DEDUPED Hot AI repos, turned into use cases Get the Monday 5

How to score a rebuilt clone against the original app with Replica Skill

Get a layout diff and a parity score for your clone.

Authorization notice. Replica Skill is dual-use. Only use it on software, sites and systems you own or have explicit permission to test.

Who it is forA developer who has rebuilt screens from a recon map and needs to see what is still missing before a release.

The problemJudging a clone by eye hides missing must-have features and layout gaps.

The outcomeA pixel-region layout diff and a parity score with a list of missing features.

What you need first

  • Claude or Claude Code installed
  • Git, to clone the repo or add the plugin
  • Python 3.8 or newer for the tools
  • Public pages, screenshots, or your own account for the app you want to study

Steps

  1. Save a screenshot of the original screen as original.png and your clone as clone.png
  2. Run python3 replica-diff/imgdiff.py original.png clone.png --out diff.png
  3. Make sure your feature matrix is at replica/features.csv
  4. Run python3 replica-diff/parity.py replica/features.csv
  5. Fix any missing must-haves, since the tool marks them as not shippable

Prompt for your agent

Using the replica-skill repo (github.com/Jakeschincariol/replica-skill), compare <original.png> with <clone.png> using imgdiff.py and score <replica/features.csv> with parity.py. Expected output: diff.png and a parity score with a missing list. Verify that the missing list includes every must-have that is not done.

Caveat. The scores only reflect the features you list in features.csv and the screenshots you provide.

Use this in your AI agent

Ask your agent

Using the Deduped MCP, get the setup steps for Jakeschincariol/replica-skill and "Score a rebuilt clone against the original app"

Add Deduped to Claude Code once

claude mcp add --transport http deduped https://dedupedai.com/mcp

Cursor, ChatGPT, others

Share: X LinkedIn Hacker News Email Copy link

Third-party repos we don't own. AI-written use cases can be wrong. Disclaimer