hackathon
Skill0 starsChanged 16 days agoMITView on GitHub
Strict phase-by-phase playbook for short hackathons (a few hours) where the pitch, demo and look matter more than depth. Walks idea → market check (ICP, competitors, market size → go / pivot / kill) → MVP scope → business model (pricing, lean canvas, scalability) → brand (name, positioning, voice, logo) → UI → build → judge-ready GitHub README → pitch deck and/or demo video (recorded in Recordly) → rehearsal, scales to the event length (2 h to 48 h) and team size (solo to 5+ with parallel tracks and a safe git workflow: worktrees, PRs, an always-demoable main), names the exact installed skill for each phase, and tracks progress in HACKATHON.md so no step gets silently skipped. Use when the user mentions a hackathon, a demo day, a time-boxed build, "what should we build", "cut the scope", "market research", "pricing", "business model", "brand voice", "make the pitch", "README", "demo video", "merge", "PR", "branch", or asks which skill to use for startup/pitch/presentation work.
4 other projects use a version of this.
Adopted by
Most recently taken up by
- chrislacey89/skillstheir own version, first seen Nov 8, 2026
- anthropics/claude-plugins-officialtheir own version, first seen Nov 8, 2026
- ChainGPT-org/chaingpt-claude-skilltheir own version, first seen Nov 1, 2026
- solanabr/startup-builder-skilltheir own version, first seen Oct 9, 2026
Other versions4 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 4
- anthropics/claude-plugins-official★ 46ksame name
- ChainGPT-org/chaingpt-claude-skill★ 7same name
- chrislacey89/skills★ 4same name
- solanabr/startup-builder-skill★ 2same 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.
The file
We split this skill into the rules it gives an agent
6 rules. 0 are common, 4 are distinctive, and 2 are very specific to this project.
0
Common
Most of the 4 edited versions we read keep these.
4
Distinctive
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 23
Not because of time, not because it seems obvious, not because the user seems to know already.
- Distinctive: line 26
Running behind is not a reason to cut: If a phase is over its timebox, say so in one line with the numbers ("phase 7 is 20 min over, 1 h 10 …
- Distinctive: line 33
Do not invent constraints: Never claim "there's no time", "this isn't needed for a hackathon" or "judges won't care" unless the user said …
- Distinctive: line 35
When unsure what the user wants, ask one short question — do not guess by dropping work.
- Very specific: line 67
Every line ends with its owner (@handle, or @all).
- Very specific: line 102
If one of these exists in your skill list, the phase must use it in addition to the skills above.
“Common” is measured: kept by at least 6 in 10 of the 4 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
Hackathon playbook
Short hackathon = judges see a ~3-minute pitch and/or a ~90-second demo video. Optimise for one working happy path + a story + a look that feels finished, and — when the judges score it — a believable business: market, pricing, how it scales. Never build breadth.
Execution contract — read first, non-negotiable
These rules override your own judgement about time, efficiency or "what matters most". Agents under perceived time pressure tend to silently drop steps; these rules exist to stop exactly that.
- Follow the phases in order and do every step of every phase. For each phase, load the named skill and do what that skill says — fully, not a summary of it.
- You never decide to skip, merge, shorten or "fast-track" a phase or a step. Not because of time, not because it seems obvious, not because the user seems to know already. Only the user can skip or reorder, and only by saying so explicitly ("skip the logo", "no video", "go straight to build").
- Running behind is not a reason to cut. If a phase is over its timebox, say so in one line with the numbers ("phase 7 is 20 min over, 1 h 10 left") and ask which phase to cut or shorten. Wait for the answer. Until then, keep doing the current phase properly.
- Every phase ends with its output on disk (the Output column below) and a
ticked line in
HACKATHON.md. A phase without its output is not done — do not start the next one. - Do not invent constraints. Never claim "there's no time", "this isn't needed for a hackathon" or "judges won't care" unless the user said it.
- When unsure what the user wants, ask one short question — do not guess by dropping work.
- When a phase points to a file in
references/(next to thisSKILL.md), read the whole file at the start of that phase and do every step in it. The steps live there to keep this file short, not because they are optional.
HACKATHON.md (create it in phase 0, update it after every phase)
# Hackathon — <project name>
Start: <HH:MM> End: <HH:MM> Deliverables: deck ☐ video ☐ live demo ☐
Profile: <2h | 4.5h | 8h | 24h | 48h> Team: <solo | 2 | 3–4 | 5+> Confirmed by user: ☐
Roles: <name/handle → role> (owner column below uses these)
- [ ] 0. Kickoff — rules, judging criteria, deliverables, profile, roles @owner
- [ ] 1. Idea check — <one-sentence problem>
- [ ] 2. Market check — depth: full ☐ lite ☐ · verdict: go ☐ pivot ☐ kill ☐ (user's call)
- [ ] 3. MVP scope — <3–4 features, critical flow>
- [ ] 4. Business model — pricing, lean canvas, scalability
- [ ] 5. Brand — BRAND.md (name, positioning, voice) + logo + DESIGN.md
- [ ] 6. UI direction — DESIGN.md direction locked
- [ ] 7. Build — happy path works end to end
- [ ] 8. Polish + feature freeze
- [ ] 9. Repo README — badges, proof table, screenshots
- [ ] 10. Pitch story
- [ ] 11. Slides (if deck)
- [ ] 12. Demo video (if video)
- [ ] 13. Rehearsal
Skipped by user: <phase — user's words — time>Every line ends with its owner (@handle, or @all). Only the user's explicit
words go under "Skipped by user". If that line is empty, every box must end up
ticked.
Phases (timeboxes shown for the 4.5 h solo profile — other lengths and team sizes: references/scale.md)
| # | Phase | Time | Use skill | Output (must exist before moving on) |
| 0 | Kickoff | 5 min (+10 git setup for teams) | — Ask: event length and team size → pick the profile and roles from references/scale.md and get the user's OK (teams of 2+: also set up references/team-git.md); judging criteria (product only? or also business model, market, scalability, viability?), pitch length, deliverables: deck, demo video, live demo, or all (default: deck + video), team roles, submission format/deadline | HACKATHON.md created, judging criteria written down |
| 1 | Idea check | 15 min | founder-validate-idea, then last30days on the problem (real complaints, quotes, numbers) | one-sentence problem + who has it + 1 real quote/number for the pitch |
| 2 | Market check | 15 min full / 5 min lite | references/market-check.md — ends with the user's go / pivot / kill decision | MARKET.md (ICP, competitors, market size, verdict) |
| 3 | MVP scope | 15 min | founder-mvp-scope (must / should / won't). Scope for the ICP and the gap found in phase 2 | MVP.md: 3–4 features max, one critical user flow, won't-have list |
| 4 | Business model | 15 min full / 5 min lite | references/business-model.md | pricing, business model and scalability added to MARKET.md |
| 5 | Brand | 20 min | references/brand.md (verbal + visual) | BRAND.md + logo SVG + DESIGN.md (2–3 colours, 1–2 fonts) |
| 6 | UI direction | 10 min, then during build | impeccable owns the direction: shape the UX/UI before code and commit to one bold, non-generic direction; typeset and colorize for fonts and palette. ui-ux-pro-max as a lookup only (--domain color, typography, google-fonts, chart or ux) for palette and font-pairing candidates — never its --design-system --persist (DESIGN.md stays the one source of truth). Then the branch for your product type in references/ui-by-product-type.md | direction + tokens written into DESIGN.md |
| 7 | Build | ~1 h 15 (full) / ~1 h 35 (lite) | normal coding, only the flow from MVP.md; seed realistic data; every UI string (headlines, buttons, empty and error states) in the BRAND.md voice — copywriting for landing/hero copy | happy path works end to end, DEMO.md with the exact click path |
| 8 | Polish + freeze | 15 min | impeccable (polish / critique / audit). Motion: impeccable's animate command picks where motion goes and builds the one focal moment; animate builds the routine UI motion (button press, dropdown, modal, drawer, toast) from its tables. Expo apps: expo-animation | no feature work after this point |
| 9 | Repo README | 15 min | references/readme.md | README.md + docs/ screenshots; repo description and topics drafted |
| 10 | Pitch story | 10 min | founder-pitch-deck (structure); tagline and key messages from BRAND.md; pull market, competition, pricing and scalability straight from MARKET.md | PITCH.md: story + speaker notes, ≤ 3 min spoken |
| 11 | Slides (if deck) | 20 min | frontend-slides (animated HTML deck in the browser; can export PDF). Need PowerPoint? use your agent's pptx skill if it has one | the deck file (+ PDF backup) |
| 12 | Demo video (if video) | 25 min | references/demo-video.md (storyboard, then Recordly) | DEMO_VIDEO.md storyboard + exported MP4 |
| 13 | Rehearsal | 15 min | grill-me to get grilled like a judge; run the pitch out loud with a timer 3× | answers to the likeliest questions — always including "how do you make money?", "how big is this?", "how does it scale?", "why you and not <competitor>?"; backup screenshots/video for a live-demo failure |
Rules of thumb
- One owner per decision: when two skills serve the same phase, the playbook says which one owns the decision and what the other adds. On a conflict the owner wins — never mix their rules or keep two sources of truth (one
DESIGN.md, oneBRAND.md). - One voice everywhere: if a sentence would not fit
BRAND.md, rewrite it. - Cut scope before writing code. If a feature is not in
DEMO.md, it does not exist. - Hardcode / seed data freely; judges do not see the backend.
- The pitch follows: problem (with a real quote or number) → who → demo → market & business (when judged) → why now → ask.
- Feature freeze at phase 8; everything after is README, deck, video and rehearsal.
- The live demo always has a backup: the exported video or screenshots.
Optional boosters — use when installed
If one of these exists in your skill list, the phase must use it in addition
to the skills above; if it does not exist, carry on without it (that is not a
skip). infographic is installed by this playbook; the rest you add yourself.
| Phase | Skill | Adds | Get it |
| 1 Idea check, 2 Market check | research | deep, cited research — use it instead of a plain last30days run, since it runs last30days itself | any deep-research skill you use |
| 1 / 3 | brainstorm | non-generic idea and feature variants (diverge → critique → converge) | Dymyt-ry/claude-code-toolbelt |
| 7 Build (dashboard) | dataviz | chart colour, mark and KPI-tile rules | built into Claude apps |
| 7 Build (landing) | landing-page-design | hero / above-the-fold / CTA layout rules | inferen-sh/skills |
| 11 Slides | pitch-deck-visuals | slide-by-slide layout and data-slide rules | inferen-sh/skills |
| 11 Slides | infographic | one strong data / process slide | installed — claude-office-skills/skills |
Other useful skills
- Research:
last30days(Reddit, X, YouTube, HN, GitHub, Polymarket; TikTok / Instagram with a ScrapeCreators key). - Ideas when stuck:
brainstormif installed, otherwisepm-brainstorm-ideas-new. - Social launch post:
social-content.
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.

