360 Secure SkillHub
360 安全技能中心

Hotel Lobby AI Prompt Director

可针对1-2名人、宠物等主体，创作原创橙色演播亭场景的表演提示与短分镜，适配Hotel Lobby风格的图片、文字、视频、动效参考类AI内容生成。流程前置版权校验，仅允许使用用户拥有合法使用权的素材，严禁复刻现有内容；支持视频/照片/文本/动作参考四种输入模式，可选MCP工具辅助创作校验，输出结构化结果，需匹配平台实时生成参数。

gpt-img-2
gpt-img-2
热度 16.67
最近更新 2026-10-09 21:03:23
分类：创作依赖：内部调用大模型 / MCP 服务应用场景：图像生成 / 视频编辑 / 社交媒体 / 模板

安装技能

复制下面的话术发给你的 Agent，它会通过 secure-skills 检索并安装该技能。

执行命令安装此 skill： npx secure-skills-cli add hotel-lobby-ai-prompt-director
Metadata
name
hotel-lobby-ai-prompt-director
description
Create original orange-booth performance prompts and short shot plans for one or two people, pets, or other subjects. Use for Hotel Lobby-style photo-to-video, image, text, or motion-reference concepts; require rights to all references and never reproduce an existing performance, song, person, setting, or sequence of frames.
metadata
{"openclaw":{"homepage":"https://aihotellobby.art/hotel-lobby-video","version":"1.0.0"}}

Hotel Lobby AI Prompt Director

Turn an authorized concept into a compact, original performance prompt. This Skill is a text-only workflow by default. Its optional MCP adds deterministic, read-only helpers and never uploads media, generates content, calls a provider, accesses an account, spends credits, or submits payments.

AI Hotel Lobby is independent and is not affiliated with or endorsed by COLORS, Quavo, Migos, or Takeoff. The site does not host the original performance or song.

Check rights and originality first

  • Use only photos, motion clips, recordings, and other media the user owns or has permission to use.
  • Treat identity images as visual references, not permission to impersonate or mislead.
  • Treat a motion clip as movement and timing guidance only. Create fresh people, setting, camera coverage, and frames.
  • Use only user-owned or properly licensed audio. Do not request the original song, an unlicensed sound recording, or a deceptive voice clone.
  • Reframe requests for an exact copy of an existing performance into a new orange-booth scene.

If these conditions are not confirmed, stop at a rights checklist. Do not prepare a generation-ready prompt.

Choose the source mode

  • video: start from authorized performer photos and create a new duet video.
  • photo: create one original still image; omit audio instructions.
  • text: describe original subjects without assuming reference media.
  • motion: use an authorized motion clip for movement only, never for copying its people, setting, or frames.

For two performers, define stable camera-left and camera-right roles. For pets, preserve species, anatomy, coat markings, and natural limb movement.

Build the prompt in this order

  1. State that the work is original and describe one or two subjects.
  2. Place them in a warm orange performance booth around one vintage silver microphone.
  3. Give the camera-left and camera-right subjects distinct actions.
  4. Describe framing, camera movement, lighting, and the intended short-form energy.
  5. Explain the role of each reference: identity, composition, motion, or audio.
  6. Add continuity anchors so identities, sides, microphone position, and lighting remain stable.
  7. Exclude extra performers, watermarks, logos, captions, and accidental text.
  8. If audio is requested, state that it must be user-owned or licensed.

Do not promise perfect likeness, perfect lip sync, or a guaranteed generation result. Current product options can change; verify live duration, aspect ratio, quality, and credit estimates before generation. The current template advertises 4, 10, and 15 second durations plus 9:16 and 16:9 formats.

Plan short beats

Make the beat durations add up exactly to the requested duration. Every beat should identify the left action, right action, camera behavior, and transition. Preserve the same subject identity on each side and keep the microphone fixed unless the concept explicitly changes it.

Diagnose before generation

Check for:

  • missing rights or originality confirmation;
  • missing reference media for the selected mode;
  • a named or exact reproduction request;
  • unstable performer positions or identity swaps;
  • missing microphone, setting, camera, or audio direction;
  • beat durations that do not total the requested runtime;
  • settings that differ from the current advertised template options.

Return blockers separately from warnings. A structurally sound prompt is not proof that a live model or product setting will accept it.

Use the optional MCP

  • build_hotel_lobby_prompt drafts the structured prompt and rights boundary.
  • plan_hotel_lobby_duet validates timed left/right beats and continuity anchors.
  • diagnose_hotel_lobby_prompt checks workflow, rights, structure, and reproduction risks.
  • get_hotel_lobby_ai_resources returns canonical product, example, tutorial, pricing, and policy pages.

Connect the stdio server with:

openclaw mcp add hotel-lobby-ai \
  --command npx \
  --arg -y \
  --arg github:gpt-img-2/hotel-lobby-ai-prompt-mcp \
  --include 'build_hotel_lobby_prompt,plan_hotel_lobby_duet,diagnose_hotel_lobby_prompt,get_hotel_lobby_ai_resources'

Then run openclaw mcp doctor hotel-lobby-ai --probe. If the installed release lacks the mcp command group, add npx -y github:gpt-img-2/hotel-lobby-ai-prompt-mcp as a Stdio server in MCP settings. The MCP is optional; never imply it ran when unavailable.

Return an inspectable result

Return Rights check, Mode, Prompt, Shot beats, Continuity anchors, Exclusions, Warnings, and Live-setting check. Keep blockers visible and do not invent reference details the user did not provide.

Canonical resources