内容创作
#video
modular-film-prompt-bible
Use for high-control film and video prompting that must combine a stable world bible, reference placeholders, real-photography fidelity, sound policy, genre fusion, and shot-level execution into one reusable creative template.
DeepseekModel
官方收录技能
质量 优秀 · 90
v1.0.0
获取
https://deepseekmodel.com/api/download.php?id=nolanx-ai-nolanx-ai-nolanx-skills-modular-film-prompt-bible-skill-md&format=skill
下载 .skill
标准格式,含 system_prompt 与 model_config,导入任意 Agent 框架即可使用
.skill 文件中 system_prompt 字段的实际内容。
name modular-film-prompt-bible description Use for high-control film and video prompting that must combine a stable world bible, reference placeholders, real-photography fidelity, sound policy, genre fusion, and shot-level execution into one reusable creative template. tags ["film-prompt","modular-template","visual-bible","shot-execution","reference-lock","sound-design"] aliases ["cinematic-prompt-bible","影视创作模板","电影提示词圣经"] agents {"allow":["planner","script_writer","image_designer","image_edit_agent","video_designer","flf_video_designer"]} Modular Film Prompt Bible Use this skill when the user wants a film-style creation template rather than a loose prompt paragraph. This skill turns a concept into two layers: a stable creative bible that stays locked executable shot or frame instructions that only describe the delta for the current beat What This Skill Solves Prevents film prompts from collapsing into generic adjective piles. Separates identity, world, medium, sound, and shot execution into distinct modules. Supports reference placeholders such as {{Portrait 1}} , {{Scene 1}} , @image 1 , or @video 1 . Works for both still-image key art and shot-by-shot video prompting. Core Principle Write prompts in this order: identity lock world and genre lock medium and realism lock sound policy per-shot execution continuity carryover Do not invert this order. If the camera language appears before the world is stable, the output tends to drift. Layer 1: Creative Bible Lock the parts that should not randomly mutate. Identity Lock Map each recurring subject to a stable reference token. Lock silhouette, age cues, species or mechanical type, facial system, costume logic, prop package, and material response. For non-human or mechanical faces, define expression states explicitly instead of writing vague emotion words. If a recurring prop exists, lock carried side, scale, outline, glow behavior, and wear state. World And Genre Lock Name the world as a specific place after a specific event, not as a mood cloud. If the brief fuses genres, define the blend operationally. Example: atompunk + western means period-tailored silhouettes, retro-futurist industrial forms, aged metal, analog control surfaces, and dusty frontier staging. State what must not modernize. Medium And Realism Lock Choose one image logic: live-action realism, large-format film, 35mm genre cinema, retro stock, premium CG with film behavior, or another single coherent medium. If realism matters, explicitly block game-like CG, plastic surfaces, floating props, and weightless movement. Describe how metal, leather, skin, smoke, dust, rain, LEDs, and water respond to light. Tie grain, halation, contrast, bloom, and motion blur to dramatic intent rather than using them as decoration. Sound Policy Decide whether the piece uses sync sound only, sparse ambience, or score-led treatment. If the request wants realism, prefer no BGM and retain only production sound, Foley, and environment. Every sound note should come from a visible or inferable source. Layer 2: Execution Prompt Once the bible is locked, describe the current frame or shot. Still Image Mode Use this section order: core mood and fidelity character or subject environment composition color and lighting style finish micro-action Video Shot Mode For each shot or time slice, define: shot objective shot size and angle composition camera position and movement subject movement expression or display-state behavior environment interaction sound in shot continuity from previous shot Do not restate the entire world bible on every shot. Reuse the lock and only describe what changes. Chinese-Friendly Section Map When the user works in Chinese, these headings map cleanly to this skill: 基础设定 : identity, world, props, genre fusion 核心氛围与画质 : realism level, capture medium, texture, grain, contrast 声音 : sync sound, ambience, Foley, BGM policy 人物 : subject lock and performance traits 场景氛围 : environment state, weather, aftermath, practical light sources 构图 : framing, dominant screen position, foreground-midground-background logic 色彩与光影 : palette, key direction, shadow density, practical motivation 画面风格 : film era, lens feel, realism guardrails 动态描述 : body behavior, gaze, servo motion, cloth or smoke response 景别 / 运镜 / 画面内容 : shot execution fields for video Performance Rules A body move must mean something: advance, recoil, stalk, circle, freeze, brace, collapse, hesitate. For mechanical characters, translate emotion into readable posture, timing, and interface-state changes. If the face is LED, masked, or stylized, define emotional beats as discrete transitions, not fuzzy mood prose. Camera Rules Choose one dominant camera intention per shot. Separate subject motion from camera motion. If the world is oppressive, use framing and depth to trap the subject. If the beat is revelatory, use push-in, tightening composition, or exposure shift with intent. If the beat is absurd, eerie, or tragicomic, let performance and framing carry it before adding style embellishment. Continuity Rules Carry forward damage state, dirt, weather, costume wear, prop orientation, and action phase. If a shot continues a previous moment, define what is inherited before defining what changes. If an interface state changes, note the exact trigger and resulting state. Guardrails Do not write "cinematic" unless you convert it into lens, light, composition, texture, and motion decisions. Do not mix incompatible era signals without explaining the fusion logic. Do not overload a shot with multiple contradictory camera verbs. Do not let sound notes drift away from visible action. Do not flatten a strong concept into generic beauty-shot language. Compact Output Pattern When building a reusable film prompt package, prefer this structure: Core premise: Recurring references: World and genre lock: Sound policy: Look bible: Shot 1: objective / framing / camera / action / sound / continuity Shot 2: objective / framing / camera / action / sound / continuity Best Fit Cases a screenshot-derived film prompt that mixes world bible and shot list robot, creature, armor, or stylized face systems that need explicit expression states genre-fusion work such as atompunk western, diesel noir, wuxia steampunk, retro-future war drama prompts that must feel like a director treatment rather than a casual text description
Agent 识别该技能的关键词,点击任意一个即可复制。
该技能未提供触发词。
下载的 .skill 包内含以下字段。
| 字段 | 说明 |
|---|---|
| format | 格式标识(skill/v1) |
| skill_id | 技能唯一 ID |
| name | 技能名称 |
| version | 版本号 |
| description | 技能描述 |
| category | 所属分类(数组) |
| trigger_words | 触发词列表 |
| tags | 标签列表 |
| source | 来源标识 |
| source_url | 来源链接(本页地址) |
| exported_at | 导出时间(每次下载生成) |
| system_prompt | 系统提示词正文 |
| model_config | 模型参数:provider / model / temperature / max_tokens / top_p |
| examples | 示例 |
| install_guide | 各平台导入说明(Coze / Dify / Claude / 自定义框架) |