best-value-ai

Use this skill when the user is deciding how to pay for AI — comparing local GPUs, API pricing, or subscription plans like ChatGPT Plus/Pro/Business, Claude Pro/Max 5x/20x, or asking questions like "which AI gives me the most for my money", "is ChatGPT Pro worth it", "should I run a local LLM", "best LLM for coding on my hardware", or "what's the cheapest way to get Claude Sonnet". Provides data-driven recommendations using quality-adjusted tokens per dollar across 34+ models, 36 hardware configurations, and measured subscription quotas from desktopcommander.app/best-value-ai.

Use this skill when the user wants to contribute a subscription quota measurement to the Best Value AI dataset. Specifically — when they say "measure my Claude Max quota", "benchmark ChatGPT Business", "contribute data to best-value-ai", "submit a measurement", or "help me measure how many tokens I actually get on my plan". Runs the standardized measurement script for Claude Code or Codex CLI, captures token/quota data, validates the output, and opens a pull request to desktop-commander/best-value-ai.