anbeime/skill / skills/video-creation-pro/video-creation-pro
video-creation-pro
Verified电商运营与自媒体创作者在为美妆、家居、智能硬件等商品制作短视频时,请使用此技能。10大AI智能体协同,从卖点提炼、分镜生图到音效字幕,一键调用COZE API自动合成高质量商品视频,彻底告别本地代码依赖,轻松实现从创作、质检到数据迭代的全流程闭环。
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 22.0 KB |
| references/agent_specs.md | markdown | 20.9 KB |
| references/audio_format.md | markdown | 3.3 KB |
| references/development_guide.md | markdown | 10.9 KB |
| references/mask_video_guide.md | markdown | 7.5 KB |
| references/quality_standards.md | markdown | 6.7 KB |
| references/storyboard_format.md | markdown | 5.6 KB |
| references/subtitle_format.md | markdown | 3.1 KB |
| references/usage_limit.md | markdown | 3.6 KB |
| references/use_cases.md | markdown | 12.7 KB |
| references/video_specs.md | markdown | 7.0 KB |
| scripts/image_quality_checker.py | script | 11.3 KB |
| scripts/video_composer.py | script | 12.2 KB |
references/development_guide.md
View on GitHub# 开发技能的核心落地建议
## 概览
本文档提供"影品智创"技能开发的核心落地建议,聚焦解决依赖包卡死+字幕问题,强制API合成。
## 1. 技能端添加关键词屏蔽
### 目标
在技能底层设置屏蔽词,若检测到智能体输出/执行中出现本地依赖包关键词,立即触发警告并阻断流程
### 屏蔽词列表
```
moviepy
opencv-python
mediapipe
pillow
numpy
cv2
Image
VideoFileClip
AudioFileClip
```
### 实现建议
```python
# 伪代码示例
def check_local_dependency(text):
"""
检测文本中是否包含本地依赖包关键词
:param text: 待检测文本
:return: (is_blocked, keyword)
"""
blocked_keywords = [
"moviepy", "opencv-python", "mediapipe", "pillow", "numpy",
"cv2", "Image", "VideoFileClip", "AudioFileClip"
]
for keyword in blocked_keywords:
if keyword.lower() in text.lower():
return (True, keyword)
return (False, None)
def handle_blocked_output(text, agent_name):
"""
处理被屏蔽的输出
"""
is_blocked, keyword = check_local_dependency(text)
if is_blocked:
print(f"【警告】检测到本地依赖包关键词:{keyword}")
print(f"来源智能体:{agent_name}")
print("已阻断流程,强制切换至纯COZE API合成流程")
return False
return True
```
## 2. 素材格式固化
### 目标
在技能中预设COZE API优选的素材格式,各智能体输出时直接按预设格式生成,无需自定义
### 预设格式规范
#### 图片素材
- 格式:JPG/PNG(COZE API优选格式)
- 分辨率:1080P
- 比例:横屏16:9或竖屏9:16
- 命名规范:镜头序号_场景名称(如01_熬夜泛红)
#### 音频素材
- 格式:MP3(API优选格式)
- 采样率:44100Hz或48000Hz
- 比特率:128kbps-320kbps
#### 字幕素材
- 格式:COZE API可直接解析的列表格式
- 包含:镜头序号、字幕文本、显示时长、格式参数、位置参数、同步参数
### 实现建议
```python
# 伪代码示例
class MaterialFormatValidator:
"""素材格式校验器"""
@staticmethod
def validate_image(image_path):
"""校验图片格式"""
valid_formats = ['.jpg', '.png', '.jpeg']
ext = os.path.splitext(image_path)[1].lower()
if ext not in valid_formats:
raise ValueError(f"图片格式不支持:{ext},仅支持JPG/PNG格式")
# 校验分辨率
img = Image.open(image_path)
width, height = img.size
if width != 1920 or height != 1080:
raise ValueError(f"图片分辨率不支持:{width}x{height},仅支持1920x1080")
return True
@staticmethod
def validate_audio(audio_path):
"""校验音频格式"""
valid_formats = ['.mp3']
ext = os.path.splitext(audio_path)[1].lower()
if ext not in valid_formats:
raise ValueError(f"音频格式不支持:{ext},仅支持MP3格式")
return True
```
## 3. API参数预设模板
### 目标
将典型视频的API合成参数保存为模板,后续同类型视频可直接调用,减少参数配置错误
### 治愈风美妆模板(玻尿酸面膜)
```json
{
"template_name": "healing_beauty_mask",
"description": "治愈风美妆面膜视频(玻尿酸面膜60秒)",
"video_params": {
"resolution": "1080P",
"aspect_ratio": "9:16",
"duration": 60,
"fps": 25,
"output_format": "MP4"
},
"transition_params": {
"type": "soft_fade",
"duration": 0.4
},
"audio_video_sync": {
"enabled": true,
"max_offset": 0.1
},
"quality_params": {
"hd_noise_reduction": true,
"color_correction": "warm_tone"
},
"subtitle_params": {
"font_family": "sans_serif",
"font_size": 28,
"color": "off_white",
"border_color": "light_pink",
"border_width": 1,
"position": {
"x": "center",
"y": 0.85,
"margin_bottom": 30,
"margin_left": 20,
"margin_right": 20
}
}
}
```
### 简约科技风模板(智能手环)
```json
{
"template_name": "minimalist_tech_watch",
"description": "简约科技风智能手环视频(30秒)",
"video_params": {
"resolution": "1080P",
"aspect_ratio": "16:9",
"duration": 30,
"fps": 25,
"output_format": "MP4"
},
"transition_params": {
"type": "tech_glow",
"duration": 0.4
},
"audio_video_sync": {
"enabled": true,
"max_offset": 0.1
},
"quality_params": {
"hd_noise_reduction": true,
"color_correction": "cool_tone"
}
}
```
### 实现建议
```python
# 伪代码示例
class TemplateManager:
"""API参数模板管理器"""
def __init__(self):
self.templates = {
"healing_beauty_mask": self.healing_beauty_mask_template,
"minimalist_tech_watch": self.minimalist_tech_watch_template
}
def get_template(self, template_name):
"""获取模板"""
if template_name not in self.templates:
raise ValueError(f"模板不存在:{template_name}")
return self.templates[template_name]
```
## 4. 字幕参数可视化校验
### 目标
在技能中添加简单的字幕参数预览功能,质检智能体可直接预览字幕位置/格式,避免因参数错误导致的合成后字幕遮挡
### 实现建议
```python
# 伪代码示例
def preview_subtitle(subtitle_params, image_size=(1080, 1920)):
"""
预览字幕参数
:param subtitle_params: 字幕参数
:param image_size: 画面尺寸
:return: 字幕位置和尺寸信息
"""
# 解析字幕参数
text = subtitle_params['text']
font_size = subtitle_params['font_size']
position = subtitle_params['position']
# 计算字幕位置
if position['x'] == 'center':
x = image_size[0] // 2
else:
x = int(image_size[0] * position['x'])
if 'y' in position:
y = int(image_size[1] * position['y'])
elif 'margin_bottom' in position:
y = image_size[1] - position['margin_bottom']
# 估算字幕尺寸
text_width = len(text) * font_size * 0.6 # 估算
text_height = font_size * 1.2
# 校验是否遮挡核心元素
# 这里需要结合实际核心元素位置信息
return {
'text': text,
'position': (x, y),
'size': (text_width, text_height),
'is_safe': True # 是否安全(不遮挡核心元素)
}
```
## 5. 合成异常兜底
### 目标
为COZE API调用添加"合成失败备用方案",若3次重试仍失败,直接将"合格图片+字幕参数+音效素材"打包输出,支持人工在COZE平台手动上传合成
### 实现建议
```python
# 伪代码示例
def api_video_synthesis_with_fallback(images, subtitle_params, audio_params, output_path):
"""
API视频合成(带异常兜底)
:param images: 图片列表
:param subtitle_params: 字幕参数包
:param audio_params: 音效素材包
:param output_path: 输出路径
:return: (success, output_path_or_fallback_package)
"""
max_retries = 3
timeout = 10 # 秒
for attempt in range(max_retries):
try:
# 调用COZE API合成
result = call_coze_api(images, subtitle_params, audio_params, timeout=timeout)
if result['success']:
# 合成成功
save_video(result['video_url'], output_path)
return (True, output_path)
except Exception as e:
print(f"API合成失败(尝试{attempt + 1}/{max_retries}):{e}")
if attempt < max_retries - 1:
# 重试
time.sleep(2)
continue
else:
# 达到最大重试次数,触发兜底方案
print("已达到最大重试次数,触发兜底方案")
fallback_package_path = output_path.replace('.mp4', '_fallback_package.zip')
create_fallback_package(images, subtitle_params, audio_params, fallback_package_path)
print(f"已生成兜底包:{fallback_package_path}")
print("请手动在COZE平台上传素材进行合成")
return (False, fallback_package_path)
def create_fallback_package(images, subtitle_params, audio_params, output_path):
"""
创建兜底包
"""
import zipfile
with zipfile.ZipFile(output_path, 'w') as zipf:
# 添加图片
for i, image in enumerate(images):
zipf.write(image, f"images/{os.path.basename(image)}")
# 添加字幕参数
subtitle_json_path = "subtitle_params.json"
with open(subtitle_json_path, 'w', encoding='utf-8') as f:
json.dump(subtitle_params, f, ensure_ascii=False, indent=2)
zipf.write(subtitle_json_path, "subtitle_params.json")
# 添加音效素材
for i, audio in enumerate(audio_params['sound_effects']):
zipf.write(audio['file_path'], f"audio/{os.path.basename(audio['file_path'])}")
# 添加背景音乐
if 'background_music' in audio_params:
zipf.write(audio_params['background_music']['file_path'], f"audio/{os.path.basename(audio_params['background_music']['file_path'])}")
# 添加README
readme_content = """
# 兜底包使用说明
本包包含视频合成所需的全部素材,请手动在COZE平台上传进行合成。
## 文件清单
- images/:分镜图片(按镜头序号排序)
- audio/:音效素材(背景音乐+场景音效)
- subtitle_params.json:字幕参数包(COZE API格式)
## 使用步骤
1. 登录COZE平台
2. 新建视频合成任务
3. 上传分镜图片(按镜头序号排序)
4. 上传音效素材
5. 复制字幕参数包内容到字幕参数配置
6. 配置API参数(1080P、9:16、60秒、25fps、柔焦渐变转场、音画同步、高清降噪)
7. 启动合成
注意事项:
- 确保素材格式正确(图片:JPG/PNG,音频:MP3)
- 确保字幕参数包格式正确(COZE API格式)
- 确保API参数配置正确
"""
with open("README.md", 'w', encoding='utf-8') as f:
f.write(readme_content)
zipf.write("README.md", "README.md")
```
## 总结
### 核心落地建议
1. **技能端添加关键词屏蔽**:检测到本地依赖包关键词立即阻断流程
2. **素材格式固化**:预设COZE API优选的素材格式
3. **API参数预设模板**:将典型视频的API合成参数保存为模板
4. **字幕参数可视化校验**:添加字幕参数预览功能
5. **合成异常兜底**:API调用失败后生成兜底包,支持人工手动合成
### 核心原则
- **API合成强制**:唯一调用COZE平台视频大模型API完成视频合成
- **流程阻断规则**:若出现本地依赖包调用行为,立即终止并切换至纯COZE API合成流程
- **API适配优先**:所有输出物均按COZE视频大模型API要求定义格式
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| receiving-code-reviewobra/superpowersUse when receiving code review feedback, before implementing suggestions, especially if feedback seems unclear or technically questionable - requires technical rigor and verification, not performative agreement or blind implementation | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| requesting-code-reviewobra/superpowersUse when completing tasks, implementing major features, or before merging to verify work meets requirements | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| using-git-worktreesobra/superpowersUse when starting feature work that needs isolation from current workspace or before executing implementation plans - ensures an isolated workspace exists via native tools or git worktree fallback | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| verification-before-completionobra/superpowersUse when about to claim work is complete, fixed, or passing, before committing or creating PRs - requires running verification commands and confirming output before making any success claims; evidence before assertions always | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |