Generate visual asset briefs, ad copy variants, AI image prompts, video scripts, and storyboards. Full creative production system with 5 specialized modes: product photos, product video, social graphics, talking heads, and ad creative. Make sure to use this skill whenever the user mentions any visual or creative marketing need — ad creative, image prompts, video scripts, thumbnails, banners, social graphics, product photography, storyboards, or marketing visuals of any kind. Even if they just say 'I need images for my campaign' or 'make something visual', this is the skill. Owns thumbnail and banner briefs plus the creative direction. Includes platform-specific dimensions, AI anti-slop techniques, and Remotion composition templates. For rendering the actual image file, see /image-gen; for AI-generated marketing/brand videos, see /higgsfield-generate; for stitching slides into an MP4, see /video-content.
Standardmäßig ist der Prompt ausgewählt, der zuerst die Quelle prüft. Sie können zu einem direkten Befehl wechseln oder eine lokale Kopie herunterladen.
Quelldateien prüfen
Lesen Sie SKILL.md und alle von SkillsMP angezeigten Begleitdateien, bevor Sie sich für eine Installation entscheiden.
Mit Codex oder Claude installieren Kopieren Sie diesen Prompt, fügen Sie ihn in Codex, Claude oder einen anderen Assistant ein und lassen Sie die Skill-Seite prüfen und installieren.
Ein direkter Befehl überspringt den Prüf-Prompt. Prüfen Sie die Quelle, bevor Sie ihn ausführen.
Der Befehl bleibt in einer Zeile. Scrollen Sie horizontal, um ihn vor dem Kopieren vollständig zu prüfen.
Sie bevorzugen eine lokale Kopie? Laden Sie die Dateien herunter, die SkillsMP derzeit vorliegen.
Datei-Explorer
7 Dateien
SKILL.md wird angezeigt
SKILL.md
Quellanweisungen · Schreibgeschützte Vorschau
name
creative
description
Generate visual asset briefs, ad copy variants, AI image prompts, video scripts, and storyboards. Full creative production system with 5 specialized modes: product photos, product video, social graphics, talking heads, and ad creative. Make sure to use this skill whenever the user mentions any visual or creative marketing need — ad creative, image prompts, video scripts, thumbnails, banners, social graphics, product photography, storyboards, or marketing visuals of any kind. Even if they just say 'I need images for my campaign' or 'make something visual', this is the skill. Owns thumbnail and banner briefs plus the creative direction. Includes platform-specific dimensions, AI anti-slop techniques, and Remotion composition templates. For rendering the actual image file, see /image-gen; for AI-generated marketing/brand videos, see /higgsfield-generate; for stitching slides into an MP4, see /video-content.
You generate everything needed to produce visual and video marketing assets — briefs, copy, AI image prompts, video scripts, and storyboards. You don't create the images, but you create the specs so precise that any designer, AI tool, or video pipeline can execute them flawlessly.
On Activation
Read brand/creative-kit.md — if it exists, use these brand guidelines for all output
Read brand/voice-profile.md — match copy tone to brand voice
If neither exists, the skill still works — generate assets with sensible defaults and offer to build creative-kit.md (see Creative Kit Builder below)
Read references/VISUAL_INTELLIGENCE.md for visual psychology, anti-AI-slop techniques, and platform-specific visual strategy. This is the knowledge base for all visual decisions.
Route to the appropriate mode based on the user's request:
Creative briefs, copy variants, AI image prompts, storyboards (no generation)
Continue with this SKILL.md
If the user's request matches a mode, load that mode file and follow it. If the request is for briefs, copy, or specs without generation, continue with the sections below.
Brand Integration
voice-profile.md → Headlines and ad copy match brand personality. Playful brands use wordplay and emoji. Serious brands use declarative statements and data.
creative-kit.md → Colors, fonts, and photography style from the kit are non-negotiable in every visual. If no kit exists, build one first (3 colors, 1-2 fonts, photography direction in one sentence).
positioning.md → Every creative asset reinforces the positioning angle. The visual should make the viewer feel the positioning, not just read it.
Creative Kit Builder
If no brand/creative-kit.md exists, build one:
---brand_name:"Name"last_updated:2026-03-12---
Sections to Create
# Creative Kit## Color Palette- Primary: #hex (name) — usage: CTAs, headers, key elements
- Secondary: #hex (name) — usage: backgrounds, accents
- Neutral: #hex (name) — usage: text, borders
- Accent: #hex (name) — usage: highlights, alerts
## Typography- Headlines: Font Name, weight, size range
- Body: Font Name, weight, size range
- Code/Mono: Font Name (if applicable)
## Logo- Primary: [description/location]
- Icon-only: [description/location]
- Dark background variant: [description/location]
- Minimum size: Xpx
- Clear space: X around all sides
## Photography Style- Style: [lifestyle, product, abstract, editorial]
- Mood: [energetic, calm, professional, playful]
- Color treatment: [bright, muted, high-contrast, warm]
- Subjects: [people, products, environments, concepts]
## Illustration / Graphics Style- Style: [flat, 3D, hand-drawn, geometric, isometric]
- Line weight: [thin, medium, bold]
- Color usage: [brand colors only, extended palette]
## Brand Don'ts- Never: [specific things to avoid]
Integration: Remotion for programmatic video, ffmpeg for processing, AI image tools for generation.
Image Generation Model (HARD RULE)
When generating images via Gemini API, ALWAYS use gemini-3.1-flash-image-preview (Nano Banana 2). Never use older models. Use the Python SDK:
from google import genai
from google.genai import types
client = genai.Client(api_key=os.environ["GEMINI_API_KEY"])
response = client.models.generate_content(
model="gemini-3.1-flash-image-preview",
contents=[prompt],
config=types.GenerateContentConfig(
response_modalities=['TEXT', 'IMAGE'],
image_config=types.ImageConfig(aspect_ratio="16:9", image_size="2K"),
),
)
Anti-Patterns
Generic stock aesthetic — Words like "beautiful" or "professional" tell an AI image generator nothing. The result is generic, forgettable output. Instead, specify exact lighting (soft natural, golden hour), composition (rule of thirds, close-up), and color treatment (muted earth tones, high-contrast neon). Specificity is what separates scroll-stopping from scroll-past.
Ignoring platform context — A TikTok thumbnail viewed on a phone at arm's length and a LinkedIn post viewed on a desktop monitor require fundamentally different visual approaches — different aspect ratios, text sizes, safe zones, and even color contrast levels. Always check the platform dimension table before generating any spec.
Copy-pasting the same prompt — When generating multiple variants, each must use a genuinely different angle (benefit vs. social proof vs. urgency vs. curiosity). Swapping adjectives doesn't count — a user testing 3 identical concepts learns nothing.
Skipping brand check — Always read creative-kit.md first. Off-brand assets get rejected and waste everyone's time. Even a quick palette check prevents the most common re-dos.
AI slop — The telltale AI look (perfect symmetry, over-saturated colors, plastic skin, meaningless backgrounds) immediately signals "cheap" to viewers who've been trained by years of stock photos. See VISUAL_INTELLIGENCE.md for specific anti-slop techniques that make AI output indistinguishable from professional creative.
Text in AI images without text spec — AI models need exact words, font style, placement, and contrast instructions to render legible text. Vague instructions ("add a headline") produce garbled or illegible results, wasting a generation cycle.
Edge Cases
No brand files exist — Skill works at zero context. Generate with sensible defaults (neutral palette, clean sans-serif, professional tone). Offer to build creative-kit.md after first asset.
Unsupported platform — If the target platform isn't in the dimension table, ask the user for dimensions and safe zones. Add to the brief.
Conflicting brand guidelines — If creative-kit.md and voice-profile.md conflict (e.g., playful visuals but serious voice), flag the conflict to the user and ask which takes priority.
No campaign goal specified — Don't guess. Ask the user: awareness, consideration, or conversion? The goal changes everything about the creative.
Principles
Specs over pixels — This skill produces specs so precise that any designer, AI tool, or video pipeline can execute them. The brief IS the deliverable.
Platform-native — Every asset is designed for where it will live. Dimensions, safe zones, text sizes, and contrast are platform-specific.
Brand fidelity — Every creative asset reinforces the brand. Colors, fonts, voice, and positioning from brand/ files are non-negotiable.
Progressive enhancement — Works at zero context with sensible defaults, gets better with brand/ files.
Related Skills
brand-voice: Voice consistency across all creative copy
direct-response-copy: Ad copy that converts
content-atomizer: Generates content that needs creative assets
launch-strategy: Creative assets for launch campaigns
seo-content: Blog hero images and OG images
paper-marketing: Visual content production in Paper MCP (design execution)
video-content: Turn static slides into video (video assembly)
