Scene Concept Art Director logo

Scene Concept Art Director

Community
AharaOoO
scene-concept-art-director

Turn a brief, approved moodboard, or existing scene image into a structured scene-art direction contract, model-ready environment prompts, reference upload rules, multi-view continuity notes, and evidence-based visual QA. Use for film, animation, game, commercial, architectural, fantasy, sci-fi, historical, interior, landscape, and repair workflows.

Overview

PublisherAharaOoO
Repositoryacheng-director-skill
Skill namescene-concept-art-director
Stars
51
Forks
5
Bundled files
15
LicenseApache-2.0
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.

  • 15 bundled files

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

  • Open source

    Published by AharaOoO on GitHub. Read the source before you install it.

Installation

Install the Scene Concept Art Director 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/AharaOoO/acheng-director-skill/tree/main/modules/assets/scene-design/external
  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/AharaOoO/acheng-director-skill.git /tmp/acheng-director-skill
mkdir -p .claude/skills
cp -r /tmp/acheng-director-skill/modules/assets/scene-design/external .claude/skills/scene-concept-art-director
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Scene Concept Art Director 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 Scene Concept Art Director 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 Scene Concept Art Director 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.

Scene Concept Art Director

This is a scene-design specialist, not a second production truth source or a general project orchestrator. It converts a brief into a reviewable scene contract and prompt artifacts. When Acheng Director is present, the scene contract is an advisory artifact: the assets owner merges approved scene facts into scene_registry and asset cards; shots owns shot-specific camera and framing; effects owns VFX and scale proofs; continuity owns persistent state. Do not edit another owner's production fields.

Route the request before writing

Choose one mode and state it in the result:

  • direction: brief or moodboard → scene-art direction contract.
  • prompt: approved contract → one complete image prompt and upload manifest.
  • repair: existing image or prompt → one narrow repair transaction; preserve named successful facts.
  • multi_view: one scene → establishing, background plate, top-down/layout, elevation, keyframe, or day/weather variants with a shared continuity ledger.
  • critique: prompt or image → rubric findings and the smallest structural repair.
  • handoff: export the contract and prompt artifacts to Acheng assets without silently editing production.json.

Do not force a moodboard, STYLE_MOTHER, a hero key art image, a camera move, or a specific image model when the request does not need it.

Inputs and assumptions

Accept a user brief, an approved moodboard-alignment/data.json, a production scene record, existing reference images, or a previous scene contract. Map moodboard summary, palette, composition, style, and relevant node fields into the scene contract; preserve their provenance and keep unresolved conflicts visible. A moodboard image is not automatically a STYLE_MOTHER.

Ask at most one question only when the missing answer changes the scene's purpose or spatial logic. Otherwise write labeled assumptions and continue. Never invent historical certainty, geography, ownership, measurements, or a reference file that was not supplied.

Required contract and artifacts

Every non-critique result starts from references/scene-contract-v1.md and produces a scene_art_direction.json (or an inline object with the same fields). The contract must include:

  • scene identity, revision, purpose, status, and source provenance;
  • thesis, location, era/season/day phase, function, power or maintenance structure, recent event, and off-screen world;
  • spatial grammar, boundaries, entrances/exits, landmarks, traversal, foreground/midground/background, and human-scale anchors;
  • materials, aging, cultural logic, motifs, human use, iconic object, weather, and atmosphere;
  • camera intent, composition, lighting sources and direction, color roles, and depth strategy;
  • references with role, preserve, exclude, real file, approval state, and hash when available;
  • generation intent, model family, input mode, execution settings, prompt status, unresolved items, and visual_status;
  • Acheng handoff owner, consumers, frozen facts, and allowed advisory paths.

For prompt, repair, and multi_view, also produce a complete *.image.txt, UPLOAD.md, index.json, and qa.json when files are requested. The prompt is independent of the conversation and contains concrete spatial, material, lighting, scale, and human-use facts. Upload instructions remain outside the prompt.

prompt_status=READY_TO_SUBMIT means the text and required inputs are present. It never means a real image was generated or visually approved. Until a human reviews an actual output, keep visual_status=UNVERIFIED.

Design method

  1. State one scene thesis and one primary visual tension.
  2. Choose a dominant spatial grammar and, only when justified, a secondary material or lighting grammar. Explain why they meet.
  3. Design the space as something people use, maintain, control, fear, repair, worship, sell, or abandon.
  4. Build foreground, midground, background, focal path, negative space, scale anchors, and an off-screen world before adding decoration.
  5. Motivate light from real sources and assign a restrained color script with one focal accent.
  6. Translate the contract into a model-specific prompt. Keep positive scene facts, reference preservation, exclusions, and execution settings separate.
  7. Run the relevant QA rubric. If the concept is generic, make one structural change, add one grounded use detail, and remove decorative noise.

References and model boundaries

Load only the references required by the route:

  • references/scene-contract-v1.md: schema, status, ownership, and handoff rules for every route.
  • references/art-direction-framework.md: thesis, cultural synthesis, material and human-use logic.
  • references/cinematic-scene-language.md: composition, camera, lighting, color, depth, and atmosphere.
  • references/prompt-recipes.md: model-neutral prompt assembly, variants, and repair transactions.
  • references/model-adapters.md: provider capability differences and execution-setting separation.
  • references/multi-view-continuity.md: shared anchors and allowed deltas across views and states.
  • references/qa-rubric.md: critique, scoring, and final gate.
  • references/acheng-handoff.md: Acheng ownership and moodboard-to-scene mapping.
  • references/usage-guide.md: user-facing copy-ready templates for every scene route and Acheng collaboration path.
  • references/cloud-song-archive-case-study.md: calibration only; never import its setting, providers, or visual facts into another project.

Do not claim that a provider supports a parameter without current evidence. Do not call a paid image or audio model unless the host and user explicitly authorize execution. Default to plan/prompt-only work.

Continuation and revision

Long work returns continuation_required, current_cursor, completed artifact paths, and the next contract section. Resume from the cursor; do not replace the contract with a summary. A revision must name the scene, field, and preserved facts. Use a delta transaction (CHANGE, PRESERVE_EXACTLY, REBUILD) for image repairs and never redesign the whole scene for a local defect.

Human-facing output

For ordinary requests show: mode, scene thesis, assumptions and blockers, contract summary, complete prompt, upload manifest, QA findings, and visual verification status. For machine requests show the full JSON, hashes, evidence, and unresolved list. Keep explanatory text outside copyable model prompts.

Style and cultural ethics

Do not imitate a living artist, director, production designer, or studio house style as a direct target. Convert references into craft attributes. Treat culture as spatial, material, climate, labor, and ritual logic rather than a decorative sticker; flag uncertainty instead of fabricating authenticity.

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 Scene Concept Art Director AI skill do?

Turn a brief, approved moodboard, or existing scene image into a structured scene-art direction contract, model-ready environment prompts, reference upload rules, multi-view continuity notes, and evidence-based visual QA. Use for film, animation, game, commercial, architectural, fantasy, sci-fi, historical, interior, landscape, and repair workflows.

Why use Scene Concept Art Director on TypingMind?

Because you install it once and use it with any model. Scene Concept Art Director 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 Scene Concept Art Director in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/AharaOoO/acheng-director-skill/tree/main/modules/assets/scene-design/external. 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 Scene Concept Art Director?

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 Scene Concept Art Director?

As many as you like. As long as a model supports skills, you can use Scene Concept Art Director with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Scene Concept Art Director AI skill free?

Yes. It is published on GitHub by AharaOoO under the Apache-2.0 license. 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 👇