agentleFS
weibo-jobweibo-job/guagua-animation-skills/visual-comedy-director/SKILL.md

visual-comedy-director

Skill2 starsChanged 11 days agoMITView on GitHub

把无对白动画故事拆成可生成、可剪辑的连续镜头表,控制喜剧节奏、空间关系与时间轴。

The file

Files
SKILL.md

视觉喜剧分镜

输入与交付

读取story.md和characters.json;输出shots.json,其顶层为episode_id、fps、target_duration_s、shots。 每镜记录shot_id、start_s、duration_s、cast、scene、shot_type、composition、camera、action、start_state、end_state、gag、sound_cues、image_path、video_path、source_in_s、source_out_s、freeze_s、status、image_status、video_status。暂未生成的路径和取用区间为null;初建时三个状态均为pending。image_status和video_status分别由关键帧与视频环节维护,取值为pending、generating、needs_review、approved、retry_needed。文字分镜通过不代表媒体通过;两种媒体状态均approved且文件与取用区间有效,才将status设为approved。

导演规则

普通镜头约3—5秒,动作约2—3秒;70秒通常15—20镜,但以动作可读和总时长为准。优先从以下选择,不必每集用满:L1双人中近景、L2表情特写、L3前景小角色与背景威胁、L4第三方反应、L5草地一排探头、L6仰拍、L7包围、L8追逐全景。 保留系列双人惊讶开场,同时在3秒内展示异常。主要爆点可安排约10/30/55秒;适当预留0.5秒期待停顿及0.2—0.3秒反应定格,不机械给每个动作插停顿。末帧约0.5秒定格纳入总时长。 仰拍增强气势,俯拍突出弱小;紧张时可推近,但追逐路线和包围关系需要全景。写清轴线、视线目标、运动方向、道具持有者、姿态和损伤状态;切镜后保持连续,跨轴需有可读过渡。 一个短镜头只承担可执行的主要动作;复杂连续事件拆镜。参考图需能支持起始姿态,不能只用表情图要求复杂全身动作。

时间语义与验收

start_s和duration_s是最终成片区间,duration_s包含定格;默认硬切,转场额外记录transition及overlap_s并纳入总长,不能重复加时。各镜区间完整覆盖目标时长,无意外空隙或重复。画面可生成性不足时回改分镜,不能靠后期长时间静帧凑数。 交付scene-generator、animation-prompt和sound-designer同一份分镜;改动时通知受影响下游并更新入点。

项目约定

本技能服务双主角反差、无对白视觉喜剧系列;默认约70秒、9:16竖屏,结尾留钩子。用户的新要求优先。保持已确认角色与画风,不把示例角色强加给其他项目。 每集在用户项目目录保存产物;镜号采用S001等稳定编号,修改内容不随意改号。沿用episode.json清单:episode_id、theme、target_duration_s、character_version、story_path、shots_path、audio_path、output_path、status。首次单独调用时只建立实际需要的文件,不要求先完成全部流程。 路径必须指向真实文件;未完成用null或“待生成”,不用虚构链接。提示词、生成素材、审核通过、导出成片是不同状态。已获授权内继续工作,不重复确认;付费工具遵循实际工具报价与授权流程。

Download now (ZIP)

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
What did you use it for? e.g. Reviewing pull requests in a Rails app

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.