
Replica Diff: Clone Parity Checker
Scores how closely your app clone matches the original's features and screen layout
- Repository
- Jakeschincariol/replica-skill
- Stars
- 7,057
- Licence
- MIT
- Files
- 3 files
What it does
Once installed, the agent can measure how close a cloned app is to an original. It weights a feature checklist (must, should, could) into a parity score, lists missing features in build order, and compares screenshots by layout while ignoring colour, so a rebrand doesn't count against the clone. It then writes a parity report with a shippable verdict and the top next steps.
Great for
- Developers rebuilding an existing app who want an honest completeness score
- Founders checking whether a clone is ready to ship before launch
- Designers who want to compare screen structure rather than exact pixels
Skip it if
- Pixel-perfect copying, since the skill explicitly discourages chasing the original's exact look (it changes the design on purpose)
- Comparing apps you have no legitimate access to, since screenshots must come from public pages or your own account
- Judging visual quality or brand design, since colour is deliberately ignored and the tool measures structure only
Try asking
- “How close is my clone to the original app?”
- “Run a parity check and tell me what's missing”
- “Diff the screens in replica/screens against replica/clone-screens”
- “Is my clone ready to ship? Write the parity report”
Install in 3 steps
- 1
Download the skill
The ZIP contains the 3 original files, the MIT licence and an attribution note.
replica-diff.zip - 2
Add it to your agent
Claude Code (all projects)
mkdir -p ~/.claude/skills && unzip replica-diff.zip -d ~/.claude/skills/Claude Code / Agent SDK (this project only)
mkdir -p .claude/skills && unzip replica-diff.zip -d .claude/skills/Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.
- 3
Ask for the task
No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.
Get better results
- Keep features.csv current, since the score is only as accurate as the rows you fill in
- Take screenshots at the same viewport (1440x900 desktop, 390x844 mobile) and in the same app state for meaningful layout scores
- Use --mode pixel only for your own regressions, not for comparing against the original
What's inside
- imgdiff.py
- parity.py
- SKILL.md
Redistributed unmodified from Jakeschincariol/replica-skill under the MIT licence. The assessment was drafted with AI from the skill's files on 2026-11-02. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.