← ClaudeAtlas

logo-makerlisted

Brand mark / wordmark / logo generator. Default ideogram-3-quality for clean embedded text (gpt-image-2 fallback). Six presets: wordmark, minimal, illustrated, typographic, geometric, emblem. Optional palette hint. N variants per call. Use when: 'logo', 'brand mark', 'wordmark', 'логотип', 'эмблема', 'фирменный знак'.
Mikefluff/skills · ★ 19 · AI & Automation · score 79
Install: claude install-skill Mikefluff/skills
<objective> Logo generator. Input: brand name + optional tagline + style preset + palette hint. Output: N stochastic variants of the same brief. Distinct from `cover-maker`: - No artist / author / medium fields — just brand - Single image (no aspect variants — logos work the same at all sizes by design) - Different model bias: `ideogram-3-quality` is the default (best embedded text rendering) - Aesthetic conventions: clean, geometric, often single-color, isolated on background Distinct from `flyer-maker`: - No event details — logos are timeless brand artifacts, not event-specific - No multi-aspect — same mark serves all platforms - Very minimal composition (logo dominates frame; no headline blocks) This skill does NOT: - Generate full brand books (typography systems, color tokens, voice guidelines) — single mark image only - Vectorize the output to SVG — that's a separate step (open the PNG in Vectornator / Illustrator / Affinity Designer + auto-trace) - Guarantee trademarkability — outputs may resemble existing marks; check with a trademark search before commercial use - Remove the background — chain with `bg-remover` if you need transparent PNG </objective> ## ROLE Read brand + style preset + palette → pick text-strong model → assemble variant prompts that strongly cue "logo isolated on white background, no shadows, vector aesthetic" → batch execute → save PNGs. ## PIPELINE 1. **Resolve brand**: - Required: `--brand "<name>"` - Optional: `--tagline "<line>"` 2