image-prompt-writing
Skill56 starsChanged 46 days agoMITView on GitHub
为 Krea2 超写实文生图(t2i)与 Qwen-Image-Edit-2511 图生图(i2i)编写中英文提示词。使用时机:为漫剧项目生成人物三视图、场景环境图、道具白底图等资产生图,或基于参考图做一致性变体。输出 prompt 与 negative_prompt。
In this repository27 other files it gives its agents.
- infinite-creationSkill
- 0715-scriptwriterSkill
- 3d-animation-short-generatorSkill
- ai-manga-plannerSkill
- brand-promo-video-generatorSkill
- cinematic-ai-comic-directorSkill
- comic-drama-generatorSkill
- co-op-game-intro-generatorSkill
- era-consistency-optimizerSkill
- h3-prompt-writingSkill
- handdrawn-live-video-generatorSkill
- hf-drama-storyboard-scriptSkill
- manga-drama-generatorSkill
- manga-full-stackSkill
- manju-director-agentSkill
- micro-drama-creatorSkill
- minimalist-product-ad-generatorSkill
- music-video-subtitle-generatorSkill
- novel-to-skitscreenplaySkill
- novel-to-storyboardSkill
- novel-to-videoSkill
- paper-collage-explainer-generatorSkill
- papercraft-stop-motion-explainerSkill
- script-masterSkill
- script-to-mangaSkill
- story-pipeline-cnSkill
- tts-voice-designSkill
The file
Files
图像提示词编写(Krea2 t2i / Qwen i2i)
为资产生图工作流编写提示词。本工程文生图走 Krea2 超写实工作流(节点 CLIPTextEncode),图生图走 Qwen-Image-Edit-2511(参考图 + 指令)。
按资产类型选尺寸与构图
- 人物 character → 16:9:角色设定图,脸部特写(五官/发型/神态/气质)+ 全身正面、侧面、背面三视图,同一人物外貌/服装/配色完全一致,简洁浅色背景,横向布局。
- 场景 scene → 16:9:环境氛围、空间透视、光线与风格,画面完整可作背景。
- 道具 prop → 1:1:纯白背景,画面只出现该道具本身,突出形状/材质/细节。
prompt 编写规则
- 结构化优于叙事化:按「主体 → 环境/背景 → 光照/材质/色调/视角」分类描述,1
3 句,80150 字。 - 先主体核心特征,再环境,最后细节(材质、光影、色调、镜头视角)。
- 明确镜头视角(特写/中景/全景、正面/侧面/背面、平视/俯视/仰视)。
- 风格与本项目 style.md 保持一致。
- 画面内不得出现任何文字、字幕、水印、logo、标语、签名。
negative_prompt 规则
通用排除:blurry, low quality, pixelated, distorted, watermark, text, subtitles, caption, letters, words, text overlay, signature, logo, oversaturated, artificial, plastic-looking。
人物类额外排除:extra fingers, deformed hands, mutated hands, fused fingers, plastic skin, over-smoothed。
输出
每个资产输出 JSON:{"prompt":"...","negative_prompt":"..."}。prompt 用英文(写实模型对英文响应更好),结构清晰、具体可执行。
调用方式
写好后用 update_asset(id, {prompt, negative_prompt}) 写入对应资产,再用 generate_asset_image(asset_id) 或 generate_assets_batch 统一提交(批量时优先,避免反复切换工作流导致模型重载)。图生图变体用 edit_asset_image(asset_id, prompt, reference_asset_id)。人物换装用 change_outfit(character_id, outfit)(生成服装资产,category=costume)。若模型支持图片输入,写提示词前先用 view_asset(asset_id) 查看参考图,基于真实图写提示词。
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

