scene-anchor-system
Skill1 starsChanged 53 days agoMITView on GitHub
Use when a game, visual novel, interactive story, or visual project needs consistent background spaces, scene bibles, asset role classification, image-generation constraints, or manifest-to-background mapping without layout drift.
In this repository20 other files it gives its agents.
- llms.txtllms.txt
- agent-skill-migrationSkill
- agents-md-verification-mapSkill
- cocos-creator-mobile-gameSkill
- game-entity-movement-presentationSkill
- game-event-directorSkill
- game-ui-panelsSkill
- godot-narrative-workflowSkill
- narrative-game-prototype-plannerSkill
- obsidian-article-writebackSkill
- offline-ux-kitSkill
- product-image-prompt-iterationSkill
- product-teardownSkill
- reporting-disciplineSkill
- screen2layersSkill
- sol-luna-local-parallelSkill
- sprite-atlas-slicerSkill
- unfinished-work-auditSkill
- windows-ai-cli-recoverySkill
- xiaohongshu-copywriterSkill
The file
We split this skill into the rules it gives an agent
2 rules. About 0 are common, 2 are distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈2
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 33
When using this skill, report:
- Distinctive: line 46
Use structured manifest/data parsing when available.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Scene Anchor System
Core Principle
Same space means same structure. Treat scene art as anchored spaces plus state changes, not as separate illustrations that only share a mood or style.
Workflow
- Locate current background assets, runtime manifest keys, design docs, screenshots, and any image-generation prompts.
- Classify every visual asset into one role:
BASE_SPACE,STATE_VARIANT,CLOSE_UP,MEMORY_CUT,EVENT_INSERT,UI_OR_REPORT_SURFACE, orEXTERNAL_SPACE. - Pick exactly one base image for each persistent space. Do not allow two competing bases for the same room, corridor, shop, street, or hub.
- Lock object anchors for persistent spaces: core furniture, windows, doors, desks, beds, major props, horizon/camera relation, and interaction hotspots.
- Define allowed changes for variants: lighting, weather, time of day, emotional overlay, temporary props, clutter state, and mild camera noise.
- Map runtime keys or manifest entries to scene roles before changing code, data, or prompts.
- Patch or create a scene bible and Codex/agent constraints before generating new art or wiring replacements.
Rules
- Variants may change atmosphere, not topology.
- Close-ups may crop or magnify a base-space object, but must not contradict the base layout when returning to the main scene.
- Memory cuts and event inserts are allowed to be compositionally different, but they must be labeled as inserts rather than persistent spaces.
- UI overlays are screen layers unless the project explicitly treats them as diegetic objects in the scene.
- Narrative variables, skip states, compression states, or report states should alter overlays, text, lighting, and availability; they should not silently move furniture or redefine the room.
- Preserve existing runtime keys unless the project already has a migration plan. Add role metadata beside current keys instead of renaming casually.
Output Contract
When using this skill, report:
- Current evidence inspected: files, manifests, screenshots, or docs.
- Asset classification table.
- One base scene per persistent space.
- Locked anchors and interaction hotspots.
- Allowed and forbidden changes for variants.
- Manifest or schema recommendations.
- Unresolved art decisions that need human confirmation.
- Validation checklist and what could not be visually verified.
Validation
Use structured manifest/data parsing when available. Inspect screenshots or actual image files before claiming visual consistency. If only filenames and docs are available, label the result as a semantic audit rather than a visual QA pass.
Reference
For a reusable scene bible template, read references/scene-bible-template.md.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

