跳到正文
cccskills
免费在 GitHub 公开

Micro-drama Video

Develop a short scripted micro-drama from an idea or approved script through character sheets, storyboard frames, animated shots, and a reviewable assembly.

中文简介正在准备中。当前显示原文说明。

实测评分 未实测

查看安装方法

包含的文件(1)

  • SKILL.md4.3 KB

SKILL.md(原文)

安装前可以查看交给 AI 代理的指令内容。

Micro-drama Video

Mission

Turn a small premise or approved script into a coherent short narrative. Keep screenwriting, character extraction, storyboard design, keyframe approval, shot animation, and assembly as visible stages so a failed shot does not erase the story or silently change the user's intent.

Use Storyboard to Video for a general narrative sequence. Use this skill when the deliverable is specifically a short drama with a beginning, escalation, turn, and ending.

Required inputs

  • premise, treatment, or approved script;
  • genre, tone, audience, language, target duration, ratio, and shot count;
  • characters, relationships, locations, props, chronology, and visual style;
  • reference images and the role of each reference;
  • dialogue, narration, music, sound, captions, and end-card intent; and
  • rights/consent, variant count, budget, reviewer, and definition of done.

If the user gives only an idea, propose a compact story arc and label all invented dialogue, character facts, and connective scenes as proposals. Do not turn a real person's likeness, private relationship, or alleged event into fiction without authorization and clear synthetic-media labeling.

Production stages

1. Develop the story

Create a one-page story brief:

logline:
promise and audience:
setup:
escalation:
turn or reveal:
resolution / end frame:
dialogue and narration:
facts that must remain true:
rights and disclosure:

Keep the story small enough for the requested duration. A micro-drama should have observable beats, not a paragraph that asks one generation call to stage multiple cuts and time jumps.

2. Extract the cast and world

Create stable IDs for characters, products, locations, and props. Record identity anchors, wardrobe, age presentation, emotional state, screen direction, and forbidden changes. For supplied people, record authorization and the allowed use; do not infer relationships or age from a photograph.

3. Build and approve the storyboard

Return one row per shot with shot_id, story beat, action, framing, camera, location, character IDs, prop state, dialogue/audio, duration, reference roles, and acceptance checks. Generate keyframes only when needed, then review story geography, continuity, text, identity, and chronology before paying to animate the sequence.

4. Animate the selected shots

Resolve the current schema through MODELS.md and muapi-video-tools.md. Prefer image-to-video or reference-driven inputs when a keyframe carries identity, product, or world continuity. Upload each local reference once, preserve its role/order, retain prediction IDs, and poll every asynchronous job.

Do not let a model's successful status stand in for a usable shot. Keep failed, blocked, and rejected branches attached to their shot IDs. Change one cause at a time when iterating.

5. Assemble and review

Combine only approved clips in story order with an explicit audio policy. Keep dialogue, music, ambience, and captions as separate decisions when the connected tools do not expose mixing or text-rendering controls. Run video QA across the whole sequence, not only the poster frame.

Output contract

Return the story brief, cast/world bible, ordered storyboard, keyframe approval state, shot parent/child table, model/endpoint, prediction IDs, output URLs, assembly status, audio/caption assumptions, QA findings, failed branches, and the smallest next iteration. Save permitted plans, manifests, receipts, and QA notes under .video/ using reports-and-receipts.md.

This skill produces a draft. It does not publish, claim that generated footage documents a real event, or grant rights to music, likenesses, voices, brands, or locations.

评价

还没有评价。欢迎分享你的使用感受。

评论

无需登录即可发表。欢迎分享使用感受、提问,或说明在哪些环境下无法运行。

还没有评论。来写第一条吧。

提交后会立即公开。请勿填写邮箱、电话等个人信息。URL 最多 2 个。

0 / 1000 字

此表单受 reCAPTCHA 保护,适用 Google 的隐私权政策和服务条款。

同一仓库的技能

Plan draft video-ad variations from supplied performance evidence, preserving the distinction between observed data and creative hypotheses.

中文简介正在准备中。当前显示原文说明。

SamurAIGPT/open-ai-video-agent82026年8月28日 更新

Create authorized presenter, avatar, and UGC-style video drafts with explicit voice, likeness, consent, and synthetic-media handling.

中文简介正在准备中。当前显示原文说明。

SamurAIGPT/open-ai-video-agent82026年8月28日 更新

Analyze a script or existing edit, identify visual gaps, and generate supplemental B-roll coverage with clear parent and factual-status labels.

中文简介正在准备中。当前显示原文说明。

SamurAIGPT/open-ai-video-agent82026年8月28日 更新

Build a consented continuity system for an authorized person, fictional character, product, or world across video shots.

中文简介正在准备中。当前显示原文说明。

SamurAIGPT/open-ai-video-agent82026年8月28日 更新

Translate creative intent into coherent cinematography, framing, camera, lens, lighting, motion, and sound direction for a video shot.

中文简介正在准备中。当前显示原文说明。

SamurAIGPT/open-ai-video-agent82026年8月28日 更新

Create a non-explicit romantic scene from authorized adult references with clear consent, identity boundaries, and synthetic-media labeling.

中文简介正在准备中。当前显示原文说明。

SamurAIGPT/open-ai-video-agent82026年8月28日 更新

查看 SamurAIGPT 的全部技能

报告此技能的问题