problem-presentation
Skill2 starsChanged 13 days agoMITView on GitHub
Use when a product owner, manager, or engineer needs a short presentation to explain one concrete problem and request one decision from a known audience.
In this repository10 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- design-taste-frontendSkill
- eraser-diagramsSkill
- codebase-onboarding-c4Skill
- copywritingSkill
- frontend-designSkill
- html-docs-standardSkill
- html-slidesSkill
- presentation-brandSkill
The file
We split this skill into the rules it gives an agent
6 rules. About 0 are common, 3 are distinctive, and 3 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.
3
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 9
Keep the problem narrow enough that the audience can act after one viewing.
- Very specific: line 19
Ask once for the company/brand name and official website URL, explaining that either may be omitted.
- Distinctive: line 27
Use about five short slides: decision question, visual explanation, evidence, consequences and options, and the exact ask.
- Very specific: line 29
Keep one message per slide.
- Distinctive: line 36
Use local/system fonts and inline or local assets only.
- Very specific: line 45
Follow the production workflow for the component handoffs and delivery commands.
≈ “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
Problem Presentation
Turn one problem into a decision-ready offline HTML deck and matching PDF. Keep the problem narrow enough that the audience can act after one viewing.
Frame the decision and brand
Identify the audience, observed problem, available evidence, and exact decision requested. Infer context from supplied material when safe; ask only when the problem or decision remains unclear. Separate observed facts, interpretation, and proposals. Label unverified claims as hypotheses. Never invent metrics, quotes, or certainty.
Ask once for the company/brand name and official website URL, explaining that
either may be omitted. Accept a local logo or brand guide instead. Use the
bundled presentation-brand component to produce a local brand.json and
logo before laying out slides. When no brand is available, use its clearly
provisional identity without another brand question.
Compose the deck
Use about five short slides: decision question, visual explanation, evidence,
consequences and options, and the exact ask. Adjust the count for clarity.
Keep one message per slide. Use the pinned visual-explainer slide workflow
and template for structure, navigation, and layout. Use pinned
diagram-design for one focused SVG only if a relationship, sequence, or gap
needs a diagram; pass factual labels and brand tokens into it. The problem,
evidence, and decision contract here takes precedence over generic style
suggestions in either upstream skill.
Use local/system fonts and inline or local assets only. Preserve the upstream slide engine while removing its remote font links. Add print rules that put every slide on its own PDF page. Keep text readable on a projector and phone; support keyboard navigation and deep links. Give consequential facts visible source labels. Include the offline CSP from the production workflow before any scripts or styles. Never place credentials or private customer data in the deck.
Check before delivery
Follow the production workflow for the
component handoffs and delivery commands. Run check-deck.mjs, export the PDF
with export-pdf.mjs, then check both. Inspect desktop, short, narrow, and
print views in a browser. Verify the PDF preserves every slide and source
label. If Chrome/Chromium cannot run, deliver the checked HTML and report the
PDF gate as failed; do not author a separate PDF with another renderer. Read
as the intended audience: can they name the problem, distinguish
evidence from inference, and answer the decision? Report any unchecked claim
or visual defect instead of calling the deck ready.
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.

