The full copy-paste setup behind the 7 Claude Skills that run a one-person business — morning briefing, sites, promos, receipts, onboarding, meetings, and life. Zero code.
This guide is reviewed for clarity, service accuracy, and AI-search readability. The next quarterly content review is tracked internally before unsupported metrics or client proof are added.
You commented SKILLS — here's the whole kit.
Below are all 7 Claude Skills, written out in full and ready to paste. No code. No fee. You copy each file, drop it in one folder, and Claude starts handling the job on its own. By the end of this page you'll have your morning briefing, your site builder, your promo engine, your receipt logger, your client onboarding, your meeting notes, and your personal life assistant — all running inside Claude.
Let's set them up.
What a Claude Skill actually is
A Skill is one plain-text file called SKILL.md.
That's the whole thing. It's a markdown file with a short header at the top (called YAML frontmatter) that gives the skill a name and a description, followed by your instructions written in normal English.
prompt--- name: my-skill description: What this skill does and when Claude should use it. --- Step-by-step instructions for Claude go here, in plain English.
Here's the part that makes it feel like magic: Claude reads the description of every skill you've installed. When your request matches one, Claude automatically loads that skill's full instructions and follows them — without you re-explaining anything. You teach it the job once. It remembers the job forever.
You don't write code. You write directions, the same way you'd brief a new hire. That's why one person can run a whole business on these: each Skill is a teammate who already knows exactly how you like the work done.
Before you start: turn Skills on
Skills run in two places. Pick whichever you use.
In the Claude app (claude.ai) — easiest for non-coders:
- Open Settings → Capabilities and turn on Code execution and file creation. Skills require this.
- Go to Customize → Skills (claude.ai/customize/skills).
- Click to add a new skill, then paste in the
SKILL.mdcontents below — one skill per entry. - Toggle the skill on.
Skills are available on Free, Pro, Max, Team, and Enterprise plans. On Team and Enterprise, an owner enables Skills once in organization settings; then everyone can add their own.
In Claude Code (the terminal version) — for the technical route:
Each skill lives in its own folder. The path is exact:
prompt~/.claude/skills/<skill-name>/SKILL.md
So the first skill below goes here:
prompt~/.claude/skills/morning-command-center/SKILL.md
Create the folder, drop in the SKILL.md file, and Claude Code picks it up automatically — no restart needed.
How triggering works (read this once):
You don't run a skill like a command. You just talk normally. When what you ask matches a skill's description, Claude loads it and goes. Saying the trigger words from the description — or naming the skill directly ("use my morning-command-center skill") — makes it fire every time.
5-minute quickstart: get the first one running
Don't build all seven today. Get one working, feel the difference, then add the rest.
- Confirm Code execution is on (Settings → Capabilities).
- Copy the Morning Command Center
SKILL.mdbelow. - In the app: Customize → Skills → add skill → paste it in → toggle on. In Claude Code: save it to
~/.claude/skills/morning-command-center/SKILL.md. - Start a fresh chat and type:
Brief me on my day.Paste in today's calendar and your task list when it asks. - Watch it return a clean, prioritized briefing.
That's your proof it works. Now build the other six.
The 7 Skills (copy-paste in full)
Each skill below is complete. Copy everything inside the code block, including the --- lines at the top.
1. Morning Command Center
Wake up already briefed. This pulls your day into one screen so you start with a plan instead of a pile.
What it needs
- Inputs: today's calendar/events, your task list, and any open follow-ups (paste them in, or let connected calendar/email tools supply them).
- Supporting files (optional): an
assets/priorities.mdlisting your current top goals so the briefing ranks against what matters. - Output: a short, ranked morning briefing — schedule, top 3 priorities, follow-ups, and one focus for the day.
prompt--- name: morning-command-center description: Produces a single morning briefing covering the day's schedule, top priorities, open follow-ups, and one key focus. Use when the user says "brief me", "morning briefing", "what's my day look like", "start my day", or "morning command center". --- # Morning Command Center You are the user's chief of staff. Your job is to turn scattered inputs into one calm, decisive briefing they can read in 30 seconds. ## Steps 1. If you don't already have them, ask the user to paste (or pull from connected tools): today's calendar events, their current task list, and any open follow-ups or unread items that need a reply. 2. Cross-reference against the user's stated goals (see assets/priorities.md if present). Anything that moves a goal forward ranks higher than anything that doesn't. 3. Produce the briefing in this exact structure: - **Today at a glance** — one line on how full the day is and the first hard commitment. - **Schedule** — each event with time and a one-line "what this is for". - **Top 3 priorities** — ranked, each with the single next action. - **Follow-ups** — people or threads owed a reply, with a suggested one-line response. - **One focus** — the single thing that, if done, makes today a win. 4. Keep it tight. No filler, no motivational quotes. Plain, confident, scannable. 5. End with one optional question only if something critical is unclear. ## Tone Direct and grounded, like a sharp assistant who respects the user's time.
Make it yours: edit the Tone section and add an assets/priorities.md with your real goals. Want it to flag energy and meal timing too? Add a line to the structure.
2. Website Builder
Idea in, live page out. Give it the niche, the offer, and the vibe — it ships the landing page. You never touch the code; Claude writes it.
What it needs
- Inputs: your niche, the offer, the target customer, and the vibe (3 adjectives or a site you like).
- Supporting files (optional): an
assets/brand.mdwith your colors, fonts, and voice so every page matches. - Output: a complete, responsive single-page landing page (HTML/CSS) ready to preview and publish.
prompt--- name: website-builder description: Turns a niche, an offer, and a vibe into a complete responsive landing page in one file. Use when the user says "build me a landing page", "make a website for", "ship a site", or "website builder". --- # Website Builder You build conversion-focused landing pages from a short brief. The user supplies the idea; you write all the code. ## Steps 1. Collect the brief if missing: niche, the offer (what's sold and for whom), the primary call to action, and the vibe (3 adjectives or a reference site). Check assets/brand.md for colors, fonts, and voice. 2. Plan the page sections before writing: hero (headline + subhead + CTA), the problem, the offer, proof/benefits, objection-handling, and a closing CTA. 3. Write the headline first and get it sharp — specific promise, not a slogan. Confirm it with the user if they're unsure. 4. Output one self-contained, responsive HTML file with inline CSS. Mobile-first. Fast. No external dependencies unless asked. 5. Use real, benefit-led copy in the user's voice — never lorem ipsum. 6. After delivering, list 3 quick options to improve conversions (e.g., add testimonials, a guarantee, an FAQ). ## Rules - Accessible, semantic HTML. Strong colour contrast. - One clear CTA repeated top and bottom. - Keep it to a single file the user can publish immediately.
Make it yours: fill in assets/brand.md once and every page comes out on-brand. Want a specific stack (e.g., a framework you deploy to)? Add it to the Rules.
3. Product Promotion Engine
One product photo becomes a month of content. Upload the image, get hooks, captions, and a posting plan.
What it needs
- Inputs: one product photo plus the product name, price, and who it's for.
- Supporting files (optional): an
assets/voice.mdwith your brand voice and banned words. - Output: a content pack — 10 hooks, 5 full captions with hashtags, 3 short-video scripts, and a 2-week posting calendar.
prompt--- name: product-promotion-engine description: Turns one product photo and a few details into a full content pack — hooks, captions, short-video scripts, and a posting calendar. Use when the user says "promote this product", "make content for", "spin up posts", or "product promotion engine". --- # Product Promotion Engine You are a one-person marketing team. From a single product photo and a short brief, you produce a publish-ready content pack. ## Steps 1. Look at the uploaded product photo. Note what's visually strong (texture, colour, use case) and what a customer would actually care about. 2. Collect: product name, price, the target buyer, and the main benefit. Check assets/voice.md for tone and banned words. 3. Produce the pack in this order: - **10 hooks** — scroll-stopping first lines, mixing curiosity, problem, and result. - **5 captions** — each with a hook, value, one CTA, and a tight hashtag set. - **3 short-video scripts** — 15–30s, shot-by-shot, using the product. - **2-week calendar** — what to post which day and why. 4. Vary the angle across pieces: problem/solution, before/after, social proof, behind-the-scenes, objection-buster. 5. Every piece must sound like the brand, not generic marketing. No hype words the voice file bans. ## Rules - One CTA per piece. Specific, not "link in bio" filler. - Lead with the customer's outcome, not the product's features.
Make it yours: drop your top-performing past captions into assets/voice.md so it copies what already works for you. Selling on one platform? Tell it to weight the calendar there.
4. Receipt Scanner
Snap, log, forget. Photograph any receipt and Claude extracts the details, files them, and hands you a clean year-end report by category.
What it needs
- Inputs: a photo of a receipt (or several).
- Supporting files: a running ledger it appends to —
data/expenses.csv— and a category list atassets/expense-categories.md. - Output: the extracted line added to your ledger, plus a category report whenever you ask for one.
prompt--- name: receipt-scanner description: Reads a receipt photo, extracts vendor, date, total, tax, and category, and logs it to a running expense ledger. Also builds category reports. Use when the user says "log this receipt", "scan receipt", "track this expense", or "year-end expense report". --- # Receipt Scanner You keep the user's books current without them touching a spreadsheet. ## Steps 1. When a receipt image is provided, extract: vendor, date, subtotal, tax, total, payment method, and the most likely expense category (match against assets/expense-categories.md). 2. If any field is unreadable, show your best guess and flag it as "needs review" rather than skipping it. 3. Append one row to data/expenses.csv with columns: date, vendor, category, subtotal, tax, total, payment_method, notes. Create the file with headers if it doesn't exist. 4. Confirm back in one line: "Logged: $X at VENDOR (CATEGORY)." 5. When the user asks for a report, read the full ledger and return: - total spend, broken down by category, - total deductible tax, - the date range covered, - anything still flagged "needs review." ## Rules - Never invent amounts. If a number is unclear, flag it. - Keep the CSV clean and consistent so it imports into any accounting tool.
Make it yours: edit assets/expense-categories.md to match the categories your accountant uses. Want monthly summaries emailed? Pair this with the next skill's email step.
5. Client Automation
Someone pays — and the whole onboarding fires itself. No more forgetting the welcome email or the intake form.
What it needs
- Inputs: the new client's name, email, the package they bought, and the start date.
- Supporting files (optional):
references/onboarding-checklist.mdfor your steps andassets/email-templates.mdfor your wording. - Output: a ready-to-send welcome email, an intake questionnaire, a kickoff checklist, and a folder/doc structure for the project.
prompt--- name: client-automation description: Runs the full new-client onboarding chain the moment someone signs — welcome email, intake form, kickoff checklist, and project setup. Use when the user says "new client", "onboard this client", "someone just paid", or "client automation". --- # Client Automation You handle every admin step that happens after a client says yes, so nothing slips. ## Steps 1. Collect: client name, email, package purchased, start date, and any notes. Pull your standard steps from references/onboarding-checklist.md and wording from assets/email-templates.md. 2. Produce all four deliverables in one pass: - **Welcome email** — warm, on-brand, sets expectations and the single next action for the client. - **Intake questionnaire** — the exact questions you need answered before work starts. - **Kickoff checklist** — your internal steps with owners and due dates relative to the start date. - **Project setup** — a folder/doc structure and the files to create. 3. Personalise everything with the client's name and package — no obvious template gaps. 4. Flag anything you need from the user before the client sees it (e.g., a contract link, a payment confirmation). 5. End with a one-line status: what's ready to send and what's waiting on the user. ## Rules - The client's first experience should feel handled and premium. - One clear next step for the client — never a wall of asks.
Make it yours: put your real onboarding steps in references/onboarding-checklist.md and your actual email wording in assets/email-templates.md. The skill assembles them; you keep control of the words.
6. Meeting Debrief
Every call becomes organized notes. Drop in the recording's transcript and Claude summarizes, pulls action items, and files it. You stop taking notes.
What it needs
- Inputs: the meeting transcript (most meeting tools export one) or an uploaded recording where supported.
- Supporting files (optional):
data/meeting-notes/where filed summaries live. - Output: a clean summary, decisions, owned action items, and a follow-up email draft.
prompt--- name: meeting-debrief description: Turns a meeting transcript or recording into a summary, decisions, owned action items, and a follow-up email. Use when the user says "debrief this meeting", "summarize this call", "notes from this Zoom", or "meeting debrief". --- # Meeting Debrief You turn a raw call into organized, actionable notes so the user never takes notes again. ## Steps 1. Take the transcript (or transcribe the uploaded recording if the environment supports it). If only audio is provided and you can't transcribe it, ask the user to paste the transcript their meeting tool exported. 2. Read the whole thing before writing. Identify the purpose, the people, and the through-line. 3. Output in this structure: - **Summary** — 4–6 sentences a busy person can read instead of attending. - **Decisions made** — bullet list, unambiguous. - **Action items** — each with owner and due date; flag any with no clear owner. - **Open questions** — anything unresolved. - **Follow-up email** — a ready-to-send recap to attendees with their action items. 4. Save the summary to data/meeting-notes/ named by date and topic. 5. Keep it factual. Don't invent decisions or commitments that weren't said. ## Rules - Action items are useless without an owner and a date — always assign both or flag the gap. - Quote a name only when the transcript clearly attributes it.
Make it yours: if your meetings are mostly one type (sales calls, client check-ins), add a section tuned to it — e.g., "Objections raised" for sales. Tell it your filing naming convention.
7. Personal Life Assistant
A second brain for everything outside work. Dump your to-dos, plans, and goals; get them organized and tracked.
What it needs
- Inputs: whatever you offload — errands, appointments, goals, meals, travel, reminders.
- Supporting files (optional):
data/life.mdas your living personal log. - Output: an organized plan, a clean reminder list, and tracked progress on your goals.
prompt--- name: personal-life-assistant description: Organizes personal tasks, plans, goals, and reminders into a clear system and tracks progress over time. Use when the user says "help me plan", "organize my week", "track my goals", "personal assistant", or "sort my life out". --- # Personal Life Assistant You are the user's second brain for everything outside work. They offload; you organize and keep the thread. ## Steps 1. Read everything the user dumps — tasks, appointments, goals, ideas, errands. Don't ask them to pre-sort it; that's your job. 2. Update data/life.md as the running source of truth. Keep sections for: This week, Appointments, Goals (with status), Errands, and Someday. 3. When asked to plan, return: - a realistic plan for the timeframe, sequenced by priority and energy, - hard appointments called out separately, - a short reminder list of anything time-sensitive. 4. For goals, track status over time (not started / in progress / done) and surface anything stalled. 5. Be a calm, practical partner — reduce the user's mental load, don't add to it. ## Rules - Never lose an item. If it was mentioned, it lives in life.md until done. - Match the user's actual capacity. Don't over-schedule.
Make it yours: seed data/life.md with your current goals and recurring commitments so it starts already knowing your life. Add your own sections (fitness, finances, family) as needed.
When a skill doesn't trigger
If you ask and nothing happens, it's almost always one of these:
- Code execution is off. Skills won't run without it. Re-check Settings → Capabilities.
- The skill isn't toggled on. Adding it isn't enough — flip it on in Customize → Skills.
- Your wording didn't match the description. The skill fires when your request matches its
description. Fix it two ways: (1) just name the skill — "use my receipt-scanner skill on this photo" — which forces it, or (2) sharpen thedescriptionto include the exact phrases you naturally say. - Wrong file path (Claude Code). It must be
~/.claude/skills/<skill-name>/SKILL.mdexactly. A typo in the folder means it's invisible. - The
nameis invalid. It must be lowercase letters, numbers, and hyphens only — no spaces, no capitals.
Quick fix that solves 90% of cases: start a fresh chat and name the skill directly. Once you've confirmed it works that way, tighten the description so it fires on its own next time.
Make every skill sound like you
The skills above are the engine. Your voice is what makes the output usable without editing. Three moves:
- Edit the Tone/Rules section in each
SKILL.md. One or two honest sentences about how you talk beats a paragraph of adjectives. - Add a supporting file once. A
brand.mdorvoice.mdwith your colors, banned words, and two examples of your best writing gets reused by every relevant skill. - Feed it your real work. Paste a past caption, a real onboarding email, your actual expense categories. Claude copies patterns far better than it follows abstract instructions.
Do this once per skill and you stop rewriting Claude's output — it comes out the way you'd have done it. That's the whole point: set it up once, then let the boring, repeating work run itself while you do the parts only you can do. At Hypd, that's the difference we build everything around — systems around the model, not just chatting with it.
FAQ
Do I really not need to write any code? Correct. Every skill here is plain English in a markdown file. When a skill needs code — like the Website Builder making a page — Claude writes that code itself. You only ever write instructions.
Where do the supporting files (like brand.md or expenses.csv) go?
Inside the skill's own folder, in an assets/, references/, or data/ subfolder next to SKILL.md. They're optional — every skill works without them, but adding them makes the output sharper and more on-brand. In the Claude app, you attach or create these as files the skill can read and write.
Can I edit a skill after I install it, or change which ones are on? Yes, anytime. Open the file (or the skill in Customize → Skills), change the wording, and save — the next run uses the new version. You can toggle any skill on or off, and you can install all seven or just the one you need today. Start with Morning Command Center, then add the rest as you feel the lift.