Communitygithub.com

Character Design Engine

Turns a Creative Brief (from aesthetic-director-core) into a complete, model-agnostic character blueprint — design thesis, silhouette architecture, visual anchors, garment engineering, material contrast, palette hierarchy, behavioral pose + camera, narrative residue, presentation format — then runs a subtraction pass. For OC, character design, fashion concept, game character sheets.

What is Character Design Engine?

The Character Design Engine is a specialized skill that transforms a Creative Brief generated by the aesthetic-director-core into a fully resolved, model-agnostic character blueprint. It enforces a strict, dependency-driven pipeline where each design decision (silhouette, anchor points, garment layers, material contrast, color hierarchy, pose, camera, narrative residue) logically follows from the previous one, eliminating design ambiguity. The skill is designed for creating original characters (OCs), game characters, fashion concepts, and character sheets, and it explicitly avoids writing model-specific prompt syntax (e.g., Anima, NAI5) — ensuring the output is a pure design specification that any adapter can translate. It includes a mandatory subtraction pass (Step 10) and an anti-mediocrity check to guarantee the design is distinctive and conceptually coherent.

Works with~Claude Code~Codex CLI~Cursor
npx skills add https://github.com/Tera-Dark/personal-agent-skills/tree/HEAD/02_creation/character-design-engine

Ask in your favorite AI

Open a new chat with this agent skill pre-loaded.

Documentation

What does Character Design Engine do?

Mission

Input: A Creative Brief from aesthetic-director-core (direction, contradiction, silhouette strategy, causal chain, moment, density map, punctum, retained weirdness, deleted elements). Output: A character blueprint — any model adapter can translate it directly without making further design decisions.

If called directly without a Brief: First run a compressed pass using aesthetic-director-core's M1 / M3 / M6 / M7 (internal completion, no need to output full Brief), then proceed with the flow below. Do not skip. Skipping results in form-filling.

If the user explicitly requests maximalist exquisite OC / maximalist character design, prioritize reading references/maximalist-oc-design-grammar.md; this mode permits high complexity, but complexity must come from silhouette, garment engineering, unified motifs, material contrast, and multi-scale detail — not random accessory stacking.

This skill does not write Anima / NAI5 syntax. This skill does not re-decide taste—taste comes from personal-identity-profile.

Why Not a Table

The previous version's output contract was Identity: / Background: / Silhouette: / Costume: / ... — thirteen fields. Fields get filled, and they get filled with the most common values.

This version outputs a chain of interdependent decisions: each subsequent item is derived from the preceding ones. If an item cannot be derived, it means the preceding decisions were not solid.

Execution Flow

For detailed techniques, see references/oc-design-system.md (six-layer structure, silhouette strategies, complexity control, anti-mediocrity check).

Step 1  Thesis        Start from the Brief's 【Direction】. One sentence. Must include a verb. Can't write a verb → return to director.
Step 2  Silhouette Strategy  Choose one of five: halo/radial · vertical spear · asymmetric cascade · cocoon/volume · fragmented orbit
                    Write spatial relationship: behind the head / wrapping the torso / descending from one shoulder
Step 3  Visual Anchors  1 primary anchor (choose one: head/chest/waist/back/object in hand) + ≤2 secondary anchors. Primary anchor = Brief's dense zone.
Step 4  Garment Engineering    Base garment → Structural garment → Signature extension → Accessory system
                    Each layer: write "shape + location + function/motion", no adjectives. Vocabulary in references/garment-lexicon.md
Step 5  Material Contrast  At least one pair: matte vs reflective / soft vs hard / fabric vs organic / rough vs smooth. Bind to specific structures.
Step 6  Color Hierarchy   Base 60–75% / Structural color 20–30% / Punctum ≤5%. Write placement for each color. Punctum = Brief's punctum.
Step 7  Pose + Camera  Pose is given by Brief's 【Moment】 and 【Causal Chain】. Camera is determined by pose type (showcase/manipulation/turn-back/ritual/vulnerability → corresponding shot).
Step 8  Narrative Residue  One object's state indicating something happened. Must be a result of her action.
Step 9  Presentation Format  clean plate / decorated key visual / editorial plate / environmental vignette
Step 10 Subtraction Pass  For every item in Steps 4–9 ask "If I delete this, does the thesis still stand?" Yes → Delete. Record what was deleted.
Step 11 Anti-Mediocrity Check  references/oc-design-system.md § 4

Output Contract: Character Blueprint

Write in coherent short paragraphs, not a blank table. Each paragraph begins with a decision, followed by its rationale. All facts the adapter needs must be expressed using noun + location + behavior.

## [Character Name or Code Name]

**Thesis**: One sentence.
**Silhouette**: Strategy + spatial relationship + the shape recognizable in a thumbnail.
**Anchors**: Primary anchor at ___; secondary anchors at ___, ___. Quiet zone at ___.
**Costume**:
  - Base:
  - Structural:
  - Signature extension:
  - Accessory system (same morphological grammar):
**Material Contrast**: ___ vs ___, located on ___.
**Color Palette**: Base ___ / Structural ___ / Punctum ___ (location).
**Appearance**: Hair (shape, length distribution, color — not "silver long straight"), eyes, skin, racial characteristics.
**Pose and Camera**: She is ___ (moment). Because ___ therefore ___ (causality). Camera position ___, gaze ___, hands ___.
**Narrative Residue**: ___
**Presentation Format**: ___
**Retained Weirdness**: ___, because ___.
**Deleted**: ___, ___, ___.
**Vetoed Directions**: ① ___ ② ___
**Locked Facts** (explicitly given by user, not to be "optimized" by adapter): ___

Multiple Designs

When the user requests N sets, each set must differ in at least two of the following five items: Thesis, Silhouette Strategy, Primary Anchor, Material Language, Action Logic. Do not produce three color variants of the same design.

Reference Image Input

When the user provides reference images, first run image-reverse-analysis to extract structure (silhouette, density distribution, motif grammar, color hierarchy, material, pose logic, negative space), then swap out all specific objects while retaining the structure. See aesthetic-director-core/references/taste-calibration-pairs.md Pair 5.

Self-Check

  • Does the thesis contain a verb?
  • Is there only one primary anchor point?
  • Does each of the four garment layers have a shape and location, with no adjectives?
  • If you remove color, is the silhouette still recognizable? If you remove decoration, does the base garment still have independent tailoring?
  • Does the pose change the position of hair / hem / props / light?
  • Is the narrative residue caused by her?
  • Is there only one punctum, with a precise location?
  • Are the "Deleted" and "Vetoed Directions" rows present?
  • Does it conflict with personal-identity-profile/references/design-dislikes.md?
  • Are the locked facts fully passed through?

References

  • references/oc-design-system.md — Six-layer structure, five silhouette strategies, pose-camera mapping, presentation formats, complexity control, anti-mediocrity check, default behavior for open-ended requests
  • references/garment-lexicon.md — Outfit archetypes, four-layer layering, asymmetrical techniques, material collision matrix, collar/sleeve/waist/skirt vocabulary (English phrases can go directly into prompts)
  • references/maximalist-oc-design-grammar.md — Dedicated design grammar for maximalist exquisite OC: macro silhouette, multi-layer garments, unified motifs, asymmetric weight, material collisions, multi-scale detail, and pose activation

Individual skills in this repo

This repo contains 13 individual skills — each has its own dedicated page.

Tera-Dark/personal-agent-skills

Creative direction layer that turns a vague request into one committed design idea before any blueprint or prompt is written. Runs a sequence of generative "moves" (find the obsession, plant a contradiction, pick from the tail, build causality, subtract, keep one strange thing) and produces a short Creative Brief with rejected alternatives. Use for any OC / character / illustration / fashion / key-visual request, whenever output feels generic, "AI-flavored", too plain, too busy, or when the user asks for taste, direction, 审美, 创意方向, 人味, 高级感, 不要AI味.

Tera-Dark/personal-agent-skills

Persistent identity and taste layer for Tera-Dark's creative work. Holds the aesthetic signature (restrained base, one strange thing, a hint of danger, fashion-grade garment construction, female-oriented OC/gacha sensibility), hard dislikes (watches, cyber/mech, random butterflies-roses-particles, adjective costumes), business goals and workflow style. Does not generate prompts. Load first for any creative, design, illustration, prompt or aesthetic-judgment task; also when the user says 我的风格, 我喜欢, 按我习惯, 个人偏好.

Tera-Dark/personal-agent-skills

Entry point for all creative requests in this skill hub. Classifies intent (OC/character design, illustration, fashion, NAI5 prompt, Anima prompt, image reverse analysis, prompt review, ComfyUI/LoRA/dataset), loads identity + aesthetic direction first, then hands off to the right specialist and model adapter. Use whenever a request involves 设计, OC, 人设, 立绘, 插画, 服装, 提示词, prompt, NAI, NovelAI, Anima, 反推, 分析图片, ComfyUI, LoRA, or when it is unclear which skill should handle a creative task.

Tera-Dark/personal-agent-skills

Model adapter that compiles a finished character or illustration blueprint into Anima-ready English prompts in a disciplined Tag block + Natural Language block format. Handles Anima-specific format contract, length budgets, positive-first output, V1 faithful / V2 enhanced modes, model profiles and artifact troubleshooting. Use when the user asks for Anima prompts, Anima 提示词, or names an Anima checkpoint. Does not design — if no blueprint exists, route through aesthetic-director-core and character-design-engine / illustration-direction first.

Tera-Dark/personal-agent-skills

Model adapter that compiles a finished character or illustration blueprint into a natural-language image prompt for models that are not Anima or NAI5 — Midjourney, DALL-E / GPT Image, Google Imagen / Gemini image, Flux, SDXL-style checkpoints, or an unnamed target. Keeps parameters out of the prompt unless the target officially supports them, and labels every model-specific claim with an evidence level. Use when the user names any other image model, says 通用提示词, MJ, Midjourney, DALL-E, Imagen, Flux, SD, 或没说用什么模型. Does not design — if no blueprint exists, route through aesthetic-director-core first.

Tera-Dark/personal-agent-skills

Turns a Creative Brief into an authored, model-agnostic illustration blueprint — the image thesis, captured moment, visual motif, environment-character relationship, camera and framing, scale contrast, negative space, physical light sources, density map, narrative residue, and one coherent surreal detail. Use for 插画, 氛围图, 竖屏, 半留白, 印象风, 故事感, key visual, poster, scene illustration, cinematic composition, or whenever a character needs to be placed into a designed moment rather than displayed on a plate. Never writes model-specific prompt syntax.

Tera-Dark/personal-agent-skills

Model adapter that compiles a finished character or illustration blueprint into NovelAI V5 community-format prompts — weighted artist stack, global style layer, scene base, char1/char2 blocks, source#/target#/mutual# interaction tags, and targeted negative steering with weight::tag:: syntax. Use when the user asks for NAI5, NovelAI, NAI提示词, tag prompt, Danbooru-style prompt. Does not design — if no blueprint exists, route through aesthetic-director-core and character-design-engine / illustration-direction first.

Tera-Dark/personal-agent-skills

Decomposes a reference image into transferable design language — silhouette and mass distribution, motif grammar, garment construction, material contrast, palette hierarchy, pose logic, negative space and presentation format — instead of a tag dump. Output feeds aesthetic-director-core (for original work) or a model adapter (for faithful reproduction). Use for 反推, 分析图片, 提取提示词, 还原风格, 参考这张图, reverse prompt, image analysis.

Tera-Dark/personal-agent-skills

Reviews an existing image-generation prompt by separating design intent from model syntax — what it is trying to say, which design layer is weak (thesis, silhouette, garment structure, causality, density, punctum), what is noise, and what a rewrite should change. Use for 优化提示词, 这个 prompt 哪里有问题, prompt review, 分析这段提示词, why does this prompt not work.

Tera-Dark/personal-agent-skills

Technical placeholder for ComfyUI workflow organization, node troubleshooting, metadata and LoRA/model integration. Use when the user mentions ComfyUI, workflow json, nodes, 工作流. Does not handle design or prompt creativity.

Tera-Dark/personal-agent-skills

Technical placeholder for image dataset organization, caption/tag management and training preparation. Use when the user mentions dataset, 数据集, 打标, captions, 训练集. Does not handle visual design decisions.

Tera-Dark/personal-agent-skills

Technical placeholder for LoRA training pipelines: dataset prep, caption organization, training configuration, evaluation. Use when the user mentions LoRA, 训练, fine-tune, 炼丹. Does not decide character concepts or artistic direction.

Tera-Dark/personal-agent-skills

Evaluates a generated image or compiled prompt against the original Creative Brief and blueprint, using a fixed rubric (identity preservation, outfit binding, spatial clarity, unrequested additions, V1/V2 consistency, output contract) plus a design-layer read (thesis, silhouette, causality, density, punctum, one strange thing). Diagnoses which layer failed and routes the fix to that layer with a single-variable change. Use on any feedback round: 太平淡, 太乱, 不像, 没人味, 这版可以, 评价一下, 哪里不对, compare versions.

Related Skills