black-line-icon-style

Ask for a minimalist black-and-white line icon — transparent background, thick rounded strokes, single centered subject, no text — and get the image on the first turn.

When to use it

Draw from a concept:

I have a theme in my head — AI workflow, goal decomposition, review loop, data analysis — and I want a single abstract symbol that stands for the whole thing.

Make a UI icon:

I'm laying out a product page and need a feature entry icon, an empty-state icon, or a small auxiliary mark that still reads at 24-48px.

Create a supporting mark:

I want something that reads as a module badge or auxiliary graphic — logo-adjacent — without committing to full brand identity work.

Return a prompt only:

I want the generation prompt itself (Chinese, English, or with the negative list) to paste elsewhere. Say "only the prompt / just the prompt / don't generate" and the skill stops at text.

Not this:

Colorful illustration, photorealistic, or 3D → different workflow; article cover with title bar → article-cover-image; infographic with text and charts → dedicated infographic workflow; formal brand logo, trademark, or VI master mark → logo design workflow or a human designer.

What you get

By default the skill generates the image directly — no upfront analysis, no prompt-only reply — unless you explicitly ask for prompt-only mode.

• The image: one black-and-white line icon, transparent background (white fallback), thick black rounded strokes, single centered subject, no text or numbers
• Pre-generation prompt check: 7-item self-check confirms the prompt names black-and-white, transparent background, rounded strokes, single subject, no-text rule, small-size readability, and the negative-prompt list before any image call goes out
• Post-generation image check: 7-item verification on the produced image; a failure sends the prompt back to tightening
• Retry policy: the same failure twice in a row triggers a simpler geometric fallback rather than an infinite loop
• Prompt-only mode: on explicit request, returns the Chinese prompt, English prompt, and negative-prompt list instead of calling image generation
• Never does: colorful illustration, 3D or shadow effects, decorative posters, text-heavy infographics, or a finalized brand logo

Prerequisites & boundaries

Prerequisites:

An image-generation tool at runtime (Codex or Claude Code image capability). A theme, concept, or use scenario from you — the skill asks for one if missing.

Adjacent skills:

Job	Route to
Article cover / horizontal title graphic	article-cover-image
Formal brand logo, trademark, VI final artwork	logo design workflow or human designer
Text-heavy infographic or knowledge card	dedicated infographic workflow

Won't accept:

• Colorful, photorealistic, or 3D/C4D imagery
• Complex posters or article covers with title bars
• Infographics with labels, charts, or knowledge cards
• Final brand logos, trademarks, or VI master marks

Fine-line edges:

• "logo-like" auxiliary mark or concept badge → triggers
• "brand logo final / trademark / VI master mark" → does not trigger, routes out
• "only the prompt" + target is a black-line icon → returns prompt text, skips image
• "only the prompt" + target is not an icon → does not trigger