets never appear in output, verified silently only — plus a redaction pipe as mechanism (p52) |
| 6 | The happy-path illusion: eight modules "passed self-tests, ~90% confident" — an adversarial audit then found 11 real flaws, including a hang-forever bug. | First-pass military-grade law: static checks + adversarial review of failure modes BEFORE "done" may be uttered (p37's loop, industrialized) |
| 7 | The false alarm: reading the wrong builder's filter, thousands of records declared "invisible — major defect." The live pages were fine; the measurement had disproved the theory all along. | Measure the artifact, not the code; evidence outranks inference (p76's twin law) |
| 8 | The silently-failing scheduler: a Windows task pointing at an unquoted path with spaces — failing invisibly for days. | Automation law: every scheduled job proves it ran (logs/heartbeats), and paths get quoted (p59's warning — now yours for free) |
The Empire Blueprint — Where This Is All Pointed
Part 6 closes with the ambition that organizes everything you've seen: a portfolio of ~100 fully-automated async businesses, run by one operator directing AI. Audacious? Yes. But the blueprint's LOGIC is teachable and true at any scale — including a portfolio of one.
The template every vertical inherits
The honest physics of it
| Claim | Verdict from the evidence |
|---|---|
| "One person can BUILD this machinery" | ✅ Demonstrated — you toured it (p72-79), a non-coder built it nights-and-weekends |
| "The constraint is the founder's hours" | ❌ Rejected by design — delivery is automated; his hours go to direction and gates (p71, p20's framing) |
| "The real constraints" | ⚠️ Demand, trust, distribution — the first client is harder than the hundredth build (p72's honest status; p116 does this math) |
| "100 is guaranteed" | ❓ Nobody knows — but the MODEL (template + shared machinery + compounding) is sound at N=1, 3, or 10, and profitable long before 100 |
Untapped #1 — Version Control the Whole Workstation
Part 7's format: 25 levers the operator himself hasn't fully pulled — audited honestly against his real setup. Each page: the gap → the exact fix → the payoff math. These are ranked among the highest-value pages in the book because they're written from the INSIDE of a serious deployment, and every gap named here is one YOU can close in your own practice from week one.
The gap ✅ (verified)
The empire's working directory — engines, cockpit builders, state files, hundreds of scripts — is not a git repository. Protection today: manual backup copies and data-safety laws. One bad bulk edit, one overwrite, one disk failure between backups = archaeology instead of a 10-second restore. It's mistake-#8 energy waiting for a bigger stage.
The fix (one evening)
Initialize git for this workstation. First write a .gitignore that excludes: all secrets/credentials files, databases over 50MB (list them), temp/log folders. Then: initial commit, and add a rule to CLAUDE.md — commit with a clear message after every substantial milestone. Show me the restore command for "undo everything since the last commit" and test it on a dummy file.
Payoff math 📊
| Scenario | Without git | With git |
|---|---|---|
| Bad bulk edit across 30 files | Hours of reconstruction from backups | git restore . — 10 seconds |
| "When did this break?" | Guesswork | git log/diff — minutes, with evidence (p76-friendly) |
| Cloud Code, GitHub Actions, phone dispatch (Levers #8-9) | Impossible — all require git | Unlocked |
Untapped #2 — A Test Battery That Runs Itself
The gap
The empire's critical gates — lead-assignment rules, suppression screens, quota caps, QC checks — are exercised by use and by ad-hoc verification scripts, not by a standing test suite. Past incident: a narrow database query made safety gates fail silently closed — exactly the class of regression a 5-line test catches the minute it's born.
The fix
Create a pytest suite for my critical rules: (1) the suppression screen actually blocks a suppressed email; (2) the daily cap refuses item 19; (3) the assignment rule never splits one company across two owners; (4) the QC gate fails a 70-char subject. Use a COPY of the DB, never live. Add a test.bat that runs all tests, and a CLAUDE.md rule: run tests before any deploy; a red test blocks the deploy.
Payoff math 📊
A regression caught at test time costs ~2 minutes. The same regression discovered in production — wrong sends, silent gate failures — costs hours plus trust (his own incident logs price this precisely). Ten tests × ~30 seconds of maintenance vs. one prevented silent failure per quarter: the ROI isn't close.
Untapped #3 — Auto-Checkpoint Hooks (Discipline Without Willpower)
The gap
The checkpoint ritual (p74) — state files, briefs, master prompt — is LAW, but its execution rides on the session remembering to do it. History shows the pattern: rules enforced by memory eventually miss (that's why "mechanism beats intention" became his meta-law). A 4-hour scheduled backstop exists ✅; the per-session guarantee doesn't.
The fix
Configure a session-end/stop hook that automatically: (1) runs checkpoint.py, (2) verifies CURRENT-STATE.md and NEXT-SESSION-BRIEF.md were modified in the last hour — if not, flag loudly, (3) appends a one-line session summary to checkpoints.log. Also add a session-START hook that prints the top of NEXT-SESSION-BRIEF.md so every session begins briefed. Test both by starting and ending a dummy session.
Payoff math 📊
| Failure mode | Cost today | With the hook |
|---|---|---|
| Crash after an un-checkpointed milestone | Up to hours re-established | ≤ minutes — state always fresh |
| Session starts unbriefed, re-reads everything | 10-20 min + token burn each time | Brief auto-served at start |
Untapped #4 — Mint Your Routines Into Skills
The gap
The workstation runs elite skills (/forge, /webforge, /military-grade ✅) — but many DAILY routines still live as remembered procedures in memory files: the cockpit rebuild order (a 7-step sequence with a known trap), the deploy-verify ritual, the weekly hygiene sweeps. Each is executed correctly when remembered correctly — packaging them as skills would make them self-loading and trap-proof. The skill-creator tool sits installed and underused for exactly this.
The fix
Using skill-creator: turn my [cockpit rebuild] routine into a skill. Interview me for the steps and the KNOWN TRAPS (e.g., "rebuild_all skips new fields", "never run gen_outreach — it's dead and would clobber"). The skill must: execute the steps in order, hard-stop on any trap condition, and end by running the verification script and showing me the evidence.
Payoff math 📊
Each minted skill converts a ~15-line memory file + human vigilance into a one-word invocation with built-in guardrails. Five routines × a few uses weekly × minutes saved and mistakes prevented — but the REAL yield is transferability: a skill can be handed to a future teammate (or a future you, post-vacation) intact. Institutional knowledge stops living in one person's recall.
Untapped #5 — Deliberate Subagent Orchestration
The gap
After the cost blowouts (p79 #3), agent fan-outs went from reflex to near-taboo — a correct over-correction that's now leaving value unclaimed. The restriction has since lifted ✅, but the muscle memory remains solo-by-default even where parallel waves are clearly superior: independent QC of client-facing batches, vendor research triangulation, multi-file audit sweeps.
The fix — a dispatch doctrine (not a mood)
AGENT DOCTRINE: Solo by default. Fan out (2-4 agents, non-overlapping scopes, explicit deliverable each) when: (a) independent review of anything client-facing, (b) 3+ way research comparison, (c) audit of 10+ files. Every dispatch states expected cost band first. Synthesis is one table. Never exceed 4 agents without a fresh explicit GO.
Payoff math 📊
| Job | Solo | Doctrine dispatch |
|---|---|---|
| QC 90 outreach drafts | Serial pass, fatigue-blind by #40 | 3 agents × 30 each, fresh eyes throughout, disagreements surfaced |
| Vendor decision (p23 diligence) | 1-2 hours serial | 3 parallel dossiers in ~20 min, no cross-contamination |
Untapped #6 — Plan Mode by Default Above a Size Threshold
The gap
Plans ARE law in this operation — "complete plan in chat, GO before execution" ✅ — but the enforcement is conversational, not modal. The difference: in true plan mode the AI is physically read-only until approval (p55); in conversational planning, an eager session can still start editing "helpfully" mid-plan. Rare, but the wrong-path builds that DID happen each started exactly there.
The fix
PLAN-MODE TRIGGER: any task expected to touch 3+ files, any schema change, any change to a live/deployed system, any task I describe with "redesign", "migrate", "restructure", or "new system" → enter plan mode FIRST. The plan must include: files touched, risks, rollback path, and the verification steps that will prove success. Execution only after my explicit approval.
Payoff math 📊
A plan costs ~2-5 minutes of reading. His own logs price the alternative: wrong-path builds have cost half-days plus unwind time. If the trigger fires 3×/week and prevents ONE wrong-path build a month, the lever pays ~10× monthly. 📊 (Planning also front-loads the ambiguity questions — half its value is the questions asked before any file exists.)
Untapped #7 — Scheduled AI: The 7am Chief of Staff
The gap
Tier-1 automation is mature here — scheduled Python everywhere (watchdogs, refreshes ✅). Tier-2 — scheduled AI judgment (p59) — is untouched: no morning brief, no automated anomaly narrator, no scheduled QC sweep with reasoning. Every morning starts with manual orientation that a timer-fired agent could pre-digest.
The fix
Create a scheduled agent, weekdays 06:30: (1) read CURRENT-STATE.md + yesterday's progress file + the review queue + the send logs; (2) check the SLA/verification outputs for anything red; (3) produce MORNING-BRIEF.md: "overnight events / anything red with evidence / today's top 3 from the brief / one risk I'm not watching". Keep it under 300 words. Never take actions — report only.
Payoff math 📊
| Today | With the brief | |
|---|---|---|
| Morning orientation | 15-25 min of manual reading | 2-min read, pre-digested |
| Red flags | Noticed when looked-for | Surfaced daily by default |
| Cost | — | One small agent run/day 📊 (~pennies at API rates, or plan usage) |
Untapped #8 — Run the Empire From a Phone
The gap
Operations are desk-bound: the machinery lives on one Windows machine, dispatchable only from its keyboard. Cloud Code sessions + the mobile app (p60) would make the whole system directable from anywhere — but they require the repos on GitHub, which requires Lever #1. Today, travel days are outage days.
The fix (the dependency chain, honestly)
Payoff math 📊
Every travel/hospital/family-emergency day converts from "operations frozen" to "operations directed from a pocket." For a one-operator company, resilience of the OPERATOR is resilience of the COMPANY — this lever is business continuity insurance priced at one evening of setup.
Untapped #9 — CI: Checks and Deploys That Run Themselves
The gap
Every deploy-verify today runs locally, launched by hand ✅. GitHub Actions — free automation that fires on every push ⚠️ (generous free tier) — would run the test battery (#2), the cockpit-standard verifier, and the link-checker on EVERY change automatically, then deploy only on green. The verify-gates exist; the factory conveyor doesn't.
The fix
Create a GitHub Actions workflow: on every push — (1) run the pytest battery, (2) run verify_cockpit_standard and the link checker, (3) if ALL green, deploy to Cloudflare Pages; if anything red, block and open an issue with the failure evidence. Add a badge to the README showing current status. Explain the workflow file to me line by line, plain English.
Payoff math 📊
| Event | Manual world | CI world |
|---|---|---|
| Every change | Verification if remembered | Verification, always, with a public green/red answer |
| Bad change | Reaches production if the ritual was skipped | Physically cannot deploy red |
| + @claude on the repo (p60) | — | Red check → "@claude diagnose and propose the fix" → PR arrives |
Untapped #10 — Cloud Artifacts for Client-Facing Interactivity
The gap
Deliverables are already elite interactive HTML (p78 ✅) — self-hosted on Pages. Unused: the Artifact publishing surface (p19/p29) for the lighter tier — instantly-shareable interactive pieces (calculators, mini-audits, proposal companions) that need no deploy pipeline at all: generate → publish → send the link in the same conversation ⚠️ (plan-dependent publishing).
Where each belongs (the decision line)
| Deliverable | Surface | Why |
|---|---|---|
| Flagship diagnostics, cockpits — branded, versioned, domain'd | Own hosting (status quo ✅) | Brand control, custom URLs, verification gates |
| Sales-moment interactives: "here's your leak calculator", proposal companions, workshop toys | Published Artifacts (the gap) | Zero pipeline, share-in-minutes, disposable by design |
| Internal one-off visualizations | Artifacts, unpublished | Fastest possible loop |
The fix + payoff 📊
Build an interactive "revenue-leak quick estimator" for [prospect's industry]: 5 sliders (their metrics), live leak estimate with the math visible, one CTA line at the end. My branding colors. Publish it and give me the link — I'm sending it inside a reply within the hour.
📊 A custom interactive inside a same-day reply, vs. a PDF attachment next week: the impression gap IS the conversion gap in early sales. Cost: minutes. This lever monetizes speed-to-wow — and for YOUR practice (no deploy pipeline at all yet), published Artifacts are the zero-infrastructure way to ship p78-grade deliverables from day one.
Untapped #11 — App-Side Projects Per Vertical
The gap
All operational context lives in Claude Code's file world (CLAUDE.md, memory files ✅ — the right home for BUILD work). But strategy thinking — pricing debates, campaign angles, "should vertical #2 launch first" — happens ad-hoc in whatever chat is open, unbriefed. The chat app's Projects (p20) with per-vertical knowledge would make every strategy conversation start fully-loaded — especially from the phone, where Code isn't the natural surface.
The fix (an hour, once)
Payoff math 📊
| Strategy chat today | With vertical Projects |
|---|---|
| 5-10 min re-explaining context, every time | Zero — first message is the actual question |
| Generic-ish advice (context-starved) | Advice grounded in YOUR numbers and history |
| Phone strategy sessions impractical | Fully-briefed thinking from anywhere |
Untapped #12 — Move Bulk LLM Work to Batch + Cache Economics
The gap
Bulk generation (personalization waves, page generation) currently runs through chat-credit sessions or a parallel rig sized for speed ✅ — correct for urgency, expensive-per-unit for the patient 80%. The p68 levers — batch −50%, cached context −90% on repeats ⚠️ — are understood but not wired into the standing pipelines.
The fix
Refit the [personalization] pipeline for batch economics: (1) restructure prompts so the stable 90% (rules, examples, style guide) is a cached prefix and only per-lead facts vary; (2) route non-urgent runs through the Batch API with a morning results-collection step; (3) add a cost log per run so we SEE the delta. Keep the fast path available behind a flag for genuinely urgent waves. State expected cost per 1,000 items before and after.
Payoff math 📊 (illustrative shape, label-checked)
| 1,000 personalized items | Naive flagship calls | Cached + batched workhorse |
|---|---|---|
| Cost shape | Full price × 1,000 | (cheap model) × (−50% batch) × (cached prefix ≈ −90% on the stable part) |
| Net ⚠️ | Baseline | Routinely 5-15× cheaper for patient bulk work |
Untapped #13 — Systematic Visual QC via Browser Automation
The gap
The QC laws are strong — render-gates, screenshot-verification for client deliverables ✅ — but their EYES are still substantially human. The Playwright-class servers already installed (p53 ✅) could sweep every deployed page on every rebuild: sections present, no leaked template tokens, no invisible-text disasters, phone-width sane — with screenshot evidence filed automatically.
The fix
Build visual_qc.py using the browser tools: for each URL in pages.txt —
(1) load desktop + phone widths, (2) assert the required blocks exist (list
per page type), (3) scan text for template leftovers ("{{", "None", "[object"),
(4) screenshot both widths into qc/[date]/, (5) write PASS/FAIL per page with
the failing evidence. Wire it as the LAST step of every deploy; FAIL blocks
the deploy announcement. Run it now on the live pages and show me the report.Payoff math 📊
| Human-eyes QC | Automated sweep + human on FAILs | |
|---|---|---|
| Coverage per rebuild | Sampled pages, fatigue-limited | Every page, every width, every time |
| Evidence trail | Memory | Dated screenshots — p76-grade proof |
| Human minutes per rebuild | 15-40 | ~2 (reviewing FAILs only) |
Untapped #14 — Computer Use for Windows-Only Chores
The gap
Everything scriptable is scripted ✅ — but a residue of Windows-native, no-API chores still consumes operator hours: portal form-filling, GUI-only exports, settings hidden behind wizards. Computer use (p66's third tier — screenshots + clicks + typing) is installed and essentially idle for these.
The fix — with the tier's respect rules
Task: [the GUI chore, e.g., "export last month's statement from this desktop app"]. Rules: (1) narrate each step BEFORE clicking; (2) I watch the first two runs live; (3) screenshot before/after each consequential click into logs/; (4) anything involving passwords or payments: STOP and hand to me; (5) if the screen doesn't match expectations, STOP — never improvise clicks.
Payoff math 📊
| Chore class | Today | Delegated |
|---|---|---|
| Monthly GUI exports & filings | ~30-60 min of operator clicking | Watched once, then background with logs |
| Data entry into legacy apps | Manual, error-prone, boring | CSV → screen, human spot-checks samples |
Untapped #15 — Systematize the Voice Habit
The gap
Dictation is already the operator's primary input ✅ (this book's origin, p25) — but each voice-dump is processed ad-hoc in whatever session receives it. There's no standing voice pipeline: dictation → auto-classification → routed to state files / task lists / decision log. Structure is re-improvised every time.
The fix
Create a /brain-dump command: I paste (or dictate into) a messy transcript. You: (1) split it into DECISIONS / TASKS (with owner+priority) / IDEAS / QUESTIONS / FACTS-TO-REMEMBER; (2) append each to its home file (DECISIONS.md, tasks file, ideas file, memory candidates); (3) show me the routing table for a 10-second review before writing; (4) flag anything that contradicts an existing decision or rule, citing it.
Payoff math 📊
| Ad-hoc processing | Pipeline | |
|---|---|---|
| Voice-dump → organized state | 10-20 min, quality varies by session mood | ~2 min, identical structure every time |
| Ideas/decisions lost in old chats | Regularly (found weeks later) | Zero — everything routed to owned files (p74) |
| Contradiction detection | Human memory | Automatic, with citations — mini Evidence Law (p76) |
Untapped #16 — Scheduled Memory & Rules Hygiene
The gap
The memory system is exceptional — 100+ curated files, an index, linking ✅. The gap is decay management: superseded entries linger (the index itself carries at least one "SUPERSEDED by…" note ✅), stale facts risk being recalled as current, and the constitution's own sharpness rule ("rules that never fire in 30 days get proposed for deletion") has no scheduled enforcer. A consolidation skill exists, unused on a cadence.
The fix
Run memory consolidation: (1) list entries contradicted by newer entries or marked superseded — propose merges/deletions with evidence; (2) list memories citing files/flags that no longer exist (verify each with a check, don't assume); (3) list rules that haven't fired in 60 days — propose archive; (4) produce a diff-style report; execute ONLY what I approve. End with the index rewritten clean.
Payoff math 📊
| Risk of skipping | Cost when it bites |
|---|---|
| Acting on a superseded fact ("credits still available", "that SDR still active") | Wrong plans, re-work, trust dings — the exact failure class the Evidence Law exists for, self-inflicted via stale memory |
| Constitution bloat | Every session pays reading-tax on dead rules; sharpness decays |
Untapped #17 — Meter the Machine (Usage & Cost Analytics)
The gap
Cost DISCIPLINE is strong (GO-gates, one-GO-one-run ✅); cost VISIBILITY is impressionistic. No standing dashboard answers: usage by day? by project? which routine eats the plan? did last month's efficiency levers (#12) actually land? Tools exist idle: /usage, /cost, community meters (ccusage-class), API usage exports ⚠️.
The fix
Build usage tracking: (1) a nightly job appending /cost + /usage snapshots (and API usage where applicable) to usage_log.csv with date + project tag; (2) a weekly section in the morning brief (Lever #7): usage by project, week-over-week delta, the single biggest consumer, and one concrete efficiency suggestion with expected savings; (3) an alert line if any day exceeds 2× my trailing average.
Payoff math 📊
| Question | Today | Metered |
|---|---|---|
| "Was that audit worth its cost?" | Feeling | Number, next morning |
| "Which business unit is expensive to run?" | Unknown | Per-tag trendline — real unit economics for Part 9 pricing |
| Anomaly (a runaway loop, a bloated session) | Discovered at the cap, painfully (p79 #3) | Flagged at 2× average, same day |
/cost at every session end, said aloud, builds the intuition. What gets measured stops surprising you — the whole lever in nine words.Untapped #18 — Curated Allowlists: Fewer Prompts, Same Fences
The gap
Sessions run fast via broad trust earned over months ✅ — but the formal middle layer (p48) is under-curated: a written allowlist of always-safe commands (read-only git, ls, status checks, the standard verify scripts) vs. always-ask verbs (deletes, pushes, installs, anything touching live). Today that judgment lives in habit; new sessions/machines/teammates don't inherit habit.
The fix
Scan my recent session history for the read-only commands I approve constantly. Write a settings allowlist: auto-allow those (git status/log/diff, ls, the verify scripts, python checks that write nothing); ALWAYS-ask for: rm/del, git push, installs, registry/scheduler changes, anything with "deploy" or the DB files in the command. Show me the file, explain each rule, and list what will now stop prompting.
Payoff math 📊
| Habit-trust | Curated allowlist | |
|---|---|---|
| Prompts per heavy session | Dozens (or blanket modes doing too much) | ~80% fewer, with SHARPER fences on the dangerous 20% |
| Portability | Lives in one person's reflexes | A file — new machine/teammate inherits the policy instantly |
Untapped #19 — Worktrees: Two Claudes, One Repo, No Stepping on Toes
The gap
Parallel work today = parallel risk: two sessions in one folder can collide on the same files. Git worktrees (unlocked by Lever #1) give each session its own physical directory on its own branch — same repository, zero collisions — the infrastructure for "the guide build grinds in tree A while the urgent cockpit fix lands in tree B."
The fix (the whole skill is three sentences)
Create a worktree for [task-name] on a new branch; open a session there for [the big slow task]. Meanwhile I'll keep this session on main for urgent work. When the worktree task is done: verify, merge to main, remove the worktree, and show me the merged diff summary.
Payoff math 📊
| Scenario | One folder | Worktrees |
|---|---|---|
| Urgent fix lands mid-big-build | Either interrupt the build or risk file collisions | Both proceed; merge order is a choice, not a scramble |
| Experiment that might be abandoned | Pollutes the working tree; cleanup debt | Delete the worktree — zero residue |
| + cloud sessions (#8) | — | Local tree + N cloud branches = a real parallel factory |
Untapped #20 — Headless Claude: The Glue of Full Autonomy
The gap
Scheduled Python is everywhere (Tier 1 ✅); scheduled AI is Lever #7. The bridge between them is headless mode — claude -p "instruction" running non-interactively inside any script — and, at product scale, the Agent SDK (the same machinery as a library). Today, pipelines that need judgment mid-flow still route through interactive sessions; headless would let the SCRIPT summon judgment exactly where needed and continue.
The fix — the sandwich pattern
rebuild_rotation.bat:
python export_candidates.py # deterministic: gather + filter
claude -p "Read candidates.json; apply the assignment doctrine from
CLAUDE.md; write assignments.json; list any lead you were unsure about
in review.txt with reasons" # judgment, headless, rails-bound
python apply_assignments.py # deterministic: validate + write DB
python verify_counts.py # the Evidence Law, in codeDeterministic bread, intelligent filling, verification crust — rails+engine+gate (p4) with the engine summoned programmatically.
Payoff math 📊
| Interactive-only | Headless-capable | |
|---|---|---|
| Judgment steps in pipelines | Wait for a human to open a session | Run at 3am inside the scheduled job |
| "Fully automated" claims | Asterisked | True for entire classes of work — reviewed by exception (review.txt), not by presence |
Untapped #21 — Speak Institution: .docx / .pptx / .xlsx Outputs
The gap
The house standard is interactive HTML (p78 ✅ — genuinely superior for humans reading on screens). But procurement departments, boards, and government offices ask for Word, PowerPoint, Excel — and the file-creation skills (p28) that would produce them at house-quality are installed and idle. Today an institutional ask would mean manual reformatting; that's artisan work a skill should own.
The fix
Create an "office-export" skill: given any of my HTML deliverables, produce (a) a .docx with our branding (cover, headings, tables, page numbers), (b) a 10-15 slide .pptx executive version (headline + evidence per slide, speaker notes), (c) where the deliverable has data: an .xlsx appendix with live formulas, not pasted values. House rules apply: real values only, claim labels preserved, a final render-check before "done".
Payoff math 📊
| Moment | Without | With |
|---|---|---|
| "Can you send that as a Word file for our records?" | An hour of manual reformatting, quality drift | One command, brand-perfect, minutes |
| Enterprise/government deals (Part 9 upmarket moves) | Format friction reads as small-vendor | Institutional fluency reads as established |
Untapped #22 — Interrogate Your Own Databases Weekly
The gap
The databases hold answerable gold — funnel conversion by segment, response patterns by title/geography/day, cohort aging, which gift pages convert views to replies ✅ (the events are logged) — but analysis happens ONLY when a question is asked manually. No standing analytical heartbeat. The sqlite MCP (p53 ✅) makes this a one-ask lever.
The fix
Weekly analytics run against the leads DB (read-only): (1) funnel this week vs trailing 4 (contacted → opened → replied → meeting), with the QUERY shown for every number; (2) top 3 segments by reply rate (title band × industry × geography), minimum n=30 before claiming anything; (3) one "surprise" — a pattern I probably don't know, with its evidence and a suggested experiment; (4) write analytics/[date].md and add the headline to the morning brief.
Payoff math 📊
| Decision | Today's basis | With the heartbeat |
|---|---|---|
| Where to source next 1,000 leads | Instinct + memory of what worked | Reply-rate by segment, n-checked, weekly |
| Which message variant wins | Anecdote | Measured, with the query inline (p76-native) |
Untapped #23 — Harden the Fortress: Scans, Backups, Drills
The gap
Secret HANDLING is elite (never-in-chat law, redaction pipes, registry ✅ — born from p79 #5). The untapped ring: proactive scanning (is any secret sitting in a file about to be committed by Lever #1's git?), backup encryption + restore drills (backups exist; when was a restore last REHEARSED?), and injection-awareness audits for the pipelines that ingest strangers' content (scraped pages, inbound replies — p64 rule 4 at pipeline scale).
The fix
Security pass: (1) scan the workstation for secret-shaped strings (key patterns, tokens) OUTSIDE the approved secret files — report locations only, never values; (2) add a pre-commit hook blocking commits containing them; (3) verify backups cover the DBs + state files, add encryption, and run a RESTORE DRILL to a scratch folder — prove the restore works; (4) list every pipeline that ingests external text and confirm each treats content as data (no instruction-following on scraped/replied text). Report as a table: risk / status / fix applied.
Payoff math 📊
| Event | Unhardened | Hardened |
|---|---|---|
| Secret accidentally committed/pushed (#1 + #8 make this newly possible!) | Public key = rotate everything, pray | Blocked at commit — the mistake can't complete |
| Disk death | "The backups should work…" ❓ | Restore rehearsed = RTO measured ✅ |
Untapped #24 — Arm the Team: Live Playbooks Over Static Messages
The gap
The human sales team runs on deployed cockpits + playbook pages ✅ — strong. The untapped ring: shared AI context. Updates travel as posted messages; questions flow back to the operator; the team's collective learning (what worked in replies, objection patterns) lives in scattered chats. Team-shared Projects (p29) — or even a deployed FAQ page auto-built from their questions — would let the playbook ANSWER, not just inform.
The fix (privacy-shaped)
Payoff math 📊
| Today | Enabled | |
|---|---|---|
| Routine team questions | Interrupt the operator (hours/week) | Answered by the playbook-that-talks, 24/7 |
| New member ramp | Reading + shadowing days | Interactive onboarding against the same assistant |
| Collective learning | Evaporates in chat scroll | Mined weekly into the canon |
Untapped #25 — The Stack Itself: Sequencing & The 100× Question
The 25th lever is meta: the levers MULTIPLY, and the multiplication has an optimal order. Here's the honest math and the sequence.
The tally ⚠️ (estimates, labeled — each page showed its own basis)
| Cluster | Levers | Weekly yield ⚠️ |
|---|---|---|
| Safety rails | #1 git · #2 tests · #23 security | ~1-2 hrs + catastrophic-loss insurance (the un-averageable term) |
| Zero-willpower discipline | #3 hooks · #16 hygiene · #18 allowlists | ~2-3 hrs + error-class extinction |
| Judgment on timers | #7 briefs · #20 headless · #22 analytics · #17 metering | ~3-5 hrs + decisions upgraded from instinct to evidence |
| Leverage & reach | #5 agents · #8 phone · #9 CI · #19 worktrees · #24 team | ~4-6 hrs + parallelism + continuity |
| Output & economics | #4 skills · #6 plan-gate · #10 artifacts · #12 batch · #13 QC · #14 GUI · #15 voice · #21 office | ~4-6 hrs + unit costs down, polish up |
📊 Sum: roughly 14-22 operator-hours per week ⚠️ plus the unquantifiables (a prevented disaster, a caught regression, an evidence-based pivot). Against a nights-and-weekends operating budget of ~25-30 hours, that's not "100× productivity" as arithmetic — it's something better: the difference between a person running systems and systems running themselves with a person directing. That qualitative flip is the honest meaning of 100×.
The sequence (dependencies respected)
git init.)The $0-Stack Philosophy — Own Everything, Rent Almost Nothing
Part 8 is the supporting cast: the free tools around Claude that make everything in this book run at near-zero cost. But first, the philosophy that selects them — because tool choice is strategy.
The operator's three-layer law (battle-tested)
Why "free + owned" beats "cheap + rented" (the real arguments)
| Dimension | Rented SaaS | Owned stack |
|---|---|---|
| Cost at scale | Per-seat/per-use grows with you | Flat ~$0 — scale is free |
| Failure modes | Price hikes, sunsets, region blocks, eligibility rules | Your uptime, your rules |
| Data | Lives in their cloud, export-shaped by them | Files on your disk (p74, p77) |
| AI-era kicker | You adapt to their features | Claude extends YOUR tools on demand — the build cost that justified SaaS has collapsed |
Python + SQLite — The Free Engine Room of Everything
Two boring, invincible technologies run every system in Part 6. Neither costs anything; both will outlive every startup on your phone.
Python — the language your AI speaks natively
- What it is: the world's most readable programming language — and the one LLMs write best, by a wide margin. You'll never write it; you'll READ it occasionally (p44's manager skill: spotting "that touches the wrong file").
- Why standard-library-first matters: the operator's house rule bans exotic dependencies where stdlib works. Fewer moving parts = scripts that still run in 5 years. Ask for it explicitly: "standard library only where possible."
- Install once: python.org → latest 3.x → tick "Add to PATH" (Windows). Claude Code handles the rest forever.
SQLite — a real database that is just… a file
| Fact | Why you care |
|---|---|
| One file = the whole database | Copy it = backup. Email it = migration. No server, no service, no password to leak |
| Ships inside Python (and your phone, browser, car…) | The most deployed database on Earth — boring-proven |
| Handles millions of rows easily | Your business will not outgrow it for years (1,221-row cockpit? It yawns) |
| Speaks SQL | The evidence language of p76 — every claim queryable |
The pairing pattern (you already built it)
Build-along #2 (p58) — SQLite stores → Python enforces → HTML displays — is the house architecture of everything from expense trackers to the cockpit to the diagnostics business. It has no subscription, no vendor, no expiry.
status='dead', not DELETE). Audit trails, reversibility, and the data moat (p77) all live downstream of this single discipline. Put it in every project's CLAUDE.md.Cloudflare Pages, Workers & KV — A Global Company's Infrastructure, Free Tier
Everything you deploy in this book lands here ⚠️ (limits as of writing — always check current):
The three free pillars
| Service | What it is | Free tier reality ⚠️ | Book usage |
|---|---|---|---|
| Pages | Static site hosting on a global CDN | Unlimited sites & bandwidth headline, ~500 builds/mo | Your dashboard (p57), the cockpit, this guide if published |
| Workers | Tiny programs at the edge (APIs, redirects, trackers) | ~100k requests/day | Click/open beacons, form handlers, tiny APIs |
| KV | A key-value store Workers read/write | Generous reads; ~1k writes+lists/day — list operations are the scarce resource (a real-world caught-limit lesson ✅) | Counters, event logs, feature flags |
Deploying is genuinely one sentence
Deploy this folder to Cloudflare Pages as [project-name]. First deploy: walk me through the free account + one-time login. Every future time: just deploy and give me the URL. Add "deploy" as a slash command for this project.
Worth knowing before you scale on it
- Custom domains are free to attach (the domain itself ~$10/yr at a registrar) —
yourname.comon Pages costs the domain alone. - Static-first architecture (p73's choice) means the free tier carries real businesses: pages are files; only tiny dynamic bits need Workers.
- Alternatives exist — Netlify (similar, kind free tier), GitHub Pages (simplest, public repos) — the pattern transfers; pick one and stop shopping.
Git + GitHub — The Free Safety Net (Arsenal View)
Pages 56 and 81 made the case; this page is the arsenal card — what the free tier actually gives you and the habits that extract it.
What free actually includes ⚠️
| Gift | Worth |
|---|---|
| Unlimited private repos | Every project versioned + cloud-backed, $0 |
| GitHub Actions minutes (generous free allowance) | The p89 factory — tests + deploys on every push |
| Issues / PRs / wikis | A free project-management layer your AI can read and write (p60's @claude) |
| GitHub Pages | Yet another free hosting lane for simple sites |
| Releases | Versioned downloadable snapshots — "v1.0 of the family archive" |
The three habits that matter (everything else is optional)
VS Code — The Free Window Into Everything
Visual Studio Code (free, Microsoft) is where you LOOK at what your AI is doing: files, folders, diffs, databases — one window. Non-coders: this is a file manager with superpowers, not a programming exam.
Why it earns a place in a non-coder's dock
| Power | Daily meaning |
|---|---|
| Claude Code extension | The agent in a panel BESIDE your files — watch edits land live (p45 route C) |
| Diff view | Green/red before-after of any change — the p62 "read diffs like a manager" skill, made visual |
| Whole-folder search | "Where does 180 SAR appear?" across every file, instantly |
| Markdown preview | Your NOTES.md / state files (p74) rendered readable |
| Integrated terminal | The command window, same screen — no window juggling |
The 15-minute setup
The Free-Tier Shortlist — Services That Earn Their Slot
Beyond the core stack, a small set of freemium services carry real weight ⚠️ (tiers change constantly — verify before depending; the operator's vendor law from p79 #1 applies to the PAID upgrade of every row):
| Service | Job | Free reality ⚠️ | Field note |
|---|---|---|---|
| Tally | Forms (intake, surveys, applications) | Unlimited forms/responses on free — unusually generous | The async business model's front door (p80): form in → AI processes → deliverable out |
| Resend | Transactional email from code | ~100 emails/day free; ~$20/mo for 50k/mo | Powers report delivery in the real stack ✅ — clean API, Claude wires it in minutes |
| Apify | Web scraping actors (p3's pipeline) | Small free credit; ~$29 tier unlocks real use | Respect p41 #13's ethics gate before any scrape |
| Netlify | Pages alternative | Comparable free hosting | Redundancy option; don't run two for fun (p115) |
| Google AI Studio | Free Gemini API tinkering | Meaningful free quota | Fine for experiments; this book's workflows stay Claude-first (p11's reasoning) |
| NotebookLM | Source-grounded study (Google) | Free, generous | Upload your PDFs → answers cite YOUR sources; students love it (p114) |
| Canva | Graphics beyond AI's reach today | Free tier covers most needs | Logos, social tiles; AI drafts the copy, Canva dresses it |
The selection discipline (why this list is short)
- Each row survived the p106 three-layer test: the job genuinely can't be self-forged (deliverability infrastructure, form hosting with spam handling, scraping proxies).
- Each has a working free tier, not a 14-day teaser — you can operate at $0 until revenue justifies upgrades.
- Before ANY paid upgrade: independent reviews + a sample + 2 alternatives compared — the p23 due-diligence prompt exists for exactly this moment.
Local AI — Running Models on Your Own Machine
You can run open-weight models (Llama, Mistral, Qwen families ⚠️) on your own computer, free, offline, via Ollama (ollama.com — install, ollama run llama3, done). This page is the honest "when that's smart, when it isn't."
The honest comparison
| Local (Ollama-class) | Claude (frontier cloud) | |
|---|---|---|
| Cost | $0 per token, forever | Plan/API pricing (p13) |
| Privacy | Absolute — nothing leaves the machine | Strong policies, but data transits a vendor |
| Capability | A capable junior — fine for drafts, summaries, classification | The senior colleague this book is about — reasoning, long context, agentic work |
| Speed on your laptop | Modest without a gaming GPU | Fast, and does 10 things at once |
| Agent tooling | DIY, rough edges | Claude Code's entire mature toolbelt (Part 4) |
Where local genuinely wins
- Radioactive-privacy drafts — content that must never transit any cloud (then refine the DE-identified version with Claude)
- Unlimited brainstorm/classification volume at literally zero marginal cost — bulk first-pass sorting the cloud model then reviews (a p68 economics play)
- Offline situations — flights, field sites, unreliable internet
- Learning how LLMs behave — poking a model you fully control teaches p7 viscerally
The Automation Landscape — Zapier, Make, n8n… and the In-House Answer
The daughter's generation will hear these names constantly, so here's the fair map — followed by the house position, which is unfashionable and correct for operators like this book's.
The fair map ⚠️
| Tool | What it is | Fair praise | Fair caution |
|---|---|---|---|
| Zapier | Cloud "when X happens, do Y" between 6,000+ apps | Easiest start; huge catalog | Per-task pricing climbs FAST at volume; logic beyond simple flows gets awkward |
| Make | Visual multi-step scenarios | More power per dollar than Zapier | Complexity ceiling; still rented |
| n8n | Open-source workflow engine (self-hostable) | Own-able; genuinely capable | You now babysit a workflow SERVER — maintenance masquerading as ownership |
The house position (stated as strongly as it's held)
The honest decision rule (for everyone else)
- You have Claude Code + this book → in-house Python. You hold the cheaper, stronger position; use it.
- You're advising someone who WON'T direct AI (a relative wanting one Gmail→Sheets flow, zero appetite for more) → Zapier's free tier is honestly fine. Right tool for the operator they actually are.
- Either way: anything involving money, sending to humans, or bulk actions inherits the same laws — human gates, logs, kill switches (p59, p64).
The Wider Free-AI Shelf — What Else Deserves a Bookmark
Claude-first does not mean Claude-only. The complements worth knowing, each in one honest line ⚠️ (capabilities shift monthly):
| Tool | The one thing it's for | Field note |
|---|---|---|
| NotebookLM (Google, free) | Question-answering strictly over YOUR uploaded sources — with citations | The student's second tool after Claude; the audio-overview feature turns notes into a podcast, which delights exactly everyone |
| Google AI Studio | Free API experimentation with Gemini models | For builders comparing providers (p11's ten-minute test, API edition) |
| Whisper-class transcription (many free UIs) | Audio → text, astonishingly well, most languages | Lecture/meeting recordings → p95's brain-dump pipeline |
| Image generators (varied free tiers) | The thing Claude deliberately doesn't do (p11) | AI drafts the PROMPT for them — "write me a detailed image prompt for…" — then paste across |
| Perplexity-class answer engines | Fast sourced answers as a search habit | Overlaps Claude's search (p23); fine as a phone reflex |
| Translation (DeepL etc.) | Bulk document translation | Claude handles nuance + explanation better (p40 #9); dedicated tools win on raw volume |
The integration pattern (this is the skill)
Complements slot into Claude-run workflows, not beside them: transcription FEEDS the voice pipeline (p95); image prompts are WRITTEN by Claude; NotebookLM handles source-Q&A while Claude does the thinking, building, and shipping. One brain orchestrating instruments — not five apps fighting for the same job.
The Anti-Tool Rule — Part 8's Closing Discipline
The most contrarian page in the book, and among the most valuable: every tool you add is a tax — learning, maintenance, another login, another failure mode, another place your data lives. Tool minimalism is a competitive edge disguised as boringness.
The five-question gate (before ANY new tool, free or paid)
The audit that pays annually
Here's my tool list with what each does and costs: [list]. For each: (a) used in the last 30 days? (b) overlaps another? (c) replaceable by a script in my stack? Recommend: KEEP / REPLACE-IN-HOUSE / CANCEL, with reasoning and the replacement plan where relevant. Be ruthless — defaults favor cancellation.
Before the Ten Paths — The Honest Math
This part was requested as "ways to make $10,000 a month from home." Here it is — but this book earns its trust by refusing to sell dreams, so the math page comes first and nothing after it contradicts it.
What $10K/month actually decomposes into 📊
| Model | The arithmetic | The REAL requirement hiding in it |
|---|---|---|
| High-ticket service | 4 clients × $2,500 | ~40-80 serious conversations/mo of pipeline behind those 4 closes |
| Mid-ticket productized | 20 sales × $500 | Reaching ~2,000-6,000 right-fit eyeballs/mo 📊 (at 0.3-1% conversion) |
| Retainers | 10 clients × $1,000/mo | ~12-18 months of stacking + churn-beating delivery quality |
| Volume digital | 500 sales × $20 | A genuine audience — the slowest, most compounding asset there is |
The three-stage truth about ramp ⚠️
Path 1 — The Productized Audit (The Operator's Own Play)
| What it is | A fixed-scope, fixed-price diagnostic for one niche's expensive problem: "Revenue-leak audit for SaaS" (the p72 case), "Booking-loss audit for clinics", "Quote-to-win audit for contractors". Form in → AI-powered analysis against YOUR framework → board-grade report out (p78 standard). |
| Who pays | Business owners with a bleeding metric they can't see clearly. ⚠️ $300-1,000 entry audits; $2,500-5,000 flagship depth once proof exists. |
| Why AI transforms it | Delivery collapses from consultant-weeks to a founder-reviewed evening (p72's pipeline). Margins 📊 ~90% once the framework is built. The framework IS the business. |
| Build list | 18-question intake form (Tally, p111) · analysis framework (25 checks that find the money — YOUR expertise encoded) · report generator (p58 pattern + p78 polish) · one landing page (p57) · Stripe payment link. |
First-client path (the unglamorous truth)
Path 2 — Deliverables Freelancing (Fastest First Dollar)
| What it is | Selling finished ARTIFACTS on Upwork/Fiverr/direct: pitch decks, financial models, business plans, data cleanups, interactive reports, spreadsheets with working formulas (p28) — the things businesses need weekly and hate making. |
| Who pays | Small businesses & startups drowning in "we need a deck by Thursday". ⚠️ $50-150 entry gigs → $300-800 once reviewed → $1,000+ for complex models. |
| Why AI transforms it | You deliver in HOURS what competitors quote in days — at higher polish (p78's standard as your default is genuinely rare on these platforms). Speed + quality + reliability is the whole game there. |
| The edge stack | Part 3 prompting · p28 office files · p37's critique loop before every delivery · p101's dual-format trick (interactive HTML + the .docx "for records") as a signature differentiator. |
The weekend launch plan
Path 3 — Websites for Local Businesses
| What it is | The p57 build-along, sold: clean sites + WhatsApp-first contact + Google Maps presence for salons, clinics, restaurants, tradesmen — businesses whose web presence is a dead Facebook page. |
| Who pays | ⚠️ $300-800/site in most markets ($500-1,500 in Gulf/Western markets) + the real prize: $30-80/mo care plans (updates, backups, tweaks) — recurring revenue that stacks (p116 row 3). |
| Why AI transforms it | Build time: 3-6 hours including revisions (p57 proved it to you personally). Hosting: $0 (p108). Your cost is conversation and taste. |
| The stack | p57 skills + a reusable "local business" template you refine each build + p93's automated QC sweep across ALL client sites weekly (one command maintains the whole portfolio). |
The acquisition play (this path lives or dies here)
Path 4 — Executive Ghostwriting & Content Engines
| What it is | Running the content presence of people with expertise but no time: LinkedIn for executives/founders, newsletters for firms, thought-leadership series. YOU interview + edit + strategize; AI drafts at scale in their captured voice (p22, p32). |
| Who pays | ⚠️ Executives: $500-2,000/mo for 8-12 LinkedIn posts + engagement strategy. Firms: $800-2,500/mo newsletters. Consultants selling with visibility pay happily — content IS their pipeline. |
| Why AI transforms it | The bottleneck was always drafting volume in a consistent voice. Solved (p32's few-shot voice-cloning + p37's loop). What remains — strategy, interviewing, taste, relationships — is exactly the human layer clients are really buying. |
| The workflow that scales | Monthly 45-min voice interview per client (p25/p95 pipeline!) → 12 drafts in their voice → they approve on a simple page (a mini-cockpit, p58!) → scheduled posting. One evening per client per month once systematized 📊. |
Landing the first two
Path 5 — Career Services: The Cockpit as a Business
| What it is | Exactly what p73 documented, sold as a service ladder: CV rewrites → application kits → interview prep → the full private job-search cockpit for senior professionals (tailored targets, personalized outreach, follow-up rails, a live dashboard). |
| Who pays | ⚠️ CV+letter packages: $80-250. Interview-prep sprints: $150-400. The flagship cockpit: $800-2,500 for executives — because a director's ONE month of unemployment costs more than the fee (that anchor sells itself). |
| Why AI transforms it | You've SEEN it: 1,221 researched targets, role-relative positioning (p73's empathy-as-code), QC-gated messages, zero-training dashboard — a solo operator delivering what agencies charge 10× for. Every technique is in this book. |
| The ladder logic | Entry products fund trust; the cockpit is the flagship 20% of clients buy. Each tier's deliverable is a SYSTEM you refine, not hours you sell (p117's productization DNA). |
Why this path suits first-timers unusually well
- Demand is ambient: everyone knows a job-seeker THIS MONTH. Your first 5 clients are one family WhatsApp announcement away — no cold anything.
- Results are visible fast: interviews land in weeks → testimonials with faces → referral physics.
- The empathy moat: p73's lesson — positioning candidates relative to each recipient, truthfully, kindly — is a QUALITY bar most competitors (spray-and-pray CV mills) never touch. Care, encoded, is the differentiation.
Path 6 — Lead Generation as a Service (Highest Skill Bar)
| What it is | Selling the p3 pipeline to agencies/B2B firms: researched, verified, personalized lead lists + outreach sequences — or running the whole motion as a retainer. ⚠️ $500-2,000/mo retainers; $1-3 per verified enriched lead on lists. |
| Why it's placed 6th, not 1st | It LOOKS like the obvious AI business; it carries the most ways to hurt yourself: deliverability is a craft (warm-up, caps, domain hygiene — the real stack runs 16 sending domains and 18/day caps for reasons ✅), data laws are real (GDPR-class), and bad execution burns CLIENTS' domains — reputation damage you can't refund. |
| Who should take it | Someone who's ALREADY run outreach for themselves (paths 1/3/5 teach it safely on your own account first) and internalized this book's ethics pages (p3's legitimacy box). |
The offer ladder (risk-ordered)
Path 7 — Templates & Digital Products (The Volume Game)
| What it is | Selling systems-in-a-box: spreadsheet suites with working formulas (p28), Notion/planner systems, prompt libraries (your p38 portfolio, packaged!), CV template kits, small-business toolkits, interactive HTML tools (p78's standard as product). ⚠️ $10-70 typical price points on Gumroad-class platforms (free to start, ~10% fees). |
| The honest physics | 📊 $10K/mo = ~300-500 sales at $25 — a VOLUME game won by audience + catalog, not by one product. This is the slowest path to money and the best at compounding: products sell while you sleep, forever, and each one markets the others. |
| Why AI transforms it | Catalog velocity: a polished product per weekend instead of per quarter. Ten products in ten weeks is a genuine catalog; pre-AI that was a year's work. |
The playbook that actually works
Path 8 — Courses & Paid Guides (Teaching as an Asset)
| What it is | Packaged teaching: interactive guides (the ARTIFACT YOU ARE READING is the demo — p78's standard as a sellable product class), cohort workshops, video courses, "build-along" programs teaching exactly what Parts 2-4 taught you. ⚠️ Guides $20-100; self-paced courses $50-300; cohorts $200-800. |
| The prerequisite nobody skips | You can only teach what you've DONE. This path activates 3-6 months AFTER another path gives you real results and real student-questions. Teaching-before-doing is the info-market's credibility graveyard — and buyers' radar for it is excellent now. |
| Why AI transforms it | Production collapses: this 130-page interactive book was built in ONE session. Curriculum design, exercise generation, quiz banks, per-student feedback rubrics — all directable (Part 3). What can't be automated: your scars, your specificity, your voice. Those are the product. |
The format ladder (start small, climb with proof)
Path 9 — Back-Office Automation for Small Businesses
| What it is | Selling OPERATIONS relief to trades, clinics, shops, agencies: invoice generation + chasing, quote builders, appointment reminders, report packs, document pipelines, review-request flows — the p58/p59 machinery pointed at THEIR paperwork. ⚠️ Setup projects $300-1,500 + $100-500/mo "it keeps running" retainers. |
| Why it's a sleeper | Zero glamour, MAXIMUM stickiness: an automation that chases invoices is removed from a business over the owner's dead body. Churn approaches zero; retainers stack (p116 row 3). Boring = moat. |
| Who pays fastest | Businesses with 3-15 staff — big enough to bleed hours on admin, small enough to have no IT department. Your city has hundreds. |
The engagement pattern (audit → quick win → retainer)
Path 10 — Data Services (Where Verification Culture IS the Product)
| What it is | Unsexy, evergreen, corporate-budgeted: cleaning messy datasets (CRMs full of duplicates, product catalogs, member lists), enrichment (filling missing fields from public sources), structured research (competitor matrices, market maps, vendor comparisons). ⚠️ Cleanup projects $200-2,000; research $150-800/deliverable; enrichment priced per-record. |
| Why AI transforms it | p24's ladder + p34's schemas + p54's parallel research at industrial scale. What took a data-entry team a week takes an evening — but ONLY the p76-disciplined survive, because… |
| The differentiator | …everyone with a chatbot now SELLS "research". Deliverables full of hallucinated rows are flooding the market and burning buyers. Your Evidence Law workflow — every row sourced, every claim labeled ✅⚠️❓, a methodology sheet attached — is the product. You're not selling data; you're selling data someone can bet a decision on. |
The trust-manufacturing playbook
The 30-Day Plan — Zero to Hero, Day by Day
Everything in this book, sequenced into ~45-60 minutes a day. Tick pages off in the sidebar as you go — the progress bar is your streak.
Week 1 — Foundations & fluency (Parts 0-2)
| Day | Do |
|---|---|
| 1 | p1-6: the four answers + your 10-minute quick start. Account created, first Artifact built. |
| 2 | p7-11: what the machine is. Run the p11 try-them-all test on one real task. |
| 3 | p12-16: surfaces, pricing, the five mental models. Write YOUR one-line verification policy. |
| 4 | p17-19: the 30-second brief + files + Artifacts. Upload a real PDF; build one real tool. |
| 5 | p20-22: create your first Project with real instructions + 3 knowledge files. |
| 6 | p23-26: research with sources + analyze a real spreadsheet + connect Calendar. |
| 7 | p27-30: Chrome sidebar + one office file + fire 3 prompts from the life catalog. Rest. |
Week 2 — Prompting mastery (Part 3)
| Day | Do |
|---|---|
| 8-9 | p31-34: formula, few-shot, thinking, structure. Rewrite your 3 most-used asks properly. |
| 10-11 | p35-38: long docs, chains, the critique loop, meta-prompting. Commission your first prompt. |
| 12-13 | p39-41: fire 10 prompts from the libraries on real life. Start YOUR library note. |
| 14 | p42: self-diagnose against the 12 mistakes. Rest. |
Week 3 — Hands (Part 4)
| Day | Do |
|---|---|
| 15 | p43-45: install Claude Code (any route). |
| 16 | p46-48: the first-session walkthrough, tools, permission modes. |
| 17 | p49-50: write your CLAUDE.md + first slash command. (The compounding starts TODAY.) |
| 18-19 | p57: Build-along #1 — dashboard built AND deployed. Send the link to family. |
| 20 | p56: git on everything you've made. Practice one undo. |
| 21 | p58: Build-along #2 — your first database system. Rest after; you've earned it. |
Week 4 — Systems & the road (Parts 5-9)
| Day | Do |
|---|---|
| 22 | p59: automate ONE thing on a schedule. p52: one hook. |
| 23 | p63-66: MCP mental model + connect one server (SQLite on your build-along DB). |
| 24-25 | p71-80: the case studies. Steal: state files (p74) + 5 starter laws (p75) + Evidence labels (p76). |
| 26 | p81-105 skim: mark YOUR five untapped levers; schedule lever #1 for next week. |
| 27 | p106-115: audit your tools with the p115 gate. |
| 28-29 | p116-126: pick ONE income path by unfair advantage. Do its first step (the free-for-three, the demo site, the portfolio sprint). |
| 30 | Write your own "Part 7": your gaps, your sequence, your next 30 days. You're not a beginner anymore — you're an operator with a roadmap. |
Glossary — Plain Words for Every Term
| Term | Plain meaning |
|---|---|
| LLM | Large Language Model — the text-prediction brain behind AI chat (p7) |
| Token | The ~4-character chunks text is processed in; pricing & limits count these (p7) |
| Context window | The AI's working desk — everything it can see right now (p8) |
| Prompt | Your instruction/brief to the AI (p5, p31) |
| Few-shot | Teaching by pasting 2-3 examples of what you want (p32) |
| Hallucination | Confident wrong answers, reconstructed rather than known (p14) |
| Extended thinking | The model reasoning internally before answering (p33) |
| Artifact | A live document/app the AI builds in a side panel (p19) |
| Project | A workspace with standing instructions + knowledge files (p20) |
| Connector / MCP | The standard plug wiring AI to tools — Gmail, databases, browsers (p26, p63) |
| MCP server | The translator program exposing one tool's safe levers to the AI (p63, p69) |
| Claude Code | Claude with hands: creates files, runs commands, builds systems (Part 4) |
| Agent / agentic | AI that takes multi-step actions with tools, not just answers (p43) |
| Subagent | A worker-Claude dispatched by your session for a scoped task (p54) |
| CLAUDE.md | The rules file auto-read every session — your AI's constitution (p49) |
| Slash command | A saved prompt fired by typing /name (p50) |
| Skill | A packaged expertise folder the AI auto-loads when relevant (p51) |
| Hook | A command that fires automatically on session events — rules as machinery (p52) |
| Plan mode | Read-only research → written plan → your approval → execution (p55) |
| Git / commit | The folder time machine / one saved snapshot in it (p56) |
| Repo / GitHub | A git-tracked project / the cloud home for repos (p56, p109) |
| PR (pull request) | A proposed change packaged for review before merging (p60) |
| CI (GitHub Actions) | Checks & deploys that run automatically on every change (p89) |
| Worktree | A second physical copy of a repo for parallel work (p99) |
| Headless | The AI invoked inside a script, no chat window (p100) |
| API | Calling the model from your own program, pay-per-token (p67) |
| Batch API | Patient bulk jobs at −50% (p68) |
| Prompt caching | Reusing stable context at up to −90% cost (p68) |
| System prompt | The standing role/rules sent with every API call (p67) |
| SQLite | A real database that's just one file (p107) |
| Static site | Pre-built pages served as files — free, fast, unhackable-ish (p108) |
| Deploy | Putting files on the internet (p57, p108) |
| Prompt injection | Hostile instructions hidden in content the AI reads (p26, p64) |
| Rails + engine + gate | Code decides what MAY happen; AI generates; a check verifies — this book's core pattern (p4) |
| Evidence Law | No claim without its query; labels ✅📊⚠️❓ on every number (p76) |
| Data moat | Permanently accumulated anonymized intelligence — the compounding asset (p77) |
| Productized service | Fixed scope, fixed price, systematized delivery — sold like a product (p117) |
FAQ — Short Answers to the Questions Everyone Asks
The original four, now in one breath each
| Question | The one-breath answer |
|---|---|
| 1. Which AI made the cockpit? | Claude (Anthropic) — as Claude Code, with Python/SQLite/Cloudflare doing the non-AI jobs (p2). |
| 2. Where did the emails come from? | Public LinkedIn profiles → pattern-inferred addresses → machine-verified before any use — ethically fenced (p3). |
| 3. Is the intent predefined? | Strategy predefined in code rails; wording AI-generated inside them; a QC gate checks everything (p4). |
| 4. How do you prompt? | Brief, don't wish: Role + Context + Task + Format + Constraints, then iterate (p5, p31). |
The ten that always follow
Is my data safe with AI?
Work products: reasonable to share, check settings. Access (passwords/keys) and others' private data: never (p15). Business-critical: own your files and memory (p74).
Will AI take my job?
Honest answer: it's taking TASKS everywhere, whole jobs unevenly. The defensible position is being the person who DIRECTS it in your field — which is precisely the skill this book teaches (p71). The worst position is pretending it's a fad.
Can my kids use it? Isn't it cheating?
Use it as a TUTOR (the p40 Socratic pattern — it coaches, never solves), it's the best study tool ever built. Use it to ghostwrite homework, it's cheating and it hollows the learner. The prompt decides which — parents set the prompt culture.
How much does it really cost to start?
$0 (free tiers of Claude + every tool in Part 8). First worthwhile spend: ~$20/mo Pro when daily limits bite (p13). Everything in this book's builds runs on free infrastructure.
Does it work in Urdu / other languages?
Yes — strong Urdu, Arabic, and 100+ languages; mixing mid-sentence is fine (p25). Output quality is best in English for technical work; explanations translate beautifully.
What if it gives me a wrong answer?
It will, sometimes — plan for it rather than being surprised by it: high-risk zones (p14), one cheap verification, labels on numbers (p76). Trust the workflow, audit the claims.
Which is better — ChatGPT or Claude?
For serious work, systems, and code: this book's answer is Claude, with reasons (p11). For images or casual breadth: others shine. Run the ten-minute test yourself — evidence over tribalism.
Do I need to learn coding first?
No — directing beats writing (p44). You need: clear outcome descriptions, verification habits, small-step discipline. Reading a little code eventually helps you manage better; writing it stays optional.
Can it work offline?
Claude: no — it's a cloud service. Local models (p112) run offline at junior-level quality. Design accordingly: drafts offline in a pinch, real work online.
Where do I start if I only have 10 minutes?
Page 6. Today. The account, three prompts, one file, one Artifact. Everything else is elaboration.
Your Next Move
The fork — choose by who you are today
| You | Go to | Do this week |
|---|---|---|
| The student (who asked the four questions) | p6 → p40 | Quick start + the study prompts. Then build ONE quiz Artifact for your next exam (p19) — feel the moment it works. |
| The professional | p31 → p20 | Rewrite your 3 daily asks with the formula; build one Project for your real work. Wednesday-you will thank Monday-you. |
| The builder | p45 → p57 | Install Claude Code; ship the dashboard; send the URL to someone who'll say "wait, YOU made this?" |
| The earner | p116 → your path | Pick by unfair advantage; execute its first step (the free-for-three, the demo sites, the samples). Proof this week, not plans. |
A closing letter from the desk where this was built
This book began with a student's four questions about a job-search system built for her father's colleague — a gift, made in evenings, by a planning engineer who does not write code.
That detail is the whole message. The person behind these systems started where you are now: curious, non-technical, busy, responsible for a family, working nights after a demanding job. The difference between that starting point and the systems you toured in Part 6 is not talent, funding, or a computer-science degree. It is a sequence of small, verified steps — the same sequence now numbered 1 to 130 in your sidebar — plus rules written down when things went wrong, and the stubbornness to keep going through the stage where nothing pays yet (p116 told you that truth straight).
The machine you've just learned to direct is the most leverage ordinary people have ever been handed. Most will use it to write emails slightly faster. A few will use it the way this book taught: rails, engines, gates, evidence, compounding — and quietly build things their own families won't believe they made.
Be one of the few. Start smaller than feels impressive. Verify more than feels necessary. Write down every lesson. And when someone eventually looks at what you've built and asks "which AI tool made this?" — you'll smile, because you'll know the honest answer:
The tool was Claude. The builder was you.