菜单
Generate or edit raster images with OpenAI's Image API from text prompts, reference images, and optional masks. Use for illustrations, portraits, concept art, product mockups, transparent assets, or controlled visual edits.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Restate the previous response plainly and concisely. Use only when the user explicitly asks for a simpler human explanation or invokes bro.
Apply correct-by-construction TypeScript engineering standards covering boundary parsing, typed failures, domain modeling, module boundaries, dependency injection, testing, and type safety. Use when writing or reviewing TypeScript.
Audit installed Agent Skills for duplicate copies, unused candidates, oversized descriptions, loaded roots, and prompt-budget pressure. Use when cleaning or debugging skill collections across common agent runtimes.
Apply concrete clarity, concision, structure, grammar, and plain-English checks when editing prose, reviewing drafts, writing reports or articles, or teaching clear writing.
Test-driven development. Use when the user wants to build features or fix bugs test-first, mentions "red-green-refactor", or wants integration tests.
Produce an implementation-ready technical specification using typed contracts, concrete module boundaries, call stacks, data flows, alternatives, file changes, and vertical red-green test slices. Use when the user asks for a tech spec or architecture handoff.
| name | chatgpt-imagegen |
| description | Generate or edit raster images with OpenAI's Image API from text prompts, reference images, and optional masks. Use for illustrations, portraits, concept art, product mockups, transparent assets, or controlled visual edits. |
| license | MIT |
| compatibility | Requires Python 3.10+, openai 2.16+, network access, and an OpenAI API key. |
Use the bundled chatgpt-img CLI for generation and edits. Resolve its path relative to this SKILL.md.
python3 -m pip install -r <skill-directory>/requirements.txt
export OPENAI_API_KEY='...'
OPENCODE_OPENAI_API_KEY is accepted as a fallback. CHATGPT_IMAGE_DEFAULT overrides the default model.
Image requests cost money. Confirm scope before generating many images or using high quality repeatedly.
<skill-directory>/scripts/chatgpt-img \
'A linocut illustration of a lighthouse in heavy rain' \
./lighthouse.png \
--size 1536x1024 \
--quality high
Single reference:
<skill-directory>/scripts/chatgpt-img edit \
./portrait.png \
'Use warm editorial lighting; preserve identity, pose, and clothing' \
./portrait-edited.png \
--quality high
Multiple references:
<skill-directory>/scripts/chatgpt-img \
--input ./portrait.png \
--input ./moodboard.png \
'Apply the moodboard lighting and palette while preserving the portrait identity' \
./portrait-styled.png
Optional masked edit:
<skill-directory>/scripts/chatgpt-img edit \
./input.png \
'Replace the masked background with a cloudy coastline' \
./output.png \
--mask ./mask.png
--model: defaults to gpt-image-2--size: supported dimensions or auto--quality: GPT Image uses auto, low, medium, or high--background: auto, transparent, or opaque--format: png, jpeg, or webp--compression: 0 through 100 for JPEG/WebP--num: number of outputs--input: repeatable reference image for edit mode--mask: optional edit mask--input-fidelity: supported by selected older GPT Image models, not gpt-image-2Use --help and edit --help for the complete interface.
State both the change and the invariants. For example, name the identity, geometry, text, logos, pose, camera angle, or materials that must remain unchanged. Describe lighting, palette, composition, and finish positively instead of relying on long negative lists.
Never expose API keys. Error output may include safe request metadata, but it must not include prompt contents or credentials.