生成@图片1前10秒的剧情，黑衣女刺客面部完全参考@图片2第一个场景：特写跟随镜头，女刺客的黑色布靴，脚尖点地；第二个场景：低角度全景，月光穿透林间薄雾，女刺客如离弦之箭从画面左侧腾空跃起，短刃直指将军咽喉；将军一手扼住她的脖颈；第三个场景：特写镜头，刺客面部特写，眼神凌厉却因窒息而泛红，牙关紧咬。需要运用到分镜和不同视角切换，让整个画面更有节奏感和电影感。

生成@图片1后10秒的剧情，黑衣女刺客掷出飞刀，飞刀划过将军喉咙，将军倒地，女子潇洒落地。随后，飞刀旋转回到黑衣女子手中，黑衣女子手持飞刀做出@图片2相同的姿势。全片需要运用到分镜和不同视角切换，让整个画面更有节奏感和电影感。仅生成打斗音效和环境的音效，不要配背景音乐。动作设计符合武侠片规范，动作干净利落。禁止出现血腥画面，无血腥暴力；不不良引导，无危险动作。（注意短刀的长度与参考图全程保持一致）

Seedance 2.0 Scene Continuation Video Prompt

A two-part prompt for Seedance 2.0 to generate a continuous video sequence based on a single reference image (Image 1). The first part generates the 10 seconds of action leading up to Image 1, featuring a female assassin in black, mimicking a famous scene (Arya Stark vs. Night King). The second part generates the 10 seconds following Image 1, showing the assassin defeating the general and striking a pose. Both parts require specific camera work, rhythm, movie-like quality, martial arts choreography, and sound effects, while strictly avoiding blood/gore and dangerous actions. The assassin's face and pose must reference a second image (Image 2).

模型