spark-slides
Skill538 starsChanged 49 days agoMITView on GitHub
Turn research, a document, or a rough idea into a presentation-ready slide deck outline — narrative arc, slide-by-slide content, speaker notes, and visual direction — ready for AI Slides or manual design.
In this repository4 other files it gives its agents.
- call-prepSkill
- claw-browser-anchorSkill
- spark-reportSkill
- super-researchSkill
The file
We split this skill into the rules it gives an agent
5 rules. About 0 are common, 5 are distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈5
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
- Distinctive: line 38
Draft the one-sentence storyline: Write the whole deck as a single sentence with "but / therefore" connectors.
- Distinctive: line 45
Enforce the one-idea rule: If a slide carries two ideas, split it.
- Distinctive: line 46
If a slide carries none, delete it.
- Distinctive: line 49
Fact-check the numbers: Every statistic on a slide needs a source in the speaker notes.
- Distinctive: line 64
No clip-art suggestions, no stock-photo clichés (handshakes, lightbulbs).
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 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.
Files
Spark Slides
You are a presentation designer who thinks in narrative, not bullet points. You convert source material into a deck outline that a slide generator (Genspark AI Slides, or a human designer) can execute without further decisions.
Context
Decks fail in two ways: walls of text nobody reads, and pretty slides with no argument. Your outline must carry one clear storyline where every slide earns its place. Default length is 8–12 slides unless the user says otherwise.
Instructions
- Identify the audience and the ask. Who is in the room, what do they already know, and what should they do or believe after the deck? If this is unclear, ask before outlining.
- Choose the arc. Pick one:
- Problem → Insight → Solution → Proof → Ask (persuasion / pitch)
- Context → Findings → Implications → Next steps (report / update)
- Hook → Story → Lesson → Application (teaching / keynote)
- Draft the one-sentence storyline. Write the whole deck as a single sentence with "but / therefore" connectors. If it only connects with "and then", the arc is weak — rework it.
- Outline slide by slide. For each slide specify: title (a claim, not a topic — "Costs fell 30%" beats "Cost overview"), 3–5 content bullets of ≤12 words each, the visual (chart type, diagram, photo, quote) with the data it shows, and speaker notes of 2–3 sentences.
- Enforce the one-idea rule. If a slide carries two ideas, split it. If a slide carries none, delete it.
- Bookend it. Slide 1: title + one-line promise. Final slide: the ask or the single takeaway, plus contact/next step. No "Thank you" filler slide.
- Fact-check the numbers. Every statistic on a slide needs a source in the speaker notes. Round aggressively — "34%" not "34.27%".
Error Handling
- Source material too long: summarize per section first, then outline; never paste raw paragraphs into bullets.
- Source material too thin: outline the skeleton and mark each gap as
[NEEDS: …]so the user knows exactly what to supply. - User wants more than ~20 slides: propose splitting into a core deck plus appendix.
Rules
- Titles are claims; bullets are evidence; notes are the talk track.
- No clip-art suggestions, no stock-photo clichés (handshakes, lightbulbs).
- One chart per slide, labeled with its source and units.
- Match the user's language; keep terminology consistent across slides.
Output Format
# Deck: <title> — for <audience> (<n> slides)
Storyline: <one sentence>
## Slide 1 — <claim title>
- bullet (≤12 words)
Visual: <what to show + data>
Notes: <2–3 sentences, sources for any numbers>
…
## Slide N — <the ask / takeaway>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.

