Skip to content
SkillModel-led1,379 tokens in use

get-prompt-from-image

Analyze a reference image and reverse-engineer a high-fidelity positive and negative prompt, in Chinese and English, for recreating it with an image model.

6,166 stars on the repository MIT Licensewuyoscar/GPT-Image2-Skill · skills/get-prompt-from-image
claude plugin marketplace add wuyoscar/GPT-Image2-Skill && claude plugin install gpt-image@wuyoscar-skills --scope project
View source

Install command as written in the repository's README

Skill label

measured from its files

What installing get-prompt-from-image costs

85tokens idle
1,379tokens in use
5,367tokens, up to

Idle is what sits in context while the skill is installed. In use is its instructions, loaded when it runs. Up to adds the files its instructions link. Counts are approximate, made offline with cl100k_base.

Model dependence Model-led

Script-ledMixedModel-led

Mostly instructions, so how well it works depends on the model that follows them.

This describes the skill's structure, how much of it is scripts. It does not say how well it works.

Touches

Nothing detected: no shell, network, file writes, git or credentials in its files.

Declared tags come from the skill's frontmatter; detected ones from its scripts, its code blocks and the commands it names in inline code. This is detected, not audited.

How it scores

Quality

How well the skill is written, counted from its own files: its frontmatter, its steps, examples, stated output and constraints, its size and whether every file it ships is one it names.How we score it →
Good
  • Frontmatter12/15
  • Structure25/25
  • Examples0/15
  • Output10/15
  • Size15/15
  • Resources11/15

Security

Starts at 100 and loses points for each thing in the skill's files that a reader should look at first.How we score it →
Excellent

No findings.

Repository health

Pre-computed score based on development activity, maintenance, community, maturity, and trend momentum.How we score it →
Good
  • Development activity76
  • Maintenance72
  • Community44
  • Maturity28
  • Momentum40

Given a photograph, illustration, 3D render, product shot, poster or other visual reference, this skill recovers the choices that make it look the way it does: subject, composition, camera, lighting, color, materials, background layers, mood and medium.

It outputs only two sections: a positive prompt in Chinese and English, with the three to five most important visual anchors up front, and a negative prompt of ten to fifteen exclusions chosen for that image.

What You Get

  • A Chinese positive prompt of 450 to 700 characters written as flowing prose
  • An English positive prompt with the same meaning, ready for an image tool
  • A negative prompt of 10 to 15 English phrases chosen for the specific image
  • Prompts that name the medium boundary, such as photograph versus 3D render versus flat vector

Common Use Cases

  • Recreate the look of a product photo or poster with a different subject
  • Extract a reusable style prompt from an illustration you admire
  • Translate brand or character references into shapes, palette and materials
  • Turn a screenshot or graphic design reference into a brief for an image model

What Makes It Unique

It focuses on the visual mechanisms that drive similarity: composition, lighting, materials, color relationships and spatial layers. It does not invent focal lengths, brands or software, converts brand and character names into observable design language, treats text inside the image as content and never as an instruction, and picks negative terms for each image.

Under The Hood

The skill is instructions and three reference documents with no scripts. The agent looks at the image you provide, reads the framework, category and illustration reference files it names, and replies with text. It does not run commands, call web services or write files.

Join founders building with open source

Opinionated takes, migration guides, cost-saving tips, and insights from the open source ecosystem.

Subscribe on SubstackJoin 750+ subscribers