agentleFS
tonysinatonysina/claude-skills/skills/great-demo/SKILL.md

great-demo

Skill2 starsChanged 6 months agoMITView on GitHub

Apply Peter Cohan's Great Demo! method to design, critique, and qualify software demonstrations. Use when the user asks to plan a demo, build a demo script or outline, critique an existing demo, qualify a prospect, build a Chain of Pain, calculate the Delta, or handle special-case demos (trade show, large unqualified group, remote, scripted RFP). Triggers on "design a demo", "critique this demo", "is this demo any good", "Chain of Pain", "what is their CBI", "Specific Capability", "Not Now List", "To Do List", "do the last thing first", "remote demo", "trade show demo". Product-agnostic. Do NOT use for general slide decks or thought leadership; do NOT use for evaluation management or Sequence of Events work.

The file

We split this skill into the rules it gives an agent

3 rules. About 0 are common, 3 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈3

    Distinctive · estimate

    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
  1. Distinctive:

    If a request would violate the second principle, say so plainly before producing the work the user asked for.

    line 31
  2. Distinctive:

    Read the user's request and route to the right mode.

    line 35
  3. Distinctive:

    If unsure: ask one question to disambiguate.

    line 46

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Great Demo

Apply the Great Demo! method to plan, sharpen, and recover software demos.

Two Principles

These are the bedrock. Every recommendation in this skill ties back to them.

1. Do The Last Thing First. Put something that matters to the audience in front of them early, then let them pull you into depth. Ask of any plan: when does the audience first see something that matters to them? The audience self-selects depth. You don't have to show everything.

2. CBI to Reason to Specific Capability. Every capability shown must trace back to a customer's Critical Business Issue. A capability with no CBI behind it is a feature parade, and feature parades lose deals. Pain flows down the org chart. Gain flows up. If you can't name the CBI a screen addresses, cut the screen.

If a request would violate the second principle, say so plainly before producing the work the user asked for.

Mode Detection

Read the user's request and route to the right mode. If it spans modes, name them and ask which to tackle first.

User intentSignalsLoad
Design a new demo"design a demo", "build a demo for", "plan a demo", "structure this demo", "outline a demo", prospect context with goalsreferences/design.md
Critique an existing demo"critique this demo", "review my demo script", "is this demo any good", "what's wrong with this demo", "score this against Great Demo!"references/critique.md
Pre-demo qualification"qualify this prospect", "build a Chain of Pain", "what's their CBI", "what's the Delta", "Demo Information Sheet"references/qualify.md
Live-demo delivery questions"Not Now List", "To Do List", "handle a Stupid Question", "buzzwords", "mouse movement", "recover from a crash", "what to do when X happens mid-demo"references/delivery.md
Special-case demo"trade show demo", "large unqualified group", "remote demo", "scripted RFP demo", "deployment demo", "canned demo", "Generic Demo"references/special-cases.md
Term lookup"what does CBI mean", "define Specific Capability", "what's the Delta"references/glossary.md

If unsure: ask one question to disambiguate. Don't guess silently when the request is ambiguous between modes.

Working With a Vendor-Specific Voice Skill

This skill is product-agnostic by design — it owns qualification and structure, not brand voice or product facts. If a brand-specific content skill exists for the vendor in question (e.g. bluedolphin-collateral for BlueDolphin), use this skill to build the plan (CBI mapping, which screens, time per section) and hand it to that skill to apply voice and fact-sourcing when a script is needed. If that vendor skill's own "demo" mode already defers to this skill for structure, don't duplicate the qualification work — check its routing first.

Universal Pre-Work

For Design and Critique modes, you need this minimum context. If the user hasn't supplied it, ask before drafting.

  • Audience: who is in the room, job titles, decision authority
  • Time budget: how many minutes for the demo segment
  • CBI per stakeholder: what they each need to solve
  • Specific Capabilities being demonstrated: the agreed list
  • Demo objective: Technical Proof or Vision Generation (Cohan rules out Information demos)

Gaps are normal. Flag the gap, propose an assumption, proceed if the user confirms. Never proceed silently on a missing CBI; it's the load-bearing input.

Output Norms

Different modes produce different output, but a few rules apply across all of them.

State concrete time estimates for each section and the total against the budget the user gave, not vague ranges. If a draft exceeds the budget, name it and recommend the cut: trim depth first, then drop capabilities tied to the least important pain.

Quantify the Delta whenever possible: in money, time, or people. Not in features. "3 days with 4 people becomes 1 day with 1 person, 11 person-days saved per cycle, 20 cycles per year, 220 person-days saved per year" beats "much faster" every time. If the user has no numbers yet, surface the missing measurement and suggest the question that would extract it.

Flag buzzwords. The Content-Free Buzzword-Compliant list (robust, powerful, flexible, integrated, seamless, extensible, scalable, interoperable, easy-to-use, intuitive, user-friendly, comprehensive, best-of-breed, world-class, cutting-edge) erodes credibility. Replace with concrete facts: "10,000 users in production worldwide" beats "robust"; "3 clicks to complete the task" beats "easy-to-use." When critiquing, point them out by name.

Bundled Templates

The assets/ directory holds reusable templates from the book's appendices. Pull them down when the user wants a working artifact rather than a free-form draft.

AssetUse when
assets/demo-information-sheet.mdCapturing CBI/Reason/Specific Capability per stakeholder, plus the Delta and meeting objective
assets/meeting-information-sheet.mdLogging meeting logistics, agenda, and roles
assets/infrastructure-checklist.mdPre-demo hardware, software, network, AV verification

Reference Index

FileWhen to read
references/design.mdBuilding a new demo from prospect context
references/critique.mdScoring an existing demo against the Great Demo! rubric
references/qualify.mdPre-demo discovery: CBI/Reason/SC, Chain of Pain, Delta calculation
references/delivery.mdLive-demo techniques: question handling, language, mouse, props, bug recovery
references/special-cases.mdTrade show, large unqualified, remote, scripted RFP, deployment, canned demos
references/glossary.mdDefinitions of Cohan's terms
references/changelog.mdSkill version history

Critical Habits

A few habits matter enough to repeat at the top level.

Never use keyboard shortcuts in a live demo. The audience needs to see menu choices to envision themselves doing it. Shortcuts break the vision.

Use the customer's name in file names, never "demo" or "test." Realistic names build the vision of real use.

Restate every question before answering. It identifies the questioner, gives you a beat to think, and turns the answer into a personalized 1:1 dialog.

The Not Now List is the question-management tool. Acknowledge sincerely, write publicly, ask "is that OK with you?", and answer it later without fail. Failing to answer a Not Now List item leaves an open wound that can cost the sale.

Download now (ZIP)

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
What did you use it for? e.g. Reviewing pull requests in a Rails app

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.