agentleFS
serhiilenivserhiileniv/ccs/skills/deck/SKILL.md

deck

Skill0 starsChanged 10 days agoMITView on GitHub

Hackathon, after /ccs:pitch. Builds the whole pitch deck (7–9 slides, demo video at the centre) as offline HTML in the house design, then screenshots and checks every slide and exports PDF. User-invoked only.

  • Installs packages

7 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

2 rules. 0 are common, 0 are distinctive, and 2 are very specific to this project.

  • 0

    Common

    Most of the 6 edited versions we read keep these.

  • 0

    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
  1. Very specific:

    Read hackathon/plan.md, facts.md, script.md (from /ccs:brief and /ccs:pitch), demo/out/ (video + poster), and screenshots the repo has.

    line 15
  2. Very specific:

    every number on a slide is in facts.md

    line 59

“Common” is measured: kept by at least 6 in 10 of the 6 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.

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

/ccs:deck

The deck supports the talk and frames the demo. It never repeats the script and never shows a number the repo can't prove.

Step 0 — inputs

Read hackathon/plan.md, facts.md, script.md (from /ccs:brief and /ccs:pitch), demo/out/ (video + poster), and screenshots the repo has. No facts.md → stop and run /ccs:pitch first. Design: the house design.

Step 1 — outline

One idea per slide, mapped to a section of script.md:

#SlideContent
1TitleProduct name + the promise (the sentence, shortened). No agenda.
2ProblemOne user's moment, one visual, at most 2 costs.
3IdeaThe one idea, ≤ 12 words, or the contrast line.
4ProductOne cropped real screenshot in a device frame, ≤ 3 callouts. Crop, never shrink.
5Demodemo.mp4 full-bleed, local file, poster frame; live fallback note in speaker notes.
6–7ProofOne wow per slide: a big number from facts.md + one caption that says what it means.
8How it's builtDiagram of the trust path ("LLM proposes, code verifies"), not the stack list.
9CloseThe sentence + QR to the live link or repo.

7 slides is fine; 9 is the ceiling. Drop slides before shrinking text. Demo-data numbers are labelled as demo data. Show the outline, then build — no approval wait unless the user asked for one.

Step 2 — build

Copy ${CLAUDE_SKILL_DIR}/template/ to hackathon/deck/ (keep vendor/ and fonts/; it must present offline). Edit only the slide sections in index.html; reuse the layouts the template already has.

  • Speaker notes per slide (<aside class="notes">) = that slide's lines from script.md.
  • Images and video copied into hackathon/deck/assets/; relative paths only.
  • Colours only from the tokens in deck.css. A new colour is a design change: stop and ask.
  • ≤ 30 words, ≤ 3 bullets, ≥ 24 px, inside the 64 px safe area.

Step 3 — QA loop

Once: cd ${CLAUDE_SKILL_DIR}/qa && npm i && npx playwright install chromium. Then: node ${CLAUDE_SKILL_DIR}/qa/qa.ts hackathon/deck/index.html --out hackathon/deck/out --minutes <limit in minutes, a number>. It renders every slide, fails on overflow, safe area, small text, word and bullet counts, fonts not loaded, broken or upscaled images, missing video, off-token colours and > 9 slides, exports the PDF and checks its page count, and prints talk time from the notes.

Then read the contact sheet and every slide at 960×540 (a projector proxy):

  • one idea, takeaway readable in 5 seconds
  • consistent alignment and spacing with the other slides
  • every number on a slide is in facts.md

Fix only failing slides; re-run. At most 3 rounds per slide; still failing → report it.

Step 4 — output

hackathon/deck/index.html (present: open in a browser, S for speaker view), hackathon/deck/out/deck.pdf, hackathon/deck/out/slide-*.png, contact-sheet.png. QA's Chromium can't play H.264, so it checks the video file and poster exist; open the deck in Chrome once to see it play. pptx in arguments → also one PPTX of full-slide images with notes (not editable; say so).

Print: slide count, talk time vs limit, QA result, and the manual list: team photo, live-state screenshots not yet captured, test on the venue projector.

What QA cannot check: whether the deck persuades, or how the venue projector renders it.

Language

lang: uk → slide text and notes in Ukrainian; product names, ids and UI labels stay as in the product.

Writing

One line where one line does. Answer first. State the fact, not the story behind it — no narrating what was tried, no paragraph restating the line above. Two things always stay: why a non-obvious decision was made, and what a check cannot do. Caveats and risks still get said, one line each.

Outcome: success = QA passes, PDF exported · partial = deck built, QA fails left · failed = no deck.

Report

Last step, every run, whatever the outcome: write .ccs/reports/<skill>/<YYYY-MM-DD>-<slug>.md at the repo root, and add .ccs/ to .git/info/exclude if missing. One line per field, no secrets, tokens or customer data. Print the path.

---
skill: <skill>
date: <YYYY-MM-DD>
repo: <owner/name>
target: <argument, one line>
outcome: success | partial | failed
learned: false
---

Used: <tools, CLIs, MCPs, files that did the work>
Found: <the result, one or two lines>
Friction: <what slowed or blocked it; a user correction; "none">

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.