demo-builder
Skill15 starsChanged 6 days agoMITView on GitHub
Demo Builder. Builds a demo showing how your product solves the problem, before the product exists. Turns a small spec into one self-contained HTML page that runs anywhere with no server: sample records, one button, every row moving through its statuses and a total counting up, with a banner that always says it is a demo with sample data. Claude can then make it look and behave more like the real thing for this buyer. Use when the user says "build a demo", "mock this up", "show how it would work", "I need something to show them", "prototype", or "make a clickable demo".
3 other projects use a version of this.
Adopted by
Most recently taken up by
- SVerITG/Metistheir own version, first seen Oct 8, 2026
- SVerITG/Metis_PHtheir own version, first seen Aug 25, 2026
- apetcu/skillstheir own version, first seen Apr 10, 2026
Other versions3 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 3
- SVerITG/Metis★ 4same name
- apetcu/skills★ 2same name
- SVerITG/Metis_PH★ 1same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository10 other files it gives its agents.
- case-studySkill
- follow-upSkill
- launch-plannerSkill
- market-researcherSkill
- objection-handlerSkill
- offer-builderSkill
- pre-saleSkill
- pricingSkill
- requirementsSkill
- whale-hunterSkill
The file
We split this skill into the rules it gives an agent
6 rules. 0 are common, 6 are distinctive, and 0 are very specific to this project.
0
Common
Most of the 3 edited versions we read keep these.
6
Distinctive
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 41
Find the moment: Ask: what is the one screen where the problem goes away?
- Distinctive: line 49
Keep it small: a demo that takes a week to build has stopped saving time.
- Distinctive: line 68
Never removes the banner. A demo that passes for a shipped product is a false promise, and it comes back as a refund.
- Distinctive: line 69
Never puts a real customer's records in a demo.
- Distinctive: line 70
Never promises a feature in a demo that the user doesn't intend to build.
- Distinctive: line 76
Never invent a result, a client or a review.
“Common” is measured: kept by at least 6 in 10 of the 3 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
demo-builder
Nobody buys a description. A demo that takes 2 minutes to watch does what a 20-page deck can't: it lets someone see their problem go away. It only has to show the one moment that matters, and it has to be honest about what it is.
${CLAUDE_SKILL_DIR} below is this skill's folder (where this SKILL.md lives).
python3 "${CLAUDE_SKILL_DIR}/demo.py" check --spec "${CLAUDE_SKILL_DIR}/demo.example.json"
python3 "${CLAUDE_SKILL_DIR}/demo.py" build --spec "${CLAUDE_SKILL_DIR}/demo.example.json" --out whale/demo/index.htmlThe spec
title your product's name
subtitle the one job it does here ("Late invoices")
button the one action ("Chase all")
total_label what the number at the bottom means ("Collected")
sample_data must be true: a demo never shows a real customer's records
statuses what each row moves through, in order (["Reminder sent", "Paid"])
rows 3 to 12: name, amount ($1,234.00), note ("71 days late")
accent optional color, like #D97757Steps
- Find the moment. Ask: what is the one screen where the problem goes away? For late invoices, it is a list of overdue invoices, one button, and the money coming in. Everything else can wait.
- Write the spec. Save it as
whale/demo.json. Use sample names and round, believable numbers from the buyer's world. Never use a real customer's data, and never the buyer's own records without their written okay. - Build it.
demo.py check, thendemo.py build. It writes one HTML file with no network calls. - Open it and press the button yourself. Then show the user. Watch for anything that looks more finished than it is, and say so.
- Make it theirs, if it helps. Claude can restyle the page for one buyer or add a second screen. Keep the banner. Keep it small: a demo that takes a week to build has stopped saving time.
- Host it. Any static host works (GitHub Pages, Netlify, Cloudflare Pages), or attach the file. Publishing it is
the user's call.
/whale-hunterputs the link in the pitch.
Showing it
- Say first: "This is a demo with sample data. Here's how it will work for you."
- Under 2 minutes. Show the problem, press the button, show the result, stop.
- Then ask, and listen: "How close is this to how you'd want it to work?" Every answer is a requirement.
Example from the reel
"Build a demo that chases late invoices for a freight company." Your Product, Late invoices, five rows from Granite
State Tile ($15,300.00, 71 days late) to Tern Bay Seafood ($5,240.00, 33 days late). "Chase all" moves each row to
Reminder sent, then Paid, and the total counts up to $48,210.00. This is demo.example.json (illustrative: every
business and amount in it is fictional).
What it never does
- Never removes the banner. A demo that passes for a shipped product is a false promise, and it comes back as a refund.
- Never puts a real customer's records in a demo.
- Never promises a feature in a demo that the user doesn't intend to build.
Consent and truth
- Nothing is sent, published or charged without the user's explicit yes. This skill only writes a file.
- A demo shows how it will work, not what already exists. Say so every time you show it.
- Never invent a result, a client or a review. The numbers in a demo are sample data and are labelled that way.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

