Skip to content
Back to search

fpv-immersive-video-prompting

Route-first FPV AI video prompting skill for Seedance, Kling, Runway, Veo, GPT Image asset packs, numbered stops, and red-line path control.

Safety scan

A static text scan, not a guarantee. Select the badge to see what was scanned and every finding.

Checked today
Badge for your README
findsafeskills safety rating

Markdown

[![findsafeskills safety rating](/api/listings/zhouwei713--fpv-immersive-video-prompting/badge.svg)](/listings/zhouwei713--fpv-immersive-video-prompting)

HTML

<a href="/listings/zhouwei713--fpv-immersive-video-prompting"><img src="/api/listings/zhouwei713--fpv-immersive-video-prompting/badge.svg" alt="findsafeskills safety rating" /></a>

Install

i

Install instructions from the README

As written by the repo's author. findsafeskills has not verified these steps.

方式一:安装到 Hermes

git clone https://github.com/zhouwei713/fpv-immersive-video-prompting.git \
  ~/.hermes/skills/creative/fpv-immersive-video-prompting

重启 Hermes,或开启新会话后即可使用:

fpv-immersive-video-prompting

方式二:作为通用 Skill 使用

如果你不使用 Hermes,也可以直接把 `SKILL.md` 放进 Claude、Codex、Cursor、OpenCode 或其他支持 Skill / long prompt 的 Agent 环境里。

方式三:只当提示词方法论参考

直接阅读:

SKILL.md
skill/references/gpt-image-asset-packs.md
skill/references/session-patterns.md

---

What was scanned

The files the safety scan read, as written by the repository's author and shown as plain text. Lines with a finding are marked.

SKILL.md · Manifest · 235 lines

View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)

  1. ---
  2. name: fpv-immersive-video-prompting
  3. description: Use when the user wants to turn a static scene, character images, aerial map, drawn path, or route-control image into an immersive first-person FPV AI video prompt, especially Seedance/Kling/Runway/Veo style one-shot videos with numbered stop markers, red-line path control, world-map flythroughs, camera route planning, variable character counts, non-human POVs such as drones, pets, robot vacuums, character references, timed interactions, dialogue, spatial audio, and negative constraints.
  4. version: 1.0.0
  5. author: Hermes Agent
  6. license: MIT
  7. metadata:
  8. hermes:
  9. tags: [video, ai-video, fpv, prompting, camera-route, character-consistency]
  10. related_skills: [gpt-image-2-prompting]
  11. ---
  12. # FPV Immersive Video Prompting
  13. ## Overview
  14. This skill turns a scene idea plus optional image references into a directed FPV AI-video prompt. The target output should feel like the viewer is moving through a playable scene, not watching a generic beauty shot.
  15. Use this for image-to-video workflows where the user wants:
  16. - one-shot first-person walkthroughs
  17. - numbered route stops or camera planning
  18. - characters reacting as the camera approaches
  19. - human or non-human POVs: guest, pet cat, drone, robot vacuum, bird, vehicle, object, spirit
  20. - variable numbers of characters or targets
  21. - strong identity and route consistency
  22. ## Core Principle
  23. Default to Chinese for all generated prompts, including GPT Image asset prompts, route-control image prompts, video prompts, negative prompts, and delivery notes, unless the user explicitly asks for English or a target tool requires English. Keep technical terms such as FPV, Seedance, Kling, Runway, Veo, one-shot, prompt, and path control in English when they are clearer.
  24. Write the output as a small playable scene specification:
  25. 1. Who or what is the camera?
  26. 2. Where does it start?
  27. 3. How many main targets are there exactly?
  28. 4. What numbered order does it visit them in?
  29. 5. How does the camera physically move?
  30. 6. What happens at each stop?
  31. 7. What must stay consistent?
  32. 8. What must never appear?
  33. ## Default Asset Workflow
  34. When the user needs GPT Image / GPT-Image-2 assets, request a multi-image asset pack rather than one crowded contact sheet. Treat requests for “生图 prompt”, “首帧图”, “参考图”, “素材包”, “用 GPT Image 生成”, or any image-prep step as an asset-pack request, not as a single first-frame prompt.
  35. For any case that needs multiple images, output one batch-generation prompt that asks GPT Image / ChatGPT to generate multiple separate images in a single response. Do not split the asset pack into many independent per-image prompts unless the user explicitly asks for that. The prompt must say: “请一次性生成 [X] 张独立图片,不要生成拼图、九宫格、contact sheet 或一张图里塞多个画面。”
  36. For close-interaction scenes with N main people/targets, always output the complete asset pack unless the user explicitly asks for only one image:
  37. 1. First-frame scene image with small numbered stop markers only: 1, 2, 3, ... N
  38. 2. N separate character/reference images, one for each main person/target
  39. 3. Optional clean first frame without numbers for safer image-to-video input
  40. Before returning any image prompt, run this count check: total images = 1 route/numbered first frame + N character references + optional 1 clean first frame. The video prompt must refer to these images by role, e.g. “图片 1 is route planning, 图片 2-4 are character references, 图片 5 is clean first frame.”
  41. Default image sets:
  42. Close-interaction / character mode:
  43. 1. First-frame scene image with small numbered stop markers only: 1, 2, 3, ...
  44. 2. One character/reference image for each main target
  45. 3. Optional clean first frame without numbers for safer image-to-video input
  46. World-route / path-control mode:
  47. 1. Aerial route-control map or wide first frame with one clear continuous red path from start to destination
  48. 2. Optional clean world/scene reference without the red path
  49. 3. Optional landmark or character references only if specific destinations need consistent appearance
  50. Prefer numbered stop markers for close-range character interactions, indoor scenes, crowded social scenes, and exact target-count workflows. GPT Image often struggles to draw one continuous physically coherent route through furniture, people, walls, railings, or water; a bad red line can mislead the video model. Numbered stops are easier to generate correctly, and the video prompt can define the movement between them.
  51. Use red-line path control for large-scale route-shaped scenes: aerial world maps, fantasy continent journeys, city-to-landmark flythroughs, racing lines, canyon/drone routes, open-world game trailers, and Seedance 2.0 path-control demos. In that mode, the route image is a planning artifact; the final video must remove all red lines, arrows, annotations, labels, and map-view appearance.
  52. ### Red-Line Path Control Rules
  53. Use this mode when the user asks for Seedance 2.0 path control, a drawn route, an aerial map flythrough, a game-world traversal, a racing line, or a large-scale drone/bird/spirit FPV journey.
  54. For the route-control image prompt:
  55. - Use a high-resolution 16:9 aerial terrain map, world map, tactical map, or wide route-planning scene.
  56. - Draw one clear continuous red route line, optionally with a subtle arrow, from start to destination.
  57. - Make the route physically plausible through visible corridors: roads, valleys, rivers, city gates, bridges, canyons, rooftops, airspace, tunnels, coastlines, or ridgelines.
  58. - Build 4-6 visually distinct route segments so the video has progression: peaceful start, transition zone, landmark/city, danger zone, final destination.
  59. - Keep the red line clean and legible. Avoid multiple competing paths unless the user explicitly wants branching routes.
  60. For the video prompt:
  61. - Say the uploaded image is a route-planning map/image, not the final look.
  62. - Say the red line/arrow/annotations are only camera-path controls and must be completely removed.
  63. - State the camera must strictly follow the drawn route geometry from start to destination.
  64. - Use invisible first-person drone, bird, spirit, vehicle, or mounted-camera POV unless the route is truly walkable.
  65. - Add timeline segments by geography, not by character stop.
  66. - Include camera language: continuous forward motion, natural banking, close passes, altitude changes, foreground parallax, progressive acceleration, smooth horizon control.
  67. - Avoid map view, visible red lines, annotations, teleporting, reverse motion, jump cuts, visible drone, guide characters, flat terrain, deformed landmarks, flickering structures.
  68. Use numbered stops instead when the video depends on close character interactions, exact character count, indoor navigation, or object-level continuity.
  69. ### Numbered Stop Marker Rules
  70. For the first-frame image prompt:
  71. - Do not draw red route lines, connecting lines, or arrows by default.
  72. - Use small numbered markers near each target or on the floor beside each target.
  73. - Markers must be ordered by reachable spatial depth, not randomly scattered.
  74. - Marker 1 should usually be closest to camera; later markers move deeper into the scene.
  75. - Each next stop must be physically reachable from the previous one without crossing walls, water, furniture, railings, people, or other obstacles.
  76. - The scene must be designed around a walkable/glideable/flyable route before placing characters.
  77. Only use a continuous red route line if the user explicitly asks for it or the specific workflow handles route lines reliably.
  78. ## Variable Count Handling
  79. If the user specifies a number of people/targets, build the whole asset plan, route, timeline, and video prompt around exactly that number. Do not default back to five.
  80. Recommended target counts:
  81. | Duration | Best count | Handling |
  82. | --- | --- | --- |
  83. | 5-8s | 1-3 targets | one clear approach, few actions |
  84. | 8-12s | 3-4 targets | short interactions, one final reveal |
  85. | 12-15s | 4-5 targets | ideal for sequential stops |
  86. | 15-25s | 5-8 targets | shorter beats and group transitions |
  87. | 25s+ | 8+ targets | split into groups/zones |
  88. If the user says “a group” without a number, choose a practical default based on duration, usually 4-5 main interaction targets. Background extras may exist only if requested and must remain secondary.
  89. Always write:
  90. “exactly [N] main characters/targets” and “do not add or remove main targets.”
  91. ## POV Identity Selection
  92. FPV does not always mean normal human eye level. If the user specifies the POV, honor it. If not, choose a POV that makes the scene more coherent or interesting.
  93. | Scene / intent | Good POV choices | Movement cues |
  94. | --- | --- | --- |
  95. | palace, mansion, apartment, party | invited guest, butler, courier, pet cat | walking sway, room turns, hand/paw glimpses |
  96. | modern living room, cafe, cozy home | pet cat, robot vacuum, visitor | low-angle gliding, furniture legs, paws/tail or wheel hum |
  97. | outdoor landscape, city, battlefield | drone, bird, kite, spirit | aerial arcs, height changes, wind, no footstep bob |
  98. | museum, exhibition, showroom | visitor, guide robot, security camera | slow walkthrough, display pauses, precise turns |
  99. | horror / mystery | child, flashlight holder, CCTV, ghost | low height, hesitant movement, limited view |
  100. | car/train/boat/aircraft | passenger, driver, mounted camera | constrained path, window motion, engine/road/water sounds |
  101. POV constraints:
  102. - Human: eye-level height, walking bob, breathing, hands/sleeves may appear.
  103. - Pet cat/dog: low height, curious head turns, paws/tail may appear, no human hand gestures.
  104. - Drone: smooth continuous flight, altitude changes, banking turns, propeller/wind sound, no walking bob.
  105. - Robot vacuum/small robot: very low height, slow floor gliding, furniture legs loom large, mechanical hum, cannot climb stairs or jump.
  106. - Bird/insect/spirit: floating arcs or gliding, unusual height, but still continuous and motivated.
  107. - Object POV: movement must follow the object’s plausible motion unless the concept is surreal.
  108. Always enforce physical limits. A cat cannot float over a table; a drone cannot pass through walls; a robot vacuum cannot jump onto a sofa; a human cannot squeeze through furniture.
  109. ## Video Prompt Structure
  110. Default structure:
  111. 1. Input usage: first frame, numbered route reference, character references, or red-line route-control image
  112. 2. Specs: aspect ratio, duration, one-shot
  113. 3. POV identity and physical movement
  114. 4. Route order: numbered target stops for close interaction, or geography segments for red-line path control
  115. 5. Compact timeline with one beat per target or one beat per route segment
  116. 6. Environment motion and sound design if useful
  117. 7. Consistency and avoid list
  118. Keep the final video prompt directly copyable. Do not impose a fixed character limit unless the user asks for one.
  119. ## Template: Video Prompt
  120. ```text
  121. 使用上传图片作为首帧、编号路线参考和 [N] 个角色/目标外观参考,生成一段 [比例]、[时长] 秒、一镜到底的 [场景类型] [POV身份] FPV 视频。首帧中的编号 [1..N] 只作为镜头停靠顺序参考,不要出现在最终画面里。最终视频不要出现数字、编号、路线、箭头、文字标签或 UI。
  122. 观众/摄像机是一位/一台 [POV身份],从 [起点] 出发,严格按编号顺序移动:[起点] → 1 [目标1] → 2 [目标2] → ... → [终点/回望]。本视频包含 exactly [N] 个主要人物/目标:[列出名称],不要增加或减少主目标。角色脸、服装、位置、身份保持参考图一致。
  123. 镜头运动必须符合 [POV身份] 的物理限制:[低机位/步行/飞行/滑行/车载等具体运动];移动连续,有真实加速减速,不能瞬移、跳切、穿过障碍物或变成其他 POV。
  124. 时间轴:
  125. 0.00-[入口时间]:[建立起点和环境动态]
  126. [时间]-[时间]:移动到 1,[目标1动作/台词/反应]
  127. [时间]-[时间]:移动到 2,[目标2动作/台词/反应]
  128. ...
  129. [结尾时间]:到达 [终点] 或回望全景,形成空间层次。
  130. 环境动态:[水波/窗光/咖啡蒸汽/灯光/风/反光/布料/屏幕等]。
  131. 避免:编号残留、路线/箭头、换脸、同脸、身份漂移、主目标数量变化、穿模、跳切、瞬移、错误 POV、过度抖动、场景漂移、低质、畸形、水印。
  132. ```
  133. ## Common Pitfalls
  134. 1. Asking GPT Image for a continuous red route line by default.
  135. - Prefer numbered stop markers. Continuous lines often become disconnected, cross obstacles, or contaminate the video.
  136. 2. Letting numbered stops scatter randomly.
  137. - First define a walkable/glideable/flyable spatial route, then place numbered stops along it.
  138. 3. Ignoring the user’s specified count.
  139. - Use exactly the requested number in assets, route, timeline, and constraints.
  140. 4. Treating every POV like a human.
  141. - Robot vacuums need floor gliding; cats need low curious movement; drones need flight; humans can show hands and walking sway.
  142. 5. Giving every person full dialogue in high-count scenes.
  143. - For more than 6 targets, use groups, gestures, glances, or quick beats.
  144. 6. Forgetting to remove markers in the final video.
  145. - Always say numbers/markers are planning references only and must not appear in final output.
  146. 7. Forgetting that red-line path control and numbered stops solve different problems.
  147. - Numbered stops are better for characters and interiors; red-line route control is better for world-scale flythroughs, racing lines, and map-to-world journeys.
  148. ## Verification Checklist
  149. Before returning the prompt, check:
  150. - [ ] Aspect ratio and duration are specified
  151. - [ ] First frame / route reference / character references are assigned
  152. - [ ] Main target count is explicit and matches the user request
  153. - [ ] POV identity is explicit or reasonably inferred
  154. - [ ] POV movement physics match the identity
  155. - [ ] Route order is explicit and physically continuous
  156. - [ ] Timeline covers the full duration and fits the target count
  157. - [ ] Each target has an action/reaction, or group beats for high-count scenes
  158. - [ ] Character/object consistency constraints are repeated
  159. - [ ] Environment secondary motion is included where useful
  160. - [ ] Negative constraints mention no route lines/arrows/numbers/markers in final video
  161. - [ ] Negative constraints mention no added/removed main targets
  162. - [ ] Red-line path-control mode is used only when the scene is route-shaped and large-scale, or the user explicitly requests it
  163. - [ ] Generated prompts are in Chinese by default unless the user explicitly requested English or a target tool requires English
  164. - [ ] Final prompt is directly copyable
  165. ## References
  166. Read `references/session-patterns.md` for examples and session-specific lessons from the initial development of this workflow.
  167. Read `references/mayz-seedance-world-route-case.md` for the Seedance 2.0 red-line world-route pattern: aerial map control image, strict drawn route geometry, map-to-world flythrough, biome progression, speed-run/racing-line mode, and theme-park ride style journeys.
  168. Read `references/public-article-angle.md` when explaining this workflow publicly in a WeChat article, X thread, tutorial intro, or demo write-up. It captures the session framing: FPV prompts are action trajectories, not just visual descriptions; numbered stops vs red-line path control; space-first design; POV physics; count-duration tradeoffs.
  169. Read `references/cafe-cat-numbered-stops-case.md` when the user asks for cafe, cozy indoor, pet-cat POV, or low-angle numbered-stop interaction scenes. It captures a 15-second / 3-person cafe route pattern and the cat-specific movement constraints that prevent drone-like motion.
README.md · README · 277 lines

View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)

  1. <div align="center">
  2. # FPV 运镜导演.skill
  3. > *把 AI 视频提示词,从“画面描述”升级成“行动轨迹设计”。*
  4. [![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](LICENSE)
  5. [![Agent Skills](https://img.shields.io/badge/Agent%20Skills-Standard-green)](https://agentskills.io)
  6. [![Runtime](https://img.shields.io/badge/Runtime-Hermes%20·%20Claude%20·%20Codex%20·%20Cursor-blueviolet)](#安装)
  7. [![AI Video](https://img.shields.io/badge/AI%20Video-Seedance%20·%20Kling%20·%20Runway%20·%20Veo-ff69b4)](#适合什么场景)
  8. <br>
  9. **一个专门为 FPV / image-to-video 设计的 AI 视频提示词 Skill。**
  10. 它不只帮你写“电影感、浅景深、光影高级”。
  11. 它会先设计镜头是谁,从哪里出发,按什么顺序经过哪些人,怎么绕过障碍物,最后停在哪里。
  12. <br>
  13. [看效果](#效果示例) · [安装](#安装) · [它解决什么](#它解决什么) · [工作原理](#工作原理) · [资产包](#gpt-image-资产包) · [English](README_EN.md)
  14. </div>
  15. ---
  16. ## 效果示例
  17. ### 咖啡厅猫咪视角
  18. 用户输入:
  19. ```text
  20. 帮我做一个咖啡厅里 3 个人依次互动的 FPV 视频提示词,15 秒,猫咪视角
  21. ```
  22. Skill 会先判断这不是红线路径场景,而是近距离人物互动。它会生成一套完整资产包,而不是只给一张首帧图。
  23. ```text
  24. 图片 1:咖啡厅猫咪低机位首帧,带 1、2、3 小编号停靠点
  25. 图片 2:靠窗女生独立人物参考图
  26. 图片 3:吧台咖啡师独立人物参考图
  27. 图片 4:角落圆桌男生独立人物参考图
  28. 图片 5:干净版首帧,去掉编号,用作真正视频首帧
  29. ```
  30. 然后输出视频 prompt:
  31. ```text
  32. 观众是一只在咖啡厅里自由走动的猫咪,镜头保持接近地面的低机位,
  33. 从咖啡厅入口旁的地垫开始,严格按编号顺序移动:
  34. 入口地垫 → 1 靠窗座位的女生 → 2 吧台前的咖啡师 → 3 角落圆桌旁的男生 → 窗边阳光下停住。
  35. 全片包含 exactly 3 个主要人物,不要增加或减少主目标。
  36. 镜头运动必须符合猫咪的身体限制,能看到桌腿、椅腿、地面纹理、人的鞋子,
  37. 可以短暂停顿、好奇转头、绕开障碍物,但不能飞,不能跳上吧台,不能穿过桌椅和人腿。
  38. ```
  39. 这类提示词的重点不是“咖啡厅很漂亮”,而是让视频模型知道猫到底怎么走。
  40. ---
  41. ### 世界地图飞行
  42. 用户输入:
  43. ```text
  44. 我想做一个 Seedance 2.0 的世界地图飞行,从雪原穿过峡谷、王城,最后到火山。
  45. ```
  46. Skill 会切换到红线路径控制模式。
  47. ```text
  48. 图片 1:16:9 奇幻大陆航拍路线规划图,一条连续红线从雪原出发,经过峡谷和王城,抵达火山口
  49. 图片 2:可选干净世界参考图,去掉红线
  50. ```
  51. 视频 prompt 会明确说明:
  52. ```text
  53. 红色路线只作为摄像机路径控制,不是最终画面内容。
  54. 最终视频不要出现红线、箭头、地图标注、文字标签、UI 或俯视地图感。
  55. 镜头必须严格沿红线几何飞行,有自然 banking、贴近地形掠过、穿越地标时的前景视差和稳定地平线。
  56. ```
  57. 同样是 FPV,一个是猫在咖啡厅里走,一个是无人机穿越大陆。两种场景不能用同一套提示词。
  58. ---
  59. ## 它解决什么
  60. 很多 AI 视频 prompt 看起来很完整,实际一生成就翻车。
  61. 常见问题是:
  62. - 镜头突然瞬移,空间断了
  63. - 人物数量一会儿多一会儿少
  64. - 首帧里的编号、红线、箭头残留在成片里
  65. - 说是猫咪视角,结果变成无人机视角
  66. - 说是一镜到底,结果中间跳切
  67. - 室内路线穿过桌子、椅子、墙和人腿
  68. - 多人物互动里,角色脸和衣服互相漂移
  69. 原因通常不是缺少风格词,而是缺少行动轨迹。
  70. FPV 视频要写清楚的不是一张图,而是一段运动:谁在看,怎么走,经过谁,在哪里停,什么东西不能变。
  71. ---
  72. ## 适合什么场景
  73. - 咖啡厅、客厅、展厅、庭院、宫殿等室内一镜到底
  74. - 3 到 8 个角色依次互动的短视频
  75. - 猫、狗、机器人吸尘器、无人机、鸟、幽灵、车辆等非人类 POV
  76. - GPT Image / GPT-Image-2 首帧和人物参考图资产包
  77. - Seedance、Kling、Runway、Veo 等 image-to-video 工作流
  78. - 红线路径控制、世界地图飞行、城市到地标、峡谷穿越、赛车线路
  79. - 想把提示词从“视觉描述”变成“镜头调度”的创作者
  80. ---
  81. ## 工作原理
  82. 这个 Skill 会把一个 FPV 视频拆成 8 个问题。
  83. ```text
  84. 1. 摄像机是谁
  85. 2. 从哪里开始
  86. 3. exactly 有多少个主要人物或目标
  87. 4. 按什么顺序经过它们
  88. 5. 每一段路线是否物理可达
  89. 6. 每个停靠点发生什么互动
  90. 7. 哪些身份、服装、位置必须保持一致
  91. 8. 哪些东西绝对不能出现在最终画面里
  92. ```
  93. 然后它会选择两种路线模式之一。
  94. ### 编号停靠点
  95. 适合近距离人物互动、室内空间、咖啡厅、客厅、派对、展厅。
  96. 这类场景里不要默认画红线。红线很容易穿过桌椅、墙面和人腿,也容易残留在视频里。
  97. 更稳定的做法是,在首帧里放小编号 1、2、3,把角色顺序标清楚。真正的移动路线交给视频 prompt 约束。
  98. ### 红线路径控制
  99. 适合大世界路线、航拍地图、城市飞行、峡谷穿越、赛车线路、Seedance 2.0 path control。
  100. 这类场景的核心是路线几何。红线可以作为路径控制,但最终视频里必须完全消失。
  101. ---
  102. ## GPT Image 资产包
  103. 如果场景里有 N 个主要人物,Skill 默认生成完整资产包。
  104. ```text
  105. 1 张带编号首帧图
  106. N 张人物独立参考图
  107. 1 张可选干净首帧图
  108. ```
  109. 以 3 人咖啡厅为例:
  110. ```text
  111. 图片 1:咖啡厅猫咪视角首帧,带编号 1、2、3
  112. 图片 2:靠窗女生参考图
  113. 图片 3:咖啡师参考图
  114. 图片 4:角落男生参考图
  115. 图片 5:干净版首帧,去掉编号和所有标记
  116. ```
  117. 这样做的目的很简单:首帧管空间,人物参考管身份,干净首帧管最终输入。
  118. 如果只给一张图,视频模型很容易把路线、角色和画面标记混在一起。
  119. ---
  120. ## 安装
  121. ### 方式一:安装到 Hermes
  122. ```bash
  123. git clone https://github.com/zhouwei713/fpv-immersive-video-prompting.git \
  124. ~/.hermes/skills/creative/fpv-immersive-video-prompting
  125. ```
  126. 重启 Hermes,或开启新会话后即可使用:
  127. ```text
  128. fpv-immersive-video-prompting
  129. ```
  130. ### 方式二:作为通用 Skill 使用
  131. 如果你不使用 Hermes,也可以直接把 `SKILL.md` 放进 Claude、Codex、Cursor、OpenCode 或其他支持 Skill / long prompt 的 Agent 环境里。
  132. ### 方式三:只当提示词方法论参考
  133. 直接阅读:
  134. ```text
  135. SKILL.md
  136. skill/references/gpt-image-asset-packs.md
  137. skill/references/session-patterns.md
  138. ```
  139. ---
  140. ## 使用方式
  141. 你可以这样说:
  142. ```text
  143. 帮我做一个咖啡厅里 3 个人依次互动的 FPV 视频提示词,15 秒,猫咪视角
  144. ```
  145. ```text
  146. 做一个宫殿庭院 FPV,一镜到底经过 5 个角色,最后到池塘边
  147. ```
  148. ```text
  149. 我有一张世界地图,想让镜头沿红线从雪原飞到火山,帮我写 Seedance 2.0 prompt
  150. ```
  151. ```text
  152. 给我一套 GPT Image 生图 prompt,要包含首帧图、人物参考图和干净版首帧
  153. ```
  154. ---
  155. ## 仓库结构
  156. ```text
  157. .
  158. ├── README.md
  159. ├── README_EN.md
  160. ├── SKILL.md
  161. ├── skill/
  162. │ ├── SKILL.md
  163. │ └── references/
  164. │ ├── gpt-image-asset-packs.md
  165. │ ├── liyue-ai-redline-fpv-case.md
  166. │ ├── mayz-seedance-world-route-case.md
  167. │ ├── public-article-angle.md
  168. │ └── session-patterns.md
  169. ├── examples/
  170. │ ├── numbered-stop-example.md
  171. │ └── redline-route-example.md
  172. └── LICENSE
  173. ```
  174. ---
  175. ## 设计理念
  176. FPV 提示词不是把一张图说得更美。
  177. 它更像在设计一个小关卡:入口在哪里,路线怎么走,角色站在哪里,镜头能不能绕过去,最后有没有一个让观众理解空间的停顿。
  178. 只要运动没设计清楚,模型就会替你设计。它设计出来的结果,通常就是瞬移、穿墙、换脸和跳切。
  179. 这个 Skill 的价值,是把这套判断流程固定下来,让每次生成 FPV 视频前都先过一遍空间、路线、角色和物理限制。
  180. ---
  181. ## English
  182. For English documentation, see [README_EN.md](README_EN.md).
  183. ---
  184. ## License
  185. MIT License.

Ratings and comments

Sign in to rate this skill.

Sign in

Free to use. If it’s helped, pass it on.