image

Trigger phrases

Phrases that activate this skill when typed to Claude Code:

• AI image generation
• generate an image
• hero image
• social media graphic
• OG image

What it does

image is a Claude Code skill from Corey Haines’s marketing-skills repo. It turns Claude into a visual content producer who picks the right tool per image type — Flux or Ideogram for stylized hero images, Nano Banana / GPT Image / Gemini for product-aware shots, Canva or Figma for templated batches — and packages outputs at the right specs for each channel. The skill activates when you mention “AI image generation”, “generate an image”, “hero image”, or “OG image”.

The output of a session is a generated image set: prompts tuned per model, output at correct dimensions for the placement (1200×630 for OG, square/4:5 for social, 16:9 for blog), and an optimization step (WebP/AVIF, compression) so the asset is web-ready, not raw render.

When to use it

Reach for it when:

• You need a blog hero or OG image and don’t want to wait on design
• You’re producing a batch of social graphics and want a templated approach
• You want a product mockup or lifestyle scene that doesn’t look stock-photo generic

When not to reach for it:

• You need brand-precise illustration with strict style guidelines — designer territory
• The image is the product (logo, full identity); use a designer

Install

The skill is distributed via Corey Haines’s marketing-skills repo. Install via the repo’s recommended path — copy the image SKILL.md into your project’s .claude/skills/image/ directory, or use the repo’s plugin install if you’ve set it up.

Once installed, the skill activates on the trigger phrases above. The first time it runs, it will check for .agents/product-marketing-context.md (or .claude/product-marketing-context.md) — populating that file with your product context first dramatically improves output quality across all of Haines’s marketing skills.

What a session looks like

A typical session has three phases:

1. Tool selection. Per image type — text-heavy graphics (Ideogram), photoreal (Flux), product-aware (GPT Image / Gemini), templated batches (Canva / Figma).
2. Prompt + iteration. A prompt structured for the chosen model, with 2–3 variations to pick from. The skill won’t give you one prompt and call it done.
3. Spec + optimization. Correct dimensions per placement, format (WebP > PNG), compression, alt-text guidelines.

The discipline that makes it work: model-fit. Asking Flux to render a logo or asking Ideogram for photoreal lifestyle is the most common failure mode; the skill picks the right tool first.

Receipts

Honest reporting on what image produces and where it has limits:

Where it works well:

• The model-selection logic saves the “I tried it on three tools and none looked right” loop
• OG image specs are correct out of the box — easy to ship without a designer review
• Optimization step keeps assets fast; AI-generated PNGs are routinely 4MB and unshippable

Where it backfires:

• Text rendering is still imperfect across most models; expect Ideogram or hand-overlay for heavy-text graphics
• Brand-style consistency across many images is hard without a fine-tuned model — the skill produces consistent enough for blog, not enough for full identity

Pattern that works: use the skill for blog heroes, OG images, and social graphics; reserve real designers for hero brand work and motion. The split keeps both throughput and craft high.

Source and attribution

Originally written by Corey Haines. The canonical SKILL.md and any supporting files live in the image folder of his marketing-skills repository.

License: MIT. You can install, adapt, and redistribute the skill, with attribution preserved.

This page documents the skill from a practitioner’s perspective. For the formal spec and any updates, defer to the source repo.