# Decision Tracing

> Reverse-engineer why a design decision was made. Move from observation to inference. Use when studying products you admire, learning from competitors, or building analytical design thinking.

- Skill: `dragoon0x/decision-tracing` (Agent Skill)
- Install (CLI): `npx skillmds@latest add dragoon0x/decision-tracing`
- Raw SKILL.md: https://api.skillmd.com/api/skills/dragoon0x/decision-tracing/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: AI & ML
- Author: dragoon0x (https://skillmd.com/u/dragoon0x)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/dragoon0x/decision-tracing

---


# Decision Tracing

Look at a decision. Figure out why it was made.

## How to use

- `/decision-tracing` Apply decision-tracing analysis to any design in this conversation.

## Constraints

### Trace Structure
- MUST follow: Observation > Decision > Probable Reason
- "The primary font is Inter" > "They chose a geometric sans-serif" > "Data-heavy product needing maximum legibility at small sizes with neutral personality"
- MUST generate at least 5 traces per design screen
- SHOULD acknowledge uncertainty: "probable reason" not "the reason"
- NEVER state intent as fact unless you have insider knowledge

### What to Trace
- Typeface selection and type system
- Color palette and color role assignments
- Layout structure and grid decisions
- Spacing and density choices
- Navigation architecture
- What's absent (what they chose NOT to include)

### Anti-Patterns
- Assuming the most flattering interpretation of every decision
- Tracing only visual decisions and ignoring structural ones
- Stating reasons as facts without marking them as hypotheses