Infographic
Transformez un paragraphe en une image — sans jamais inventer de données
Le processus complet du texte à l'infographie : structuration de l'information, hiérarchie visuelle, couleurs et liste de vérification. Aucun chiffre inventé, jamais.

Installez-la en 3 étapes
Sans inscription. Sans outils en ligne de commande. Choisissez votre IA, copiez une ligne, collez-la.
Vous préférez le faire vous-même ? Achetez la compétence ($4.99), téléchargez le ZIP depuis votre confirmation d'achat, décompressez-le dans le répertoire indiqué ci-dessus, redémarrez votre session IA — même résultat. Vous l'avez déjà achetée ? Votre assistant IA peut la télécharger pour vous.
Preuves concrètes
Des personnes réelles ont accompli exactement ce travail avec des agents IA.
Port Claude Code skills to Codex
Lire le cas →Cas d’usageFix app bugs, harden security and ship fixes autonomously
Lire le cas →Cas d’usageTurn a Telegram message into researched, designed social posts
Lire le cas →Contenu du pack
Un fichier SKILL.md (le guide pratique) + un modèle de départ. Aperçu :
Infographic
When to use
- The user drops in an article or a chunk of text and asks "make this a graphic" for their blog, socials, or a deck.
- The user has a set of numbers (sales, survey results, comparisons) and wants them understood at a glance.
- The user is writing an explainer or tutorial and needs one overview graphic that maps the whole piece.
- The user needs a "conclusion-first" one-pager for a report or pitch instead of a Word doc.
- The user already has a draft infographic and wants a quality review against a checklist.
Workflow
- Collect the material, lock the theme (Input: source text/data → Action: distill → Output: one-sentence theme + outline)
Read everything first, then say in one sentence what this graphic is really about; list 3–5 core points. If you can't state the theme clearly, don't move on — a fuzzy theme guarantees a messy graphic. - Layer the information (Input: outline → Action: split into three layers → Output: layered draft)
Headline layer: one conclusion sentence (a graphic gets exactly one core conclusion);
key-points layer: 3–5 points, each ≤ 20 words, facts only, no adjectives;
evidence layer: numbers, sources, footnotes in small print. Lock the words before touching visuals — never in reverse order. - Pin down every number (Input: each number in the draft → Action: trace → Output: data list)
Annotate every number with three things: the value, its definition (e.g. "YoY vs QoQ", reporting period), and its source. If the source text gives no source, mark "Source: user-provided, unverified" — never invent a plausible-looking number. This is the red line of this skill; see Notes. - Set the visual hierarchy (Input: layered draft → Action: rank weight → Output: layout sketch in words)
Give each layer a visual weight so the most important thing is biggest and seen first.
Rules: headline layer largest, topmost; key-points layer in colored/icon zones, one zone per point;
evidence layer smallest, at the bottom. One reading path only (top-to-bottom or left-to-right) — never make the reader guess where to look first. - Colors and charts (Input: layout sketch → Action: pick colors + chart types → Output: palette + chart list)
Palette: 1 primary + 1 accent + neutral base (white/light grey/cream), max 4 colors total. Match chart to data type, no exceptions: trends → line, shares → donut (≤ 5 categories), comparisons → bars, processes → arrow steps. Delete any decorative chart with no data behind it. - Produce and self-check (Input: finished prompt/layout → Action: generate + check off → Output: final or revision list)
Generate from the prompt skeleton in template.md (or lay out by hand from the layered draft), then go through the template's checklist item by item. If it fails, go back to step 4/5 — never ship with known problems.
Aperçu du modèle
1. Layered draft skeleton (lock the words first, visuals later)
Headline layer (exactly 1 core conclusion for the whole graphic)
- Core conclusion:
Key-points layer (3–5 points, ≤ 20 words each, facts only)
Le pack de compétence lui-même (SKILL.md + modèle) est rédigé en anglais.


