DrawBot Designer

Create professional posters, layouts, and graphics with automatic enforcement of design principles.

Quick Start

1. Read design-vocabulary.md to translate user intent → design choices
2. Use templates/ to start implementation
3. Check examples.md for working patterns
4. See reference.md for API details

Design Thinking (BEFORE Code)

Before touching templates, commit to a bold creative direction:

1. Purpose & Context

• What problem does this solve? (inform, persuade, celebrate, warn?)
• Who sees this? (executives, students, general public, insiders?)
• Where does it live? (wall, hand, screen, street?)

2. Aesthetic Direction

Pick an extreme. Mediocrity is the enemy. Choose from:

Direction	Characteristics
Swiss Modernism	Grid worship, Helvetica, mathematical precision, asymmetry
Punk Zine	Photocopied texture, ransom-note type, deliberate chaos
Japanese Minimalism	Vast emptiness, one perfect element, asymmetric balance
Constructivist	Diagonal energy, red/black, bold geometry, propaganda feel
Psychedelic	Melting forms, vibrating colors, horror vacui
Corporate Brutalism	Oversized type, stark contrast, confrontational
Art Deco	Geometric ornament, gold accents, symmetry, luxury
Editorial/Magazine	Pull quotes, drop caps, sophisticated grid breaks
Bauhaus	Primary colors, geometric shapes, form follows function
Vernacular/Found	Hand-painted signs, imperfect type, authentic roughness
Tech Noir	Dark themes, neon accents, high contrast, terminal aesthetic
Organic/Natural	Flowing forms, earth tones, hand-drawn elements

3. The Memorable Thing

Ask: "What's the ONE thing someone will remember 5 seconds after looking away?"

• A massive word?
• An unexpected color?
• A jarring juxtaposition?
• Perfect silence (whitespace)?

If you can't answer this, the design isn't ready.

4. Then Implement

Only after committing to direction → open design-vocabulary.md → map to technical choices → use the design system.

---

Workflow: Natural Language → Design

Step 0: Interpret Intent (BEFORE coding)

When user says something like "create a bold modern poster":

1. Ask the design thinking questions above
2. Open design-vocabulary.md
3. Look up mood words: "bold" → high contrast, large title; "modern" → asymmetric, sans-serif
4. Identify content type: poster → announcement pattern
5. Commit to a memorable element
6. Then implement using the design system

Example Translation

User says: "Make an elegant invitation for a gala"

Design thinking:

• Purpose: Celebrate, make guests feel special
• Audience: Wealthy donors, formal crowd
• Direction: Art Deco meets Japanese Minimalism
• Memorable thing: One word in gold, swimming in cream space

Technical mapping:

• "elegant" → symmetric grid, serif type, muted colors, balanced whitespace
• "invitation" → announcement pattern (WHAT → WHEN → WHERE)

Result:

• Grid: 12×8, centered
• Font: Didot or similar high-contrast serif
• Colors: Warm cream, charcoal, gold accent
• Structure: Event name (HUGE, gold) → Date → Venue → RSVP
• The memorable thing: "GALA" in 200pt gold, nothing else competing

When to Use This Skill

Activate when the user requests:

• Posters, layouts, graphics, or editorial designs
• Modifications to existing DrawBot code
• Typography or layout advice
• DrawBot script creation or debugging

Core Principles

This skill automatically enforces:

• Typography: Hochuli's Detail in Typography (microtypography)
• Layout: Müller-Brockmann grid systems (macrotypography)
• Hierarchy: CRAP principles (Contrast, Repetition, Alignment, Proximity)

Mandatory Workflow

CRITICAL: Always use the design system. Never write manual calculations.

Step 1: Setup

import sys
from pathlib import Path
sys.path.insert(0, str(Path(__file__).parent.parent / "lib"))

import drawBot as db
from drawbot_grid import Grid
from drawbot_design_system import (
    POSTER_SCALE,        # or MAGAZINE_SCALE, BOOK_SCALE, REPORT_SCALE
    get_output_path,
    draw_wrapped_text,
    setup_poster_page
)

Step 2: Create Page and Grid

# Page setup (creates canvas automatically)
WIDTH, HEIGHT, MARGIN = setup_poster_page("letter", margin_ratio=1/10)

# Grid (automatically reads canvas size)
grid = Grid.from_margins(
    (-MARGIN, -MARGIN, -MARGIN, -MARGIN),
    column_subdivisions=12,
    row_subdivisions=16
)

Step 3: Use Semantic Coordinates

# ✅ CORRECT: Grid coordinates
header = (*grid[(0, 14)], *grid*(12, 2))  # Full width, top 2 rows

# ❌ WRONG: Manual calculations
header = (MARGIN, HEIGHT - 200, WIDTH - MARGIN*2, 150)

Step 4: Wrap Text Properly

# ✅ CORRECT: Point-based wrapping
draw_wrapped_text(text, x, y, width, height, font, size)

# ❌ WRONG: Character-count heuristics
wrapped = textwrap.wrap(text, width=70)

Step 5: Save with Portable Paths

# ✅ CORRECT: Works on any machine
db.saveImage(str(get_output_path("output.pdf")))

# ❌ WRONG: Hardcoded path
db.saveImage("/Users/you/...")

Resources

This Skill Directory

• examples.md - Working code examples
• reference.md - Complete API reference
• filters.md - Image effects, textures, blend modes
• templates/ - Starting templates:
  • minimal_poster.py - Simple poster
  • two_column.py - Magazine layout
  • card_layout.py - Color-coded cards

Project Files

• Complete examples: ../../examples/
  • minimal_poster_example.py - 80-line working poster
  • longitudinalbench_poster_v7.py - 352-line production poster
• Documentation: ../../docs/
  • design-system-usage.md - Complete usage guide
  • layout-design-principles.md - Grid theory, CRAP, decision matrices
  • typography-style-guide.md - Hochuli's spacing, line length, readability
  • print-production-checklist.md - Bleed, CMYK, export settings
• Assets: ../../assets/ - 1,807 textures

Typography Scales

Context	Scale	Base	Ratio	When to Use
Poster	POSTER_SCALE	18pt	1.5	Posters, displays
Magazine	MAGAZINE_SCALE	11pt	1.25	Magazines, newsletters
Book	BOOK_SCALE	11pt	1.2	Books, long-form
Report	REPORT_SCALE	12pt	1.25	Reports, docs

Access sizes:

scale = POSTER_SCALE
scale.caption   # 12pt
scale.body      # 18pt
scale.h3        # 27pt
scale.h2        # 40.5pt
scale.h1        # 60.75pt
scale.title     # 91.125pt

Color Palette Generation

Generate harmonious color palettes from a single base color:

from drawbot_design_system import (
    generate_color_palette,
    hex_to_rgb,
    check_contrast_ratio,
    get_accessible_text_color
)

# Generate palette from base color
base = hex_to_rgb("#2E86AB")  # Ocean blue
palette = generate_color_palette(base, harmony="complementary")
# Returns: {'background': (light), 'text': (dark), 'accent': base, 'accent2': complement}

# Available harmonies:
# - "complementary": Base + opposite (high contrast)
# - "analogous": Base + neighbors (harmonious)
# - "triadic": Base + 2 evenly spaced (vibrant)
# - "split_complementary": Base + split opposite (balanced)
# - "monochromatic": Base + light/dark variations (subtle)

# Check accessibility
ratio, level = check_contrast_ratio(palette['text'], palette['background'])
print(f"Contrast: {ratio:.1f}:1 ({level})")  # "Contrast: 8.2:1 (AAA)"

# Get accessible text for any background
text_color = get_accessible_text_color(palette['accent'])

OpenType Features

Enable professional typography features:

from drawbot_design_system import (
    set_opentype_features,
    get_available_opentype_features
)
import drawBot as db

# Check what features a font supports
db.font("Adobe Garamond Pro")
features = get_available_opentype_features()
print(features)  # {'liga': 'Standard ligatures', 'smcp': 'Small capitals', ...}

# Enable small caps + oldstyle figures
set_opentype_features(['smcp', 'onum'])
db.text("The Quick Brown Fox 1234", (x, y))

# Common features:
# - 'liga': Standard ligatures (fi, fl, ff)
# - 'smcp': Small capitals
# - 'onum': Oldstyle figures (varying heights)
# - 'tnum': Tabular figures (monospaced)
# - 'frac': Automatic fractions (1/2 → ½)
# - 'ss01'-'ss20': Stylistic sets

Variable Fonts

Control variable font axes dynamically:

from drawbot_design_system import (
    set_font_variation,
    get_font_variation_axes
)
import drawBot as db

# Check available axes
db.font("Skia")  # Or any variable font
axes = get_font_variation_axes()
print(axes)  # {'wght': {'minValue': 100, 'maxValue': 900, ...}, ...}

# Set axis values
set_font_variation(wght=600, wdth=85)  # Semi-bold, slightly condensed
db.text("Variable Font Text", (x, y))

# Common axes:
# - wght: Weight (100=thin, 400=regular, 700=bold, 900=black)
# - wdth: Width (50=condensed, 100=normal, 200=extended)
# - slnt: Slant (-90 to 90 degrees)
# - ital: Italic (0=upright, 1=italic)
# - opsz: Optical size (optimize for display vs text)

Print Production

Set up print-ready documents with bleed:

from drawbot_design_system import (
    setup_print_page,
    validate_print_ready,
    PRINT_PRESETS
)
import drawBot as db

# Set up page with bleed (US standard: 0.125")
specs = setup_print_page("letter", include_bleed=True)

# Draw background to FULL canvas (into bleed area)
db.fill(0.2, 0.4, 0.8)
db.rect(0, 0, specs['canvas_width'], specs['canvas_height'])

# Position content relative to trim + safe margin
content_x = specs['trim_x'] + specs['safe_margin']
content_y = specs['trim_y'] + specs['safe_margin']

# Use CMYK for print colors
db.cmykFill(0.6, 0.4, 0.4, 1.0)  # Rich black

# Validate before export
is_valid, warnings = validate_print_ready()
for w in warnings:
    print(w)

# See docs/print-production-checklist.md for complete guide

Grid Patterns

Full Width Section

header = (*grid[(0, 14)], *grid*(12, 2))  # All columns, 2 rows

Two Columns

left = (*grid[(0, 1)], *grid*(5, 13))    # Cols 0-4
right = (*grid[(7, 1)], *grid*(5, 13))   # Cols 7-11
# Columns 5-6 = automatic gutter

Stacked Sections

section1 = (*grid[(0, 10)], *grid*(12, 3))  # Rows 10-12
section2 = (*grid[(0, 6)], *grid*(12, 3))   # Rows 6-8
section3 = (*grid[(0, 2)], *grid*(12, 3))   # Rows 2-4

Decision Matrix

Content	Grid	Typography	Line Length
Poster	12×16	POSTER_SCALE	20-30 CPL
Magazine	12×8	MAGAZINE_SCALE	45-50 CPL
Book	6×8	BOOK_SCALE	60-65 CPL
Report	12×8	REPORT_SCALE	50-60 CPL

Verification Checklist

Before finalizing code, verify:

• Imports from lib/ directory
• Used setup_poster_page() BEFORE grid
• Grid created with Grid.from_margins()
• Used pre-defined scale (POSTER_SCALE, etc.)
• All layout uses grid coordinates (not manual math)
• All text uses draw_wrapped_text() (not textwrap.wrap)
• Paths use get_output_path() (not hardcoded)
• No fontSize approximations

Error Prevention

Never Do This (Technical)

❌ Hardcode page size in grid ❌ Use character counts for wrapping ❌ Approximate line height with fontSize ❌ Manual calculations instead of grid ❌ Truncate text with [:N] ❌ Hardcode absolute paths

Always Do This (Technical)

✅ Let grid read canvas size ✅ Use point-based wrapping ✅ Use real font metrics ✅ Use semantic grid coordinates ✅ Draw all text that fits ✅ Use portable path helpers

---

Avoid "AI Slop" Aesthetics

CRITICAL: These patterns make work look machine-generated. Avoid them:

Typography Anti-Patterns

❌ Helvetica Neue for everything — It's fine, but predictable. Try Akzidenz-Grotesk, Univers, or something unexpected. ❌ Safe font pairings (Montserrat + Open Sans) — Too common. Be bolder. ❌ Centered everything — Easy but lazy. Asymmetry creates energy. ❌ Even text sizes (24pt, 18pt, 14pt) — Use the scale ratios; they create proper tension. ❌ Justified text in narrow columns — Creates rivers of whitespace.

Color Anti-Patterns

❌ Blue gradient on white — The default "tech startup" look. ❌ Purple/pink gradients — Overused in AI-generated content. ❌ Gray text on white (#666 on #fff) — Low contrast, looks washed out. ❌ Rainbow gradients — Almost never appropriate. ❌ Equal color distribution — Use 70-20-10, not 33-33-33.

Layout Anti-Patterns

❌ Perfect symmetry everywhere — Real design has controlled asymmetry. ❌ Everything centered vertically AND horizontally — Creates static, lifeless compositions. ❌ Uniform margins — Vary them intentionally; tension creates interest. ❌ Clip art shapes (perfect circles, rounded rectangles) — Too geometric, too safe. ❌ Stock photo compositions — If it looks like a template, it is.

The "AI Look" Checklist

If your design has 3+ of these, reconsider:

• Centered text blocks
• Blue/purple color scheme
• Generic sans-serif font
• Perfect geometric shapes
• Even spacing everywhere
• No memorable focal point
• Could be any company's poster

The Antidote

Make ONE bold choice that a template wouldn't:

• Extremely large type (bigger than feels comfortable)
• Extremely small type (challenges the reader)
• Unexpected color (not blue, not purple)
• Deliberate asymmetry
• Texture or grain
• One element that breaks the grid

Common Tasks

Create a Poster

1. Copy templates/minimal_poster.py to ../../examples/ or your working directory
2. If copying to examples/, update the import path to use .parent.parent instead of .parent.parent.parent.parent.parent
3. Update title, subtitle, body text
4. Change filename in get_output_path()
5. Run: uv run python your_file.py

Two-Column Layout

1. Copy templates/two_column.py to ../../examples/
2. Update the import path if needed (see Create a Poster above)
3. Update column content
4. Adjust grid if needed (more/fewer rows)

Color-Coded Cards

1. Copy templates/card_layout.py to ../../examples/
2. Update the import path if needed (see Create a Poster above)
3. Modify card titles and descriptions
4. Change colors if desired

Debug Layout Issues

1. Add grid.draw(show_index=True) before saveImage()
2. View PDF to see grid structure
3. Verify grid coordinates
4. Remove grid.draw() for final output

Troubleshooting

Problem	Cause	Fix
Text overflows	Character-count wrapping	Use draw_wrapped_text()
Grid doesn't match	Created before newPage()	Create page FIRST
Wrong sizes	Wrong scale for context	Use POSTER_SCALE for posters
Paths don't work	Hardcoded paths	Use get_output_path()

See reference.md for detailed troubleshooting.

Example Session

User: "Create a poster for an AI conference"

Response:

1. Copy templates/minimal_poster.py
2. Modify content:
   • Title: "AI Conference 2025"
   • Subtitle: "Advancing Machine Learning"
   • Body: Conference details
3. Save as ai_conference.py
4. Run: uv run python examples/ai_conference.py

Result: Professional poster following all design principles automatically.

Progressive Disclosure

Start with templates, refer to examples and reference as needed:

1. Start: templates/minimal_poster.py
2. Learn: examples.md for patterns
3. Reference: reference.md for API details
4. Effects: filters.md for images/textures (when needed)
5. Deep dive: ../../docs/ for theory

Claude loads resources progressively as needed (saves context).

Force Variation

No two designs should look the same. Before finalizing, ask:

Variation Prompts

• "What if the title was 3x larger?"
• "What if I used only one color?"
• "What if text was at the bottom instead?"
• "What if margins were twice as wide?"
• "What if I removed everything except the essential?"
• "What if this was for the opposite audience?"

Rotation Mandates

Cycle through these to avoid convergence:

Fonts (never default to the same one):

• This project: Serif
• Next project: Geometric sans
• Next: Humanist sans
• Next: Slab serif
• Next: Display/decorative

Color temperature:

• This project: Warm palette
• Next: Cool palette
• Next: Neutral with warm accent
• Next: Monochrome

Layout energy:

• This project: Asymmetric, dynamic
• Next: Symmetric, calm
• Next: Vertical emphasis
• Next: Horizontal bands

---

The Creative Mandate

DrawBot is capable of extraordinary creative work.

The design system handles the rules—grids, scales, metrics. That's the foundation.

But foundations are invisible. What people see is the vision on top.

Don't settle for "correct." Aim for unforgettable.

The difference between good and great:

• Good: Follows the grid, uses the scale, wraps text properly
• Great: Does all that AND has one moment of genuine surprise

Every poster should have that moment. Find it before you write code.

Trust the system. Then transcend it.

---

Remember

The design system makes it impossible to violate design principles. That's the floor, not the ceiling.

Hochuli gives you the rules. Müller-Brockmann gives you the grid. You give it soul.

Version History

• v1.2.0 (2025-12-31): Added design thinking, aesthetic directions, anti-patterns, variation mandates
• v1.1.0 (2025-12-31): Added color harmony, OpenType/variable fonts, print production
• v1.0.0 (2025-10-18): Initial Agent Skill with curated content