AI Skill Report Card

Designing Digital Color Systems

B+78·Nov 20, 2026·Source: Extension-page
12 / 15

Define a product line in three parts, then generate deliverables:

  1. Concept: Name the system + target use case (e.g., "Subsurface Scattering System" for character skin rendering).
  2. Architecture: List components (palettes, parameters, layer structures, blend modes) and quantify them (e.g., "12 base palettes × 8 swatches, 3 light temperature variants").
  3. Packaging: Export as .swatches/.ase/.gpl + JSON metadata, bundle with a PDF guide, and tag for marketplace SEO.

Run the build script (see Workflow Step 4) to batch-export all palettes with embedded metadata in one pass.

Recommendation▾
Add a second example showing a print-safe CMYK product line to demonstrate the stated distinction between rendering vs. reproduction palettes
13 / 15
Progress:
- [ ] Step 1: Market landscape scan — identify underserved niche + pricing tier
- [ ] Step 2: Design 2-3 product line architectures (character/print/UI are common splits)
- [ ] Step 3: Write extraction prompts to turn raw research/notes into structured specs
- [ ] Step 4: Build automation script to export swatch files + metadata at scale
- [ ] Step 5: Attach SEO tags and write actionable launch next-steps

Step 1 — Market Landscape & Pivot Strategy Identify where demand exceeds supply (e.g., anime/character shading packs, archival print-safe CMYK sets, planner/UI pastel kits). State the commercial pivot in one sentence: "Shift from generic palettes to parametrically-structured, technically-documented systems that justify premium pricing."

Step 2 — Product Architecture For each product line, define:

  • System Components: named sub-systems (e.g., base tone, shadow ramp, highlight ramp, rim-light, SSS translucency layer)
  • Palette & Parameter Structure: exact counts and relationships — hue shift rules, value steps, saturation curves, CMYK/RGB/Hex triplets
  • Keep architectures orthogonal: one for illustration/rendering, one for print/reproduction fidelity, one for UI/flat design — each serving a different buyer.

Step 3 — Extraction Prompts Write reusable prompts (for Gemini/Claude/GPT notebook tools) that convert messy research into the structured format above. Each prompt should specify: input type, required output schema (JSON/table), and the exact fields to extract (hex, HSB, use-case tag, pairing rules).

Step 4 — Technical Packaging & Automation Build a script that takes a structured palette definition (JSON/CSV) and outputs:

  • .swatches / .ase / .gpl files per software target
  • A metadata sidecar (JSON) with name, hex, RGB, CMYK, and tags
  • Optional preview PNG swatch sheet

Step 5 — SEO Tags & Launch Next Steps Generate 10-15 marketplace search tags mixing broad + niche terms. End with 3-5 concrete next actions (e.g., "Render 3 sample illustrations using Product Line A," "List on Gumroad with bundle pricing").

Recommendation▾
Include an actual sample JSON metadata schema or the full export_swatches.py script rather than just describing/referencing it
14 / 20

Example 1: Input: "I want a palette system for anime-style skin shading that looks professional." Output:

Product Line A: Character Illustration & Subsurface Scattering System
System Components:
- Base Tone Layer (1 swatch per skin tone, 6 tones)
- Core Shadow Ramp (3-step value gradient, warm-shifted)
- Subsurface Bounce Layer (red/orange translucency overlay, 20-40% opacity guide)
- Rim Light / Occlusion Layer (cool-shifted highlight, 2 swatches)
Palette & Parameter Structure:
- 6 skin tones × 8 swatches each (base, 3 shadow steps, SSS bounce, 2 highlights, 1 occlusion)
- Hue shift rule: shadows shift +10° toward red, highlights shift -5° toward yellow
- Exported as HSB triplets for easy in-app adjustment

Example 2: Input: "Turn this into a swatch export script." Output: A Python script accepting a JSON list of {name, hex, cmyk, tags} and writing an Adobe .ase binary plus a metadata.json sidecar, with a CLI example: python export_swatches.py --input palette.json --out ./build/

Recommendation▾
Show a concrete 'bad' output (e.g., flat hex list) alongside the good example to reinforce the pitfalls section with contrast
  • Always separate rendering-use palettes (illustration/SSS) from reproduction-use palettes (print-safe CMYK) — buyers search for these differently.
  • Quantify everything in deliverables ("48 swatches," "6 tone families") — specificity sells.
  • Embed metadata (hex, CMYK, RGB, use-case tags) in every export file, not just a separate readme.
  • Design extraction prompts with explicit output schemas so results are copy-paste ready into the build script.
  • Tag products with both broad terms ("color palette") and niche terms ("subsurface scattering skin shading") for discoverability.
  • Don't ship palettes as flat hex lists with no structural logic (ramps, shift rules) — this reads as amateur.
  • Don't mix illustration and print-safe palettes in one product; color gamuts and buyer intent differ.
  • Don't skip the metadata sidecar — software-specific files alone (.ase) aren't searchable or remixable by buyers.
  • Don't over-scope one product line; 3 tight, well-documented systems outsell 1 sprawling mega-pack.
0
Grade B+AI Skill Framework
Scorecard
Criteria Breakdown
Quick Start
12/15
Workflow
13/15
Examples
14/20
Completeness
15/20
Format
14/15
Conciseness
13/15