スキル一覧に戻る

ornament-style-mono

pjt222
更新日 Yesterday
3 閲覧
17
2
17
GitHubで表示
メタaidesign

について

このClaudeスキルは、開発者が古典的な歴史的様式に基づいたモノクロの装飾パターンをAIで生成することを支援します。装飾的なボーダー、メダリオン、フリーズのラインアートやシルエット表現を作成するための構造化されたガイダンスを提供します。参照用画像の作成や、生成AIを通じた歴史的装飾の探求が必要な際にご利用ください。

クイックインストール

Claude Code

推奨
メイン
npx skills add pjt222/agent-almanac -a claude-code
プラグインコマンド代替
/plugin add https://github.com/pjt222/agent-almanac
Git クローン代替
git clone https://github.com/pjt222/agent-almanac.git ~/.claude/skills/ornament-style-mono

このコマンドをClaude Codeにコピー&ペーストしてスキルをインストールします

ドキュメント

Ornament Style — Monochrome

Design monochrome ornamental patterns by combining art historical knowledge of classical ornament with AI-assisted image generation. Every design rooted in specific historical period and motif tradition from Alexander Speltz's The Styles of Ornament (1904).

When Use

  • Creating decorative borders, medallions, friezes, or panels in single color
  • Exploring historical ornament styles through generative AI
  • Producing line art, silhouette, woodcut, or pen-and-ink renderings of classical motifs
  • Generating reference imagery for design, illustration, or educational materials
  • Studying structural grammar of ornamental traditions across cultures and periods

Inputs

  • Required: Desired historical period or style (or "surprise me" for random selection)
  • Required: Application context (border, medallion, frieze, panel, tile, standalone motif)
  • Optional: Specific motif preference (acanthus, palmette, meander, arabesque, etc.)
  • Optional: Rendering style preference (line art, silhouette, woodcut, pen-and-ink, engraving)
  • Optional: Target resolution and aspect ratio
  • Optional: Seed value for reproducible generation

Steps

Step 1: Select Historical Period

Choose period from classical ornament taxonomy. Each period has characteristic motifs and structural principles.

Historical Ornament Periods:
┌───────────────────┬─────────────────┬──────────────────────────────────────────┬──────────────────────┐
│ Period            │ Date Range      │ Key Motifs                               │ Mono Suitability     │
├───────────────────┼─────────────────┼──────────────────────────────────────────┼──────────────────────┤
│ Egyptian          │ 3100–332 BCE    │ Lotus, papyrus, scarab, winged disk,     │ Excellent — bold     │
│                   │                 │ uraeus, ankh                             │ geometric forms      │
├───────────────────┼─────────────────┼──────────────────────────────────────────┼──────────────────────┤
│ Greek             │ 800–31 BCE      │ Meander/Greek key, palmette, anthemion,  │ Excellent — high     │
│                   │                 │ acanthus, guilloche, egg-and-dart        │ contrast geometry    │
├───────────────────┼─────────────────┼──────────────────────────────────────────┼──────────────────────┤
│ Roman             │ 509 BCE–476 CE  │ Acanthus scroll, rosette, grotesque,     │ Very good — dense    │
│                   │                 │ candelabra, rinceau, trophy              │ carved relief style  │
├───────────────────┼─────────────────┼──────────────────────────────────────────┼──────────────────────┤
│ Byzantine         │ 330–1453 CE     │ Interlace, vine scroll, cross forms,     │ Good — flat          │
│                   │                 │ basket weave, peacock, chi-rho           │ silhouette style     │
├───────────────────┼─────────────────┼──────────────────────────────────────────┼──────────────────────┤
│ Islamic           │ 7th–17th c.     │ Arabesque, geometric star, muqarnas,     │ Excellent — pure     │
│                   │                 │ tessellation, knotwork, calligraphic     │ geometric abstraction│
├───────────────────┼─────────────────┼──────────────────────────────────────────┼──────────────────────┤
│ Romanesque        │ 1000–1200 CE    │ Interlace, beast chains, chevron,        │ Very good — heavy    │
│                   │                 │ billet, zigzag, inhabited scroll         │ carved stone quality │
├───────────────────┼─────────────────┼──────────────────────────────────────────┼──────────────────────┤
│ Gothic            │ 1150–1500 CE    │ Trefoil, quatrefoil, crocket,           │ Very good — tracery  │
│                   │                 │ finial, tracery, naturalistic leaf       │ and window patterns  │
├───────────────────┼─────────────────┼──────────────────────────────────────────┼──────────────────────┤
│ Renaissance       │ 1400–1600 CE    │ Grotesque, candelabra, putto,           │ Good — engraving     │
│                   │                 │ medallion, festoon, cartouche           │ and woodcut styles   │
├───────────────────┼─────────────────┼──────────────────────────────────────────┼──────────────────────┤
│ Baroque/Rococo    │ 1600–1780 CE    │ C-scroll, S-scroll, shell, asymmetric   │ Moderate — complex   │
│                   │                 │ cartouche, garland, ribbon              │ forms benefit from   │
│                   │                 │                                          │ color for depth      │
├───────────────────┼─────────────────┼──────────────────────────────────────────┼──────────────────────┤
│ Art Nouveau       │ 1890–1910 CE    │ Whiplash curve, organic line, lily,     │ Excellent — defined  │
│                   │                 │ dragonfly, femme-fleur, sinuous vine    │ by line quality      │
└───────────────────┴─────────────────┴──────────────────────────────────────────┴──────────────────────┘
  1. User specified period? Confirm and note its characteristic motifs
  2. "Surprise me"? Select randomly — weight toward periods with "Excellent" mono suitability
  3. Note 2-3 primary motifs associated with period for use in prompt construction

Got: Clearly identified period with 2-3 candidate motifs and understanding of why period's ornament works well (or presents challenges) in monochrome.

If fail: User requests period not in table (e.g., Celtic, Aztec, Art Deco)? Research its ornamental vocabulary using WebSearch or WebFetch and construct equivalent entry with motif list and mono suitability assessment before proceeding.

Step 2: Analyze Motif Structure

Understand structural grammar of chosen motif before constructing prompt.

  1. Identify symmetry type:

    • Bilateral (mirror across one axis — most organic motifs)
    • Radial (rotational — rosettes, medallions, star patterns)
    • Translational (repeating unit — friezes, borders, tessellations)
    • Point (central focus radiating outward — compass roses, mandalas)
  2. Identify geometric scaffold:

    • Circle-based (rosettes, medallions, roundels)
    • Rectangle-based (panels, metopes, cartouches)
    • Triangle-based (pediment fills, spandrels)
    • Band-based (friezes, borders, running ornament)
  3. Identify fill pattern:

    • Solid (silhouette, no internal detail)
    • Line-filled (hatching, cross-hatching, parallel lines)
    • Open (outline only, negative space dominant)
    • Mixed (outline with selective internal detail)
  4. Identify edge treatment:

    • Clean boundary (contained within frame)
    • Organic bleed (motif extends beyond or dissolves at edges)
    • Interlocking (connects to adjacent units — for repeating patterns)

Got: Structural description like "bilateral symmetry, band-based scaffold, line-filled, interlocking edges" that informs prompt.

If fail: Motif structure unclear? Look up visual references using WebSearch for "[period] [motif] ornament" and analyze first few results. Speltz's original plates are public domain and widely available online.

Step 3: Construct Monochrome Prompt

Build text prompt for Z-Image generation using period, motif, and structural analysis.

Prompt Template:

[Rendering style] of [motif name] ornament in the [period] style,
[composition type], monochrome, black and white,
[structural details from Step 2],
[application context], [additional qualifiers]

Rendering Style Options:

  • detailed line art — clean vector-like lines, no fills
  • black silhouette — solid black forms on white ground
  • woodcut print — bold carved lines with wood grain texture
  • pen-and-ink illustration — fine lines with hatching for depth
  • copperplate engraving — precise parallel lines creating tonal gradation
  • stencil design — connected negative space, no floating islands

Composition Qualifiers:

  • symmetrical, centered, repeating pattern, border design
  • isolated motif on white background, continuous frieze
  • within a circular frame, filling a rectangular panel

Monochrome Constraint (always include):

  • monochrome, black and white, no color, no shading (for pure line art)
  • monochrome, black and white, high contrast (for silhouette)
  • monochrome, black and white, fine hatching for depth (for engraving style)

Example Prompts:

  • detailed line art of Greek meander border pattern, continuous frieze, monochrome, black and white, geometric precision, repeating unit, classical antiquity style
  • black silhouette of Egyptian lotus and papyrus ornament, symmetrical panel design, monochrome, black and white, high contrast, temple decoration style
  • pen-and-ink illustration of Art Nouveau whiplash curve with lily motif, vertical panel, monochrome, black and white, sinuous organic lines, Alphonse Mucha influence

Got: Prompt of 20-40 words specifying rendering style, motif, period, composition, monochrome constraint.

If fail: Prompt too vague? Add structural specifics from Step 2. Too complex (over 50 words)? Simplify by removing adjectives, keeping only structural essentials. Z-Image responds best to clear, specific prompts — avoid abstract or conceptual language.

Step 4: Configure Generation Parameters

Select resolution and generation parameters appropriate to application context.

Resolution by Application:
┌────────────────────┬─────────────────────┬────────────────────────────────┐
│ Application        │ Recommended         │ Rationale                      │
├────────────────────┼─────────────────────┼────────────────────────────────┤
│ Medallion / Roundel│ 1024x1024 (1:1)     │ Radial symmetry needs square   │
│ Tile / Repeat Unit │ 1024x1024 (1:1)     │ Square for seamless tiling     │
│ Horizontal Frieze  │ 1280x720 (16:9)     │ Wide format for running border │
│ Vertical Panel     │ 720x1280 (9:16)     │ Portrait format for columns    │
│ Wide Border        │ 1344x576 (21:9)     │ Ultrawide for architectural    │
│ General / Flexible │ 1152x896 (9:7)      │ Balanced landscape format      │
│ Large Detail       │ 1536x1536 (1:1)     │ Higher res for fine line work  │
└────────────────────┴─────────────────────┴────────────────────────────────┘
  1. Select resolution based on application context
  2. Set steps to 8 (default) for initial generation; increase to 10-12 for fine line detail
  3. Set shift to 3 (default) unless experimenting
  4. Choose random_seed: true for exploration or random_seed: false with specific seed for reproducibility
  5. Record all parameters for documentation

Got: Complete parameter set ready for generation: resolution, steps, shift, seed strategy.

If fail: Unsure about resolution? Default to 1024x1024 (1:1) — works for most ornamental contexts and fastest to generate.

Step 5: Generate Image

Invoke Z-Image MCP tool to produce ornament.

  1. Call mcp__hf-mcp-server__gr1_z_image_turbo_generate with:
    • prompt: constructed prompt from Step 3
    • resolution: from Step 4
    • steps: from Step 4
    • shift: from Step 4
    • random_seed: from Step 4
    • seed: specific seed if random_seed is false
  2. Record returned seed value for reproducibility
  3. Note generation time

Got: Generated image and seed value. Image should show recognizable ornamental forms in monochrome.

If fail: MCP tool unavailable? Verify hf-mcp-server configured (see configure-mcp-server or troubleshoot-mcp-connection). Tool available but returns error? Simplify prompt and retry. Generated image entirely abstract with no ornamental character? Prompt needs more specific structural language — return to Step 3.

Step 6: Evaluate Against Style Criteria

Assess generated image against four criteria.

Monochrome Ornament Evaluation Rubric:
┌─────────────────────┬───────────────────────────────────────────────────────┐
│ Criterion           │ Evaluation Questions                                  │
├─────────────────────┼───────────────────────────────────────────────────────┤
│ 1. Symmetry         │ Does the design exhibit the intended symmetry type?   │
│                     │ Is it visually balanced? Are repeating elements       │
│                     │ consistent?                                           │
├─────────────────────┼───────────────────────────────────────────────────────┤
│ 2. Monochrome       │ Is the image truly black and white? Are there         │
│    Fidelity         │ unwanted grays, colors, or gradients? Does the        │
│                     │ rendering style match the request?                    │
├─────────────────────┼───────────────────────────────────────────────────────┤
│ 3. Period Accuracy  │ Would this design be recognizable as belonging to     │
│                     │ the specified period? Are the motifs period-           │
│                     │ appropriate? Does it avoid anachronistic elements?    │
├─────────────────────┼───────────────────────────────────────────────────────┤
│ 4. Detail Level     │ Is the level of detail appropriate for the rendering  │
│                     │ style? Line art should have clean lines; woodcut      │
│                     │ should show bold strokes; engraving should show       │
│                     │ systematic hatching.                                  │
└─────────────────────┴───────────────────────────────────────────────────────┘
  1. Score each criterion: Strong (clearly meets), Adequate (partially meets), Weak (does not meet)
  2. Note specific observations for each criterion
  3. 3+ criteria score Strong? Design successful
  4. 2+ criteria score Weak? Return to Step 3 for prompt refinement

Got: Scored evaluation with specific observations. Most first-generation images score Adequate on 2-3 criteria.

If fail: All criteria score Weak? Prompt may be too abstract or too complex. Simplify to most essential elements: one motif, one rendering style, explicit "monochrome black and white" constraint. Consider switching to period with higher mono suitability.

Step 7: Iterate or Finalize

Refine design through targeted iteration or accept result.

Iteration Strategies:

  1. Seed-locked refinement: Keep same seed, adjust prompt slightly — evolves composition while maintaining basic structure
  2. Random exploration: Use random_seed: true with same prompt — produces variations on same concept
  3. Prompt evolution: Modify specific elements (change rendering style, add/remove motif details, adjust composition)

Iteration Budget: Limit to 3 iterations per design concept. Result not satisfactory after 3 iterations? Reconsider period/motif combination or rendering style fundamentally.

  1. Evaluation in Step 6 indicates specific weaknesses? Adjust prompt to address them:
    • Weak symmetry → add "perfectly symmetrical" or "mirror symmetry"
    • Color leaking → add "pure black and white, no gray tones, no color"
    • Wrong period feel → add specific period reference artists or monuments
    • Insufficient detail → increase steps to 10-12, add "highly detailed"
  2. Regenerate using Step 5
  3. Re-evaluate using Step 6
  4. Accept when 3+ criteria score Strong or iteration budget exhausted

Got: Improved image after 1-2 iterations, or decision to accept current best result.

If fail: Iteration not improving results? Fundamental prompt concept may not translate well to model. Try different motif from same period, or switch rendering style entirely (e.g., from line art to silhouette).

Step 8: Document the Design

Create complete record of final design for reproducibility and reference.

  1. Record:
    • Period: Historical period name and date range
    • Motif: Primary motif(s) used
    • Rendering Style: Line art, silhouette, woodcut, etc.
    • Final Prompt: Exact prompt that produced accepted image
    • Seed: Seed value for reproduction
    • Resolution: Resolution used
    • Steps/Shift: Generation parameters
    • Evaluation: Brief notes on four criteria scores
    • Iterations: Number of iterations and key changes made
  2. Note any art historical observations — how generated design compares to historical examples
  3. Suggest potential applications: print, digital border, textile pattern, etc.

Got: Reproducible record allowing exact image to be regenerated and design lineage understood.

If fail: Documentation feels excessive? Minimum: record final prompt and seed — these two values sufficient to reproduce image.

Key Motifs Reference

The following motifs appear across multiple historical periods and form core vocabulary of classical ornament:

  • Acanthus: Deeply lobed leaf; Greek origin, dominant in Roman and Renaissance ornament
  • Palmette: Fan-shaped leaf cluster; Egyptian and Greek, ancestor of anthemion
  • Anthemion: Alternating palmette-and-lotus frieze; Greek, endlessly adapted
  • Guilloche: Interlocking circles forming chain; ancient, universal
  • Meander / Greek Key: Angular spiral forming continuous band; quintessentially Greek
  • Arabesque: Infinitely extending vegetal scroll; Islamic, non-representational by principle
  • Trefoil / Quatrefoil: Three/four-lobed forms within circle; Gothic tracery
  • Rosette: Radially symmetric flower form; universal across all periods
  • Scroll (C and S): Spiraling forms; Baroque and Rococo signature elements
  • Grotesque: Fantastical human-animal-vegetal hybrid; Roman, revived in Renaissance
  • Interlace / Knotwork: Woven bands without beginning or end; Celtic, Islamic, Byzantine
  • Lotus: Stylized water lily; Egyptian origin, spread across Asian ornament traditions

Checks

  • Specific historical period selected with rationale
  • Motif structure analyzed (symmetry, scaffold, fill, edge treatment)
  • Prompt includes explicit monochrome constraint ("black and white" or equivalent)
  • Prompt specifies rendering style (line art, silhouette, woodcut, etc.)
  • Resolution matches application context
  • Generated image evaluated against 4-point rubric
  • Seed value recorded for reproducibility
  • Final design documented with prompt, seed, parameters

Pitfalls

  • Omitting monochrome constraint: Z-Image defaults to color. Without explicit "monochrome, black and white" in prompt, you get color output. Add constraint early in prompt, not as afterthought
  • Over-specifying prompt: Prompts over 50 words tend to produce confused results. Keep to one motif, one rendering style, one composition type. Quality comes from clarity, not quantity
  • Ignoring period grammar: Each period has structural rules. Gothic trefoils inside Egyptian frames, or Baroque scrolls in Greek meander borders, produce visual incoherence. Stay within period vocabulary
  • Expecting vector output: Z-Image produces raster images. For true vector line art, generated image serves as reference for manual tracing, not final production asset
  • Skipping structural analysis: Jumping from period selection to prompt without analyzing motif structure produces generic "decorative" results rather than historically grounded ornament

See Also

  • ornament-style-color — polychromatic companion to this skill; adds color palette definition and color-to-structure mapping
  • meditate — focused attention and visual imagination practices can inform ornamental composition
  • review-web-design — design review principles (visual hierarchy, rhythm, balance) apply directly to ornamental composition

GitHub リポジトリ

pjt222/agent-almanac
パス: i18n/caveman/skills/ornament-style-mono
0
agentsagentskillsai-assisted-developmentclaude-codeskillsteams

関連スキル

content-collections

メタ

このスキルは、Content Collections(Markdown/MDXファイルを型安全なデータコレクションに変換するTypeScriptファーストのツール)の本番環境でテストされた設定を提供します。Zodバリデーションによる型安全性を実現し、ブログ、ドキュメントサイト、コンテンツ重視のVite + Reactアプリケーション構築時にご利用ください。Viteプラグインの設定、MDXコンパイルから、デプロイ最適化、スキーマバリデーションまで、すべてを網羅しています。

スキルを見る

polymarket

メタ

このスキルは、開発者がPolymarket予測市場プラットフォームを活用したアプリケーション構築を可能にします。API統合による取引や市場データの取得に加え、WebSocketを介したリアルタイムデータストリーミングにより、ライブ取引や市場活動を監視できます。取引戦略の実装や、ライブ市場更新を処理するツールの作成にご利用ください。

スキルを見る

creating-opencode-plugins

メタ

このスキルは、開発者がコマンド、ファイル、LSP操作など25種類以上のイベントタイプにフックするOpenCodeプラグインを作成することを支援します。JavaScript/TypeScriptモジュール向けに、プラグイン構造、イベントAPI仕様、および実装パターンを提供します。カスタムイベント駆動ロジックでOpenCode AIアシスタントのライフサイクルをインターセプト、監視、または拡張する必要がある場合にご利用ください。

スキルを見る

sglang

メタ

SGLangは、高性能なLLMサービングフレームワークであり、RadixAttentionプレフィックスキャッシュを活用したJSON、正規表現、エージェントワークフロー向けの高速で構造化された生成を特長とします。特にプレフィックスが繰り返されるタスクにおいて、大幅に高速な推論を実現し、複雑な構造化出力やマルチターン対話に最適です。制約付きデコードが必要な場合や、広範なプレフィックス共有を伴うアプリケーションを構築する場合は、vLLMなどの代替案ではなくSGLangを選択してください。

スキルを見る