agentleFS
OKHP3OKHP3/skillz/.agents/skills/okhp3-replit-build-in-public/SKILL.md

okhp3-replit-build-in-public

Skill3 starsChanged 2 months agoMITView on GitHub

Full playbook for a Replit Designathon competition submission — canvas organisation for judges, artifact selection, dual deployment (Replit Publish + GitHub Pages), session-to-session narrative continuity, and building-in-public commit discipline. Activate for "designathon", "competition", "building in public", "judges", "showcase", "submission", "dual deploy", "GitHub Pages + Replit", or any request to prepare a Replit project for public or competitive review.

The file

We split this skill into the rules it gives an agent

12 rules. About 0 are common, 9 are distinctive, and 3 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈9

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 3

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Always call this after organising for judging.

    line 94
  2. Distinctive:

    Every interaction the user tries in the first 30 seconds works

    line 114
  3. Distinctive:

    Every Publish is a production deploy — test the main artifact locally before clicking.

    line 141
  4. Very specific:

    Read the okhp3-vite-github-pages skill for the full runbook.

    line 147
  5. Distinctive:

    After any feature is merged:

    line 156
  6. Distinctive:

    Verify the GitHub Actions deploy completed successfully (green checkmark on the commit).

    line 157
  7. Distinctive:

    Update README.md if any measured figures (RTP, test count, decision count) changed.

    line 159
  8. Very specific:

    Write a handoff document (<DATE>-HANDOFF.md or append to an existing HANDOFF.md) that covers:

    line 169
  9. Distinctive:

    Every measurement, figure, or claim that appears in the README, pitch deck, or overkillhill.com project page must be traceable to a source:

    line 180
  10. Distinctive:

    Test counts: cite the exact command that produced the number

    line 183
  11. Distinctive:

    Commit after every meaningful milestone — not after every file save, not after every three days.

    line 217
  12. Distinctive:

    Use this before any public review or Designathon submission.

    line 223

≈ “Common” is an estimate: at least 1 in 5 of the 372,429 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

okhp3-replit-build-in-public

OverKill Hill P³ · overkillhill.com · github.com/OKHP3

Building in public on Replit is not just about shipping code — it is about making your process visible. This skill covers the full arc: how to organise what you have built so judges can evaluate it quickly, how to deploy it to two public surfaces, how to maintain narrative continuity across many build sessions, and how to make your commit history a readable story.


Scope

In scopeOut of scope
Canvas as a judge showcase boardCanvas callback mechanics — read the canvas skill
Artifact selection for judgesGitHub Pages mechanics — read okhp3-vite-github-pages
Dual deployment setupReplit deployment mechanics — read the deployment skill
Session handoff conventionsMulti-artifact monorepo rules — read okhp3-replit-multi-artifact
Commit-message disciplineIndividual artifact build instructions

The two modes of the canvas

Building mode — the canvas is a work surface. Iframes are scattered, some are building, labels are missing. This is fine during development; judges never see it.

Judging mode — the canvas is a showcase board. Every iframe is live, zones are labeled and numbered, a focusCanvasShapes call navigates judges to the layout automatically.

Switch to judging mode before every public review. The canvas is the first thing a judge sees in the Replit workspace.


Canvas organisation for judges

Hero row

The top strip of the board. Place your three primary deliverables here — the live interactive artifact (game, app, tool), the portfolio/pitch deck, and the showcase video. These are the artifacts judges interact with directly.

y = 100
x positions: 0, (w1 + 80), (w1 + 80 + w2 + 80), ...

Section label above the hero row:

y = 20, h = 60, color = grey
text = "SYSTEM ARTIFACTS — live app · portfolio deck · cinematic video"

Zone rows

Below the hero row, create labeled zones for design artifacts — mockups, system specs, gameplay exploration, mobile previews, storybook components. Number them in reading order.

Zone label convention:

  • text shape, color: "grey", h: 60, placed 40 units above the zone's frame baseline
  • Text pattern: "01 — ZONE NAME", "02 — ZONE NAME", etc.

Standard gutters: 80 px between sibling frames, 160 px between zones.

Final step: focusCanvasShapes

Always call this after organising for judging. Include all frame IDs and label IDs. animateMs: 500 produces a polished animated pan.

This is what judges see the moment they toggle on the canvas preview — make it count.


Artifact selection guide

Judges interact with three artifact types directly. Build all three for a strong submission.

ArtifactKindWhat judges evaluate
Interactive appweb or mobileDoes it work? Is it polished? Does the mechanic feel good?
Pitch deckslidesIs the design story clear? Does it show process, not just output?
Showcase videovideoDoes it communicate the experience to someone who hasn't played it?

What makes each artifact strong

Interactive app:

  • Launches immediately with no blank state
  • Every interaction the user tries in the first 30 seconds works
  • Mobile-responsive if the competition has a mobile track
  • Real data or seeded data — no lorem ipsum, no "coming soon" cards

Pitch deck:

  • Tells the process story: what was hard, what decisions were made, what was measured
  • Includes a simulation/measurement slide if you ran any quantitative analysis
  • Visual language consistent with the app's brand
  • 10-14 slides: opener, problem/opportunity, the build, the interesting decisions, the measurements, what's next, closer

Showcase video:

  • 60-120 seconds — under 30 s feels thin, over 180 s loses judges
  • Announcer or narrated voice-over explaining the app's appeal in plain language
  • Real in-game footage, not slide captures
  • Synced sound effects and music bed
  • Ends with a clear call to action (the URL, the repo, or the live app)

Dual deployment: Replit Publish + GitHub Pages

Maintain two live public URLs — one Replit-hosted, one GitHub Pages. This demonstrates continuous deployment practice and gives judges two independent ways to reach the project.

Replit Publish

  1. Click Publish in the Replit Publishing pane after any significant milestone.
  2. The published URL is at <project-name>.replit.app — get the exact URL from the deployment skill.
  3. Every Publish is a production deploy — test the main artifact locally before clicking.

GitHub Pages

For a Vite-based web artifact:

  1. Read the okhp3-vite-github-pages skill for the full runbook.
  2. The GitHub Actions deploy workflow triggers on every push to main.
  3. Set base in vite.config.ts to /<repo-name>/ for the Pages subdirectory.
  4. The Pages URL is https://<github-username>.github.io/<repo-name>/.

Keep both URLs in README.md. Judges who browse the GitHub repo should be able to reach the live app in one click without going to Replit.

Sync discipline

After any feature is merged:

  1. Verify the GitHub Actions deploy completed successfully (green checkmark on the commit).
  2. Click Publish in Replit if the change is significant enough to warrant a production deploy.
  3. Update README.md if any measured figures (RTP, test count, decision count) changed.

Session handoff conventions

A Designathon build runs across many sessions, potentially with different AI agents in each. Without handoff documents, narrative continuity breaks — the next session starts from scratch instead of where the last one left off.

What goes in a session handoff

Write a handoff document (<DATE>-HANDOFF.md or append to an existing HANDOFF.md) that covers:

  1. What was accomplished — features shipped, decisions made, measurements taken
  2. Current state — which workflows are running, what the git status is, what URL is live
  3. Open items — anything that wasn't finished, decisions that need a ruling
  4. Next action — the single most important thing the next session should do first, stated as one sentence

Keep handoffs short. A handoff that is longer than one page will not be read. The full history is in git.

Provenance discipline

Every measurement, figure, or claim that appears in the README, pitch deck, or overkillhill.com project page must be traceable to a source:

  • Simulation results: cite the script, the seed range, and the spin count
  • Test counts: cite the exact command that produced the number
  • RTP figures: cite which fleet produced them and what player model was assumed
  • Decision counts: cite the exact DECISION-LOG.md row range

Why. Judges who fact-check find discrepancies. Documentation that quotes a test run cannot drift. Documentation that quotes a previous document drifts every time it is copied.


Building-in-public commit discipline

Your commit history is visible on GitHub. Judges and potential collaborators read it. Make it a readable story.

Good commit messages

feat: add Phoebe's Lap Quest as UniGlee act 5
fix: restore showcase links to single-row icon labels on mobile
docs: update full-game RTP to 98.1% from external multi-agent validation
chore: prune 4 stale subrepl-* remotes from git remote

Pattern: <type>: <present-tense description of what changed and why it matters>

Types: feat, fix, docs, chore, refactor, test, style

What to avoid

  • "update" — update what?
  • "fix bug" — which bug?
  • "WIP" — staging area, not a public story
  • Merge commits that reference task IDs with no other context

Commit frequency

Commit after every meaningful milestone — not after every file save, not after every three days. A good rhythm is 2-5 commits per session. Each commit should be independently meaningful.


Building-in-public checklist (pre-submission)

Use this before any public review or Designathon submission.

Canvas:

  • All iframes are "live" (no "building" or "modifying" states)
  • Hero row has the three primary deliverables
  • Each zone has a numbered text label above it
  • focusCanvasShapes called on all frames and labels

Artifacts:

  • Interactive app launches with no blank states
  • Pitch deck tells the process story, not just the output
  • Showcase video runs end to end without error

Deployment:

  • Replit Published URL is live
  • GitHub Pages URL is live
  • Both URLs are in README.md
  • GitHub Actions deploy is green on the latest commit

Documentation:

  • README.md figures (test count, RTP, decision count) match the latest measurements
  • Session handoff written for continuity
  • DECISION-LOG.md is current (no open items that should be settled)

About

Built from retrospective analysis of the Glee-fully Chai Chasers Designathon project at overkillhill.com. By Jamie Hill · OverKill Hill P³ · github.com/OKHP3 MIT License — free to use, fork, and adapt. A nod to the source is appreciated.

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.