recreate-social-photo

Recreate the composition, camera feel, lighting, color treatment, and mood of a supplied social-photo reference with a supplied everyday source photo while preserving source identity. Use for referenc

SKILL.md

GitHub

Recreate Social Photo

Catalog

  • Display name: Recreate Social Photo
  • Categories: Social Media, Portrait
  • Status: Ready
  • Route label: GPT Image 2 consistency-sensitive edit
  • Tone: rose
  • Short description: Turn an everyday photo into a reference-led social photo.

What this skill does

  • Maps source identity and reference visual grammar separately.
  • Recreates framing, pose energy, lens feel, light, grade, texture, and mood without copying reference identity.

How to use

Provide one supplied person photo and one visual reference; state protected traits.

Recreate a social photo

Use my portrait as the person and image 2 only for composition, light, color, and mood.

Workflow

  1. Label identity source and visual reference; never swap roles.
  2. Inspect crop, viewpoint, subject scale, gesture, light direction, palette, contrast, grain, and background rhythm.
  3. Run gpt-image v1.0 with quality: "medium", imageSize: "2K", the supplied references, the requested or source-preserving ratio, batchCount: 1, and one result.
  4. Preserve identity; reproduce visual grammar rather than the reference face, logos, or text.
  5. Reject identity drift, generic studio posing, filter-only output, collage, duplicates, or copied text.

User-facing output

  • Media type: One recreated social photograph
  • Default quantity: 1
  • Content per image: One coherent social-photo frame
  • Default layout: Single image; no collage
  • Model policy: GPT Image 2 Medium/2K for consistency-sensitive final image editing
  • Downstream use: Personal and brand social posts