comfyui-retro-anime
Compare original and translation side by side
🇺🇸
Original
English🇨🇳
Translation
ChineseProvided by TippyEntertainment
由TippyEntertainment提供
This skill is designed for use on the Tasking.tech agent platform (https://tasking.tech) and is also compatible with assistant runtimes that accept skill-style handlers such as .claude, .openai, and .mistral. Use this skill for both Claude code and Tasking.tech agent source.
本技能专为Tasking.tech Agent平台设计,同时也兼容支持技能式处理程序(如.claude、.openai、.mistral)的助手运行时。可将其用于Claude代码及Tasking.tech Agent源码。
ComfyUI Retro Anime Skill
ComfyUI复古动漫技能
You control workflows in that generate images, videos/movie frames, sound effects, and voices using ComfyUI.
c:\anime-creatorYour job is to turn natural-language requests into concrete prompts that follow one global style template so all assets look and feel like they belong to the same late‑90s / early‑2000s anime universe.
你可以控制中的工作流,通过ComfyUI生成图像、视频/电影帧、音效及语音。
c:\anime-creator你的任务是将自然语言请求转换为符合统一风格模板的具体提示词,确保所有资产都看起来属于同一个90年代末/2000年初的复古动漫宇宙。
Global Prompt Template
全局提示词模板
For all creations (characters, movie frames, sound effects, voices, images),
start from this base template:
Retro anime style of the late 1990s
and early-2000s, full
body shot [GENDER and RACE] [PHYSICAL DESCRIPTION], retro anime screen still --ar 16:9 --v 7.0
针对所有创作内容(角色、电影帧、音效、语音、图像),请从以下基础模板开始:
90年代末至2000年初的复古动漫风格,全身镜头 [性别与种族] [外貌描述],复古动漫画面截图 --ar 16:9 --v 7.0
Filling the placeholders
占位符填充规则
-
[GENDER and RACE]
Short phrase, e.g.:young Japanese womanBlack teenage boyLatina girlmiddle-aged white man
-
[PHYSICAL DESCRIPTION]
1–2 short clauses covering:- body type
- hair (style/color)
- clothing / outfit
- key props or vibe
Example prompts:
Retro anime style of the late 1990s and early-2000s, full body shot young Japanese woman with short black hair, blue school uniform and messenger bag, retro anime screen still --ar 16:9 --v 7.0Retro anime style of the late 1990s and early-2000s, full body shot tall Black man with dreadlocks, green bomber jacket and headphones, retro anime screen still --ar 16:9 --v 7.0
You may optionally add a scene clause after the physical description for
movie frames (e.g., “standing on a rainy neon-lit city street at night”) while
keeping everything else unchanged.
-
[性别与种族]
简短短语,例如:年轻日本女性黑人少年拉丁裔女孩中年白人男性
-
[外貌描述]
1-2个简短分句,涵盖:- 体型
- 发型/发色
- 服装/穿搭
- 关键道具或氛围
提示词示例:
90年代末至2000年初的复古动漫风格,全身镜头 年轻日本女性,留着黑色短发,身着蓝色校服并背着邮差包,复古动漫画面截图 --ar 16:9 --v 7.090年代末至2000年初的复古动漫风格,全身镜头 高大黑人男性,留着脏辫,身着绿色飞行员夹克并戴着耳机,复古动漫画面截图 --ar 16:9 --v 7.0
针对电影帧,你可以选择在外貌描述后添加场景分句(例如:“站在雨夜中霓虹闪烁的城市街道上”),其余部分保持不变。
Modalities
多模态适配
You use the same character/scene description across modalities so they feel
coherent.
你需要在不同模态中使用相同的角色/场景描述,确保内容连贯统一。
1. Images (Characters & Frames)
1. 图像(角色与帧)
- Use the template directly as the main positive prompt.
- For characters, keep backgrounds simple unless specified.
- For movie frames, add a scene or action clause:
..., running through a crowded train station, retro anime screen still --ar 16:9 --v 7.0
- Keep [GENDER and RACE] and [PHYSICAL DESCRIPTION] identical across multiple frames of the same character so design stays consistent.
- 直接使用模板作为主要正向提示词。
- 针对角色,除非有特别说明,否则背景需保持简洁。
- 针对电影帧,添加场景或动作分句:
..., 跑过拥挤的火车站,复古动漫画面截图 --ar 16:9 --v 7.0
- 同一角色的多帧画面中,[性别与种族]和[外貌描述]必须完全一致,确保设计连贯性。
2. Sound Effects
2. 音效
-
Still anchor the description in the same retro-anime world.
-
Use the character/scene text as context, then specify the sound:Example (internal text for the SFX model):
Retro anime style of the late 1990s and early-2000s. Full body shot young Japanese woman with short black hair in a school uniform running through a rainy city street. Generate the diegetic soundscape that matches this anime screen still: footsteps splashing in puddles, distant traffic, soft rain.
-
Ensure the mood and energy match the described shot (calm, tense, action, etc.).
-
仍需以同一复古动漫世界的描述为基础。
-
将角色/场景文本作为上下文,然后指定音效:示例(SFX模型内部文本):
90年代末至2000年初的复古动漫风格。全身镜头 身着校服的年轻日本女性留着黑色短发,在雨中的城市街道奔跑。生成与该动漫画面匹配的自然音景:脚步溅起水花的声音、远处的交通声、轻柔的雨声。
-
确保氛围与能量水平和描述的画面匹配(平静、紧张、动作感等)。
3. Voices
3. 语音
-
Use the same character description and era/style as context.
-
Specify:
- gender/age,
- emotional tone,
- language/accent,
- speaking style (e.g., “typical late-90s shounen protagonist”):
Example internal prompt:Retro anime style of the late 1990s and early-2000s. Full body shot teenage white boy with messy blond hair, school uniform and skateboard. Generate his voice: energetic male teen, slightly raspy, expressive, Japanese-accented English, sounds like a late-90s shounen anime protagonist.
-
Use the same description whenever this character speaks again.
-
使用相同的角色描述及年代/风格作为上下文。
-
指定:
- 性别/年龄,
- 情绪语调,
- 语言/口音,
- 说话风格(例如:“典型的90年代末少年动漫主角”):
示例内部提示词:90年代末至2000年初的复古动漫风格。全身镜头 留着凌乱金发的白人少年,身着校服并拿着滑板。生成他的语音:充满活力的少年男性,略带沙哑,富有表现力,带日语口音的英语,听起来像90年代末少年动漫主角。
-
当该角色再次发声时,使用相同的描述。
When to Use This Skill
技能适用场景
Use this skill when the user asks to:
- Create or update anime characters.
- Generate movie frames/scenes, storyboards, or key art.
- Produce sound effects or voices tied to these characters/scenes.
- Maintain a cohesive retro‑anime aesthetic across a project.
Do not use this skill for:
- Non-anime styles (realistic photos, Western cartoons, UI mockups, logos).
- Assets that must match a different, explicitly specified art direction.
当用户提出以下需求时,使用本技能:
- 创建或更新动漫角色。
- 生成电影帧/场景、故事板或关键美术作品。
- 制作与这些角色/场景相关的音效或语音。
- 在项目中保持统一的复古动漫美学风格。
请勿在以下场景使用本技能:
- 非动漫风格(写实照片、西方卡通、UI原型、标志)。
- 必须匹配其他明确指定的艺术方向的资产。
Workflow
工作流程
-
Parse the request
- Identify: character(s), scene, mood, modality (image, frame, sfx, voice).
- If gender, race, or physical description are missing or ambiguous, ask 2–3 clarifying questions.
-
Construct the base prompt
- Fill and
[GENDER and RACE].[PHYSICAL DESCRIPTION] - Add optional scene/action clause for frames and audio.
- Preserve suffix for visual generations.
--ar 16:9 --v 7.0
- Fill
-
Map to modality
- Images/frames: use prompt directly.
- SFX/voices: reuse the same descriptive text as context, then add explicit audio instructions.
-
Maintain consistency
- For existing characters, reuse the same gender/race/physical description and only adjust pose, scene, or emotion per request.
- Keep era and style language (retro late‑90s / early‑2000s anime) unchanged.
-
解析请求
- 确定:角色、场景、氛围、模态(图像、帧、音效、语音)。
- 如果性别、种族或外貌描述缺失或模糊,提出2-3个澄清问题。
-
构建基础提示词
- 填充和
[性别与种族]。[外貌描述] - 针对帧和音频,添加可选的场景/动作分句。
- 保留视觉生成的后缀。
--ar 16:9 --v 7.0
- 填充
-
适配模态
- 图像/帧:直接使用提示词。
- 音效/语音:复用相同的描述文本作为上下文,然后添加明确的音频指令。
-
保持一致性
- 针对已有的角色,复用相同的性别/种族/外貌描述,仅根据请求调整姿势、场景或情绪。
- 保持年代与风格描述(90年代末/2000年初复古动漫)不变。
Output Format
输出格式
When this skill is invoked, respond with a concise, structured object:
-
For images/frames:
- :
typeorimageframe - : final text prompt string
prompt - (optional): stable identifier if provided
character_id
-
For sound/voice:
- :
typeorsfxvoice - : full descriptive text
context_prompt - (if applicable)
character_id
This output will be passed into the corresponding ComfyUI workflow nodes.
调用本技能时,请返回简洁的结构化对象:
-
针对图像/帧:
- :
type或imageframe - : 最终文本提示词字符串
prompt - (可选):若提供则使用稳定标识符
character_id
-
针对音效/语音:
- :
type或sfxvoice - : 完整的描述文本
context_prompt - (若适用)
character_id
该输出将传入对应的ComfyUI工作流节点。