What problem it solves

Frontend design is one of the hardest domains for AI — the gap between "technically correct" and "looks designed" is almost entirely tacit knowledge. This skill surfaces the visual judgment that senior designers know but rarely articulate, helping Claude produce UI designs with intentional aesthetic choices instead of generic defaults.

Who uses this workflow

Frontend developers and designers who want Claude to make deliberate design decisions (font pairing, color relationships, layout rhythm) rather than converging on template-like output (Inter + blue-500 + shadow-md + centered everything).

How it works

Three systems grounded in Michael Polanyi's tacit knowledge theory:

• Knowledge Filter — Skips what tutorials teach, extracts implicit designer reasoning
• Five Design Lenses — Gestalt diagnostics, subsidiary-focal awareness, tacit translation (feeling → concrete fix)
• Aesthetic Judgment Encoding — Six anti-convergence patterns that prevent AI default regression

Attribution

Based on Michael Polanyi's tacit knowledge framework, with insights from Harry Collins's taxonomy and the Dreyfus skill acquisition model. Aesthetic encoding patterns drawn from Anthropic's frontend-design research.

Example

Same prompt with and without the skill:

• Without: Blue gradients, Inter font, centered everything, 6 equal-width cards
• With: Deep green palette, Playfair+Inter pairing, 60/40 editorial split, deliberate asymmetry

Live A/B demo