Applixus Oss Benchmark Marketplace
- 1 repo stars
- Updated July 16, 2026
- Applixus-A-S/applixus-oss-model-comparison-skill
llm-ux-benchmark is a Claude Code plugin that rebuilds a reference site per LLM and scores outputs with a shared rubric so developers can compare which model fits which UX rebuild task.
About
applixus-oss-benchmark marketplace ships the llm-ux-benchmark Claude Code plugin that assigns real rebuild tasks to LLMs and scores results with a transparent rubric and leaderboard. Use it when you need evidence for which model handles frontend rebuild and UX quality for a given site instead of anecdotal preference. Commands cover rebuild, Codex scoring with optional PageSpeed input, and cross-model comparison tables documented in the README.
- /llm-ux-benchmark:rebuild recreates a URL as vanilla static HTML/CSS/JS per model
- Shared spec and rubric in lib/RUBRIC.md; scoring run neutrally by Codex
- /llm-ux-benchmark:compare builds a leaderboard across scorecards
- Measures technical quality, PageSpeed, UI fidelity, and readability
Applixus Oss Benchmark by the numbers
- Data as of Aug 10, 2026 (Skillselion catalog sync)
/plugin marketplace add Applixus-A-S/applixus-oss-model-comparison-skillAdd your badge
Show developers this marketplace is listed on Skillselion. Paste this into your README.
| repo stars | ★ 1 |
|---|---|
| Last updated | July 16, 2026 |
| Repository | Applixus-A-S/applixus-oss-model-comparison-skill ↗ |
How do you compare LLMs on a real rebuild task with neutral scoring instead of asking which model is generically better?
Benchmark LLMs by rebuilding a target site as vanilla static HTML/CSS/JS, scoring with Codex, and comparing models on a shared rubric and leaderboard.
Who is it for?
Developers evaluating GLM, Opus, Kimi, or other backends on identical site-rebuild benchmarks with Codex judging and leaderboard output.
Skip if: Teams that already picked a model and do not need rubric-based rebuild comparisons or static-site reproduction benchmarks.
What you get
Per-model rebuild folders, scorecards, optional PageSpeed inputs, and a comparison leaderboard under runs/ per the README.
Plugins in this marketplace
1 plugin - install individually after you add the marketplace.
Recommended Marketplaces
Abel30567 Skillsabel30567-skills provides support for development workflows, enabling teams to work more efficiently.
Ably Agent Skillsably-agent-skills is the official Ably Claude Code marketplace that packages realtime messaging expertise into the ably plugin with using-ably and debugging-with-ably-cli skills.1
Agent Brain MarketplaceAgent Brain Marketplace is a Spillwave Solutions Claude Code plugin for code-first RAG document and source search.116
AgentkeyAgentkey is a Claude Code marketplace plugin for AgentKey, described in the manifest as a unified API for real-time external data spanning web search, social media, crypto and blockchain data, and web scraping.585
Ai Discoveryai-discovery is a Claude Code marketplace for llm integration.3
AliveAlive Context is a Claude Code plugin that keeps coding agents context-aware across sessions, preserving architecture notes, prior edits, and decisions so multi-day implementation work stays coherent.125FAQ
What task do models perform?
Rebuild a given site from scratch as vanilla static HTML/CSS/JS with no build step per the README.
Who scores UI and code quality?
Codex acts as neutral judge; you supply PageSpeed after deploy via --pagespeed N.
What is the composite score?
Equal-weighted average of technical quality, PageSpeed, UI quality, and readability per lib/RUBRIC.md.