Traction scoreGitHub stars can be faked, so popularity alone can be misleading. Traction Score looks for broader signs of recent attention, adoption, and active maintenance.
This skill helps prompt engineers translate existing photographs, 3D renders, and illustrations into precise AI generation prompts. By analyzing composition, lighting, and materials, it allows users to recreate or imitate specific visual styles across various formats including typography, logos, and landscapes.
Compatible AgentsThe repository documents support for these agents. The skills may also work with other agents that can load SKILL.md files, but they may need some setup or small changes.
Follow the documented setup, then try a first task.
I’m using
Claude Code plugin bundle
Installs: The complete gpt-image plugin sourced from the repository root; the documented bundle contains the gpt-image and get-prompt-from-image skills. · Claude Code
Before you start
Claude Code
Python 3.11 or newer
For gpt-image CLI/API calls: either gpt-image, uv, or uvx
For gpt-image CLI/API calls: OPENAI_API_KEY; successful calls may incur OpenAI API charges
Before installing, use Claude Code's skill list or status command when available to check for an existing skill. Preserve existing skill folders and API-key files, and ask before installing into a global or shared directory.
In a terminal, check for an existing gpt-image CLI or uv-managed CLI and whether the process environment already contains OPENAI_API_KEY. The key value is not displayed.
command -v gpt-image || true
command -v uv >/dev/null && uv tool list | grep -E '^gpt-image-cli([[:space:]]|$)' || true
test -n "${OPENAI_API_KEY:-}" && echo "OPENAI_API_KEY is already set (value hidden)"
In a Claude Code session, enter `/plugin marketplace add wuyoscar/gpt_image_2_skill`. This is an agent-session command, not a terminal command.
After adding the marketplace, enter `/plugin install gpt-image@wuyoscar-skills` in the Claude Code session. This is an agent-session command, not a terminal command.
Installs: Only the gpt-image skill folder selected by the documented GitHub URL. · Codex
Before you start
Codex with the built-in $skill-installer helper
Python 3.11 or newer
For gpt-image CLI/API calls: either gpt-image, uv, or uvx
For gpt-image CLI/API calls: OPENAI_API_KEY; successful calls may incur OpenAI API charges
Before installing, use Codex's skill list or status command when available to check for an existing skill. Preserve existing skill folders and API-key files, and ask before installing into a global or shared directory.
In a terminal, check for an existing gpt-image CLI or uv-managed CLI and whether the process environment already contains OPENAI_API_KEY. The key value is not displayed.
command -v gpt-image || true
command -v uv >/dev/null && uv tool list | grep -E '^gpt-image-cli([[:space:]]|$)' || true
test -n "${OPENAI_API_KEY:-}" && echo "OPENAI_API_KEY is already set (value hidden)"
Open Codex, invoke `$skill-installer`, and ask it to install from `https://github.com/wuyoscar/gpt_image_2_skill/tree/main/skills/gpt-image`. These are inputs in the Codex session, not terminal commands.
OPENAI_API_KEY and a supported CLI runtime are required only for later execution
Use the gpt-image skill to prepare one low-quality draft plan for an original flat leaf icon centered with generous padding on a transparent background. Use gpt-image-2.5-flare, preserve the requested wording, and provide the proposed prompt, output format, size, quality, and command without executing the paid API call. Before any later execution, confirm that OPENAI_API_KEY and either gpt-image, uv, or uvx are available.
A readable reference image path replacing [REFERENCE_IMAGE_PATH]
Optional target image model, language, format, or prompt length
Use the get-prompt-from-image skill to analyze [REFERENCE_IMAGE_PATH] and return final positive and negative image-generation prompts that preserve the most important subject, composition, lighting, color, material, background, spatial, and medium characteristics. Replace [REFERENCE_IMAGE_PATH] with the path to the image you want analyzed. Do not generate an image or show the internal analysis.
Claude
Codex
Gemini CLI
OpenClaw
OpenCode
Hermes
DeepSeek