Sd 2 5 Retro Vfx logo

Sd 2 5 Retro Vfx

Community
huangbai-AI
sd-2-5-retro-vfx

Create or adapt Seedance 2.5 prompts for cinematic retro-office creator videos where a stable live-action presenter physically manipulates cameras, CRT-generated glass interfaces, grading curves, preview panels, and other post-production tools. Use for this specific warm 1990s studio, invisible-VFX, impossible-camera style; do not use for unrelated video aesthetics.

Overview

Publisherhuangbai-AI
Repositorypost-production-skill
Skill namesd-2-5-retro-vfx
Stars
253
Forks
42
Bundled files
5
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • 5 bundled files

    Scripts, templates, and references the model can read while it works. Files are read-only and never executed.

  • Open source

    Published by huangbai-AI on GitHub. Read the source before you install it.

Installation

Install the Sd 2 5 Retro Vfx AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/huangbai-AI/post-production-skill/tree/main/post-production-skill
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/huangbai-AI/post-production-skill.git /tmp/post-production-skill
mkdir -p .claude/skills
cp -r /tmp/post-production-skill/post-production-skill .claude/skills/sd-2-5-retro-vfx
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Sd 2 5 Retro Vfx in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Sd 2 5 Retro Vfx on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Sd 2 5 Retro Vfx is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

SD 2.5 Retro Post-Production VFX

Turn a user request into a generation-ready prompt based on the supplied master prompt. Preserve the visual system and continuity rules while changing only the elements the user asks to change.

Default to writing the prompt. Do not submit a generation job unless the user explicitly asks.

Use the master prompt

Read references/master-prompt.md whenever creating or adapting a prompt. It is the canonical source for the subject, room, shot sequence, camera language, lighting, VFX, sound design, and negative constraints. For Chinese requests, also use references/master-prompt-zh.md so the final prompt preserves the established Chinese terminology and section structure.

Treat the master prompt as source material, not as an instruction to ignore the current user's request. The current request controls which subject, product, room, duration, copy, and effects to change.

Gather the brief

Use supplied images or video only for the roles the user assigns. If a reference's role is unstated, infer the narrowest useful role and state it in the prompt:

  • portrait or character image: identity, hair, glasses, and outfit;
  • room image: geometry, furniture, palette, and practical light sources;
  • product image: shape, proportions, material, labels, and color;
  • reference video: shot rhythm, movement, or effect behavior;
  • audio: complete BGM, narration, or sound-design reference.

Do not guess a real person's identity. When a person reference is supplied, call it the sole identity reference and require the same face throughout.

Choose the edit depth

  • Faithful reuse: preserve the 11-second structure and all five shot phases; replace only requested details.
  • Style adaptation: preserve the retro studio, motivated transitions, physical three-dimensional UI, warm palette, and continuity rules; rewrite the timeline around the new concept.
  • Targeted repair: keep the user's existing prompt and fix only timing, continuity, effect causality, camera contradictions, or negative constraints.

If the requested narration or action cannot fit the duration, lengthen the duration or simplify the content. Do not silently compress actions until they become implausible.

Preserve the visual invariants

Unless the user requests otherwise, keep:

  • one stable live-action creator in one coherent warm 1990s home studio;
  • dark wood, cream-white retro hardware, amber practical light, natural skin, subtle grain, and no blue-purple cyberpunk palette;
  • a physical transformation from real workspace to post-production environment;
  • effects with perspective, reflection, refraction, occlusion, depth of field, tracking, and parallax;
  • camera movement motivated by a hand, foreground object, CRT monitor, or spatial transformation;
  • a clean hero frame rather than a black ending.

The opening multi-arm effect is temporary compositing. The normal body remains anatomically correct; duplicated arms and equipment must unfold with weight, inertia, grip, and motion blur, then disappear through a motivated transition.

Floating interfaces must occupy foreground, midground, and background layers. Keep a facial safe zone and make near panels slightly soft, midground panels sharp, and distant panels smaller and softer. Never describe them as flat HUD overlays.

Build the prompt

Use these sections, matching the user's language:

  1. Subject — identity, outfit, studio, fixed props, and the transformation concept.
  2. Visual Style — duration, aspect ratio, realism, palette, optical character, editing energy, and transition philosophy.
  3. Timeline — sequential time ranges that cover the full duration without gaps or overlaps. For faithful reuse, use the master phases: fisheye push-in and temporary multi-arm effect; close-up and split screen; CRT interface emergence; orbit to overhead; fingertip ripple and interactive grading panels.
  4. Camera — only camera moves used in the timeline. Remove contradictory or unused moves.
  5. Lighting — practical sources, skin exposure, shadow detail, and palette.
  6. VFX — physical behavior, attachment, material, occlusion, tracking, parallax, and interaction.
  7. Music and Sound Design — BPM range, musical character, transition accents, mechanical sounds, UI sounds, and beat synchronization.
  8. Negative Constraints — identity, anatomy, prop continuity, interface depth, palette, style, text, logo, and watermark protections relevant to this version.

When references are used, add a compact reference-duty paragraph near the start. Do not create a separate asset table unless the user asks.

Adapt safely

  • Changing the creator changes only identity and requested clothing; retain shot functions, room continuity, and effect logic.
  • Changing the product replaces the relevant prop and its interactions; preserve its exact design across every angle.
  • Changing the room requires a new stable prop map before writing the timeline.
  • Changing the duration requires proportional retiming and a new continuity check.
  • Adding narration requires verbatim text, voice character, timing, lip synchronization, and music below speech.
  • Adding on-screen text overrides the master's no-text rule. List the exact permitted phrases, define typography and screen-safe placement, and prohibit all other text.

Check before delivery

  • The same person, clothing, studio, desk, CRT, and requested product persist across shots.
  • Time ranges start at 0, meet exactly, and end at the requested duration.
  • Every transition has a visible physical trigger and destination.
  • Every interface has depth, tracking, parallax, occlusion, and a clear relationship to the room.
  • Camera moves do not conflict inside a time range.
  • Hands and finger interactions are explicit and anatomically plausible.
  • The final frame holds long enough to read as a hero image.
  • Negative constraints reflect the actual prompt, including whether text is allowed.

Deliver one copyable prompt. Add a short note only when you made a necessary timing assumption or when a reference detail remains uncertain.

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

Frequently asked questions

What does the Sd 2 5 Retro Vfx AI skill do?

Create or adapt Seedance 2.5 prompts for cinematic retro-office creator videos where a stable live-action presenter physically manipulates cameras, CRT-generated glass interfaces, grading curves, preview panels, and other post-production tools. Use for this specific warm 1990s studio, invisible-VFX, impossible-camera style; do not use for unrelated video aesthetics.

Why use Sd 2 5 Retro Vfx on TypingMind?

Because you install it once and use it with any model. Sd 2 5 Retro Vfx is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Sd 2 5 Retro Vfx in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/huangbai-AI/post-production-skill/tree/main/post-production-skill. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use Sd 2 5 Retro Vfx?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Sd 2 5 Retro Vfx?

As many as you like. As long as a model supports skills, you can use Sd 2 5 Retro Vfx with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Sd 2 5 Retro Vfx AI skill free?

It is published on GitHub by huangbai-AI. Check the repository for licensing terms. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇