mcpField
V
version-archaeology

version-archaeology

github.com/dragoon0xDesign & UI677 installsUpdated 35w ago
Repository

About version-archaeology

A design-analysis skill for tracing how a product changes across versions. It compares redesigns to identify lasting decisions, experiments that were reversed, and shifts between cosmetic and structural changes.

Key features

  1. 01Compare at least three versions
  2. 02Track changes and continuities
  3. 03Identify reverted experiments
  4. 04Find decisions that persist
  5. 05Assess cosmetic versus structural change

Use cases

  1. 01Analyzing a product’s redesign history
  2. 02Finding enduring design decisions
  3. 03Learning from reverted redesign experiments

Install

Copy the command and run it in a terminal.

$npx skills add https://github.com/dragoon0x/taste-skills --skill version-archaeology

After installing, type/version-archaeologyin Claude Code to confirm it loaded.

Details

From the project README

Version Archaeology

The decisions that survive redesigns are the load-bearing walls.

How to use

  • /version-archaeology Apply evolutionary analysis to a product's design history.

Constraints

Analysis Method
  • MUST compare at least 3 versions of the same product over time
  • MUST document for each version: what changed, what stayed, what was tried and reverted
  • MUST identify the persistent decisions (things that survived multiple redesigns)
  • SHOULD note the ratio of cosmetic changes to structural changes
What Persistence Means
  • Elements that survive 3+ redesigns are core taste decisions
  • Elements that change every redesign are cosmetic/trend-following
  • Elements that were added and then removed reveal failed experiments
  • MUST treat persistent decisions as evidence of what the team considers essential
Anti-Patterns
  • Only looking at the visual surface (layout changes) and missing structural persistence (information architecture)
  • Assuming newer versions are always better
  • Ignoring reverted changes (these are some of the most informative data points)

Comments

Markdown supported