agentleFS
LogicOberLogicOber/Worldview-Skills/skills/core-engine/film-style-library/SKILL.md

film-style-library

Skill2 starsChanged 50 days agoApache-2.0View on GitHub

Lock a film or game-video project to one visual language before asset production. Use the user's references when available or create a small project-specific comparison board when an authorized image tool is present. Produces the style header, asset-sheet modifier, negative constraints, and model routing consumed by downstream Skills.

The file

Show its rules and text

We split this skill into the rules it gives an agent

13 rules. About 0 are common, 12 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈12

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Check every later reference against the lock before it enters generation.

    line 24
  2. Distinctive:

    Prefer evidence in this order:

    line 28
  3. Distinctive:

    Only present directions that fit the story and delivery format.

    line 47
  4. Distinctive:

    If the user already supplied consistent references, confirm the inferred direction in one sentence and continue.

    line 66
  5. Distinctive:

    When the project has no suitable references and image generation is available:

    line 70
  6. Distinctive:

    Create no more than four candidates.

    line 73
  7. Very specific:

    If image generation is unavailable or unauthorized, use text cards and set reference to null.

    line 79
  8. Distinctive:

    Use only tools verified in the current environment.

    line 102
  9. Distinctive:

    Before any image becomes a generation reference:

    line 106
  10. Distinctive:

    If it matches, record the check and continue.

    line 109
  11. Distinctive:

    If it conflicts, stop before spending quota.

    line 110
  12. Distinctive:

    If the style is reopened, invalidate downstream assets and prompts that depended on the earlier lock.

    line 111
  13. Distinctive:

    Do not generate production characters, locations, or props before the style lock exists.

    line 120

≈ “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.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Film Style Library

Why this stage exists

Words such as “animated,” “cinematic,” and “realistic” leave too many production decisions open. Line weight, material response, lighting, motion cadence, lens behavior, and surface detail can produce very different results inside the same label. A small visual comparison usually resolves that ambiguity faster than another paragraph of adjectives.

The first locked asset also influences every later reference. If its visual language is wrong, characters, locations, props, and shots inherit the mismatch. Lock style before producing those assets.

What this Skill does

  1. Inspect the request, existing project, and any user-supplied references.
  2. Infer a likely direction when the evidence is already strong; otherwise compare three or four relevant directions.
  3. Lock the selected direction in films/<slug>/style.json.
  4. Emit four fields every downstream Skill can reuse:
    • style_header for image and video prompts;
    • sheet_modifier for character, location, and prop sheets;
    • negative_block for the main forms of visual drift;
    • model_routing for tools actually available in the project.
  5. Check every later reference against the lock before it enters generation.

Build the comparison from project evidence

Prefer evidence in this order:

  1. approved user-supplied images;
  2. an existing project style guide or locked production asset;
  3. project-specific comparison frames made with an authorized image tool;
  4. concise text cards when no image evidence can be created.

Do not ship or depend on a permanent universal image library. A generic card can carry the wrong subject, era, palette, or production assumptions into a new project. Comparison frames belong to the current project and must keep subject, framing, action, aspect ratio, and major palette constant. Change only the visual-language fields the user is choosing.

Useful direction families include:

DirectionShape and surfaceLight and colorMotion and camera
Photographic cinemaNatural anatomy and materialsMotivated light and restrained gradePhysical lenses and controlled camera mass
Real-time game cinematicGame-ready geometry and PBR surfacesWorld lighting that can survive runtimeReadable staging and camera collision awareness
Graphic 2D animationDesigned silhouettes, line weight, and cel regionsFlat palettes and hard shadow groupsHeld poses, smears, impact frames, selective camera motion
Stylized 3D animationSimplified proportions and shaped materialsSoft bounce and controlled gradientsClear arcs, expressive timing, stable spatial continuity
Documentary observationUnstaged texture and practical detailAvailable light and muted correctionRestrained handheld framing led by events

Only present directions that fit the story and delivery format. Do not make the user choose between unrelated scenes and call that a style choice.

Conversation pattern

User: I want an animated short about two rival swordsmen.

Agent: Which direction is closest to what you see?
  A. Graphic 2D — sharp silhouettes, cel regions, impact frames
  B. Stylized 3D — shaped materials, soft bounce, expressive timing
  C. Photographic — natural materials, motivated light, physical lenses
  Reply with a letter, or send a reference image.

User: A

Agent: Locked: Graphic 2D. Every character sheet, location plate, prop sheet,
and shot prompt for this project will use that visual language.

If the user already supplied consistent references, confirm the inferred direction in one sentence and continue. Ask only when the evidence conflicts or two materially different directions remain plausible.

Create project-specific comparison frames

When the project has no suitable references and image generation is available:

  1. Record the exact subject, framing, action, aspect ratio, and major palette shared by every candidate.
  2. Create no more than four candidates.
  3. Change only shape, surface, light, line, motion treatment, and lens behavior.
  4. Save candidates under films/<slug>/style-comparison/ with their generation record.
  5. Describe each visible difference in one sentence.
  6. Record the selected file in style.json; rejected candidates are not downstream references.

If image generation is unavailable or unauthorized, use text cards and set reference to null. Do not claim that a verbal choice has the same evidentiary strength as a selected image.

Style lock file

Write films/<slug>/style.json:

{
  "style_id": "graphic-2d",
  "style_name": "Graphic 2D animation",
  "reference": "films/<slug>/style-comparison/selected.png",
  "style_header": "Graphic 2D animation with designed silhouettes, clean line weight, hard cel-shadow groups, controlled impact frames, and stable spatial continuity.",
  "sheet_modifier": "Production sheet with the locked silhouette, line, palette, and cel-shadow rules on a plain neutral background.",
  "negative_block": "No photographic skin, no generic 3D render, no painterly texture drift, and no changes to the locked palette or proportions.",
  "model_routing": {
    "sheets": "<available image tool or manual workflow>",
    "video": "<available video tool or runtime>",
    "video_draft": "<optional available draft path>"
  },
  "locked_at": "<ISO-8601 timestamp>"
}

Use only tools verified in the current environment. A missing provider, model, credential, or budget is a capability gap, not permission to invent availability.

Guard later references

Before any image becomes a generation reference:

  1. Compare silhouette, proportions, surface language, palette, lighting, and line treatment with the selected reference.
  2. If it matches, record the check and continue.
  3. If it conflicts, stop before spending quota. Name the mismatch and offer two clear choices: rebuild that asset inside the lock or deliberately reopen the project style.
  4. If the style is reopened, invalidate downstream assets and prompts that depended on the earlier lock.

Output

  • films/<slug>/style.json
  • optional films/<slug>/style-comparison/ candidates and generation record
  • one confirmation line naming the selected direction
  • a list of any existing assets that conflict with the lock

Do not generate production characters, locations, or props before the style lock exists.

Download now (ZIP)

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
What did you use it for? e.g. Reviewing pull requests in a Rails app

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.