Add EdgeRunner

[christopherkarani]

Summary

• Add EdgeRunner to the AI section
• Fast, local LLM inference for Apple Silicon built in Swift and Metal from the ground up
• Supports GGUF models (Qwen3, etc.) with Metal-optimized kernels delivering ~230+ tok/s decode on M3 Max
• Placed right after Espresso (same author, complementary project — EdgeRunner does inference, Espresso compiles for ANE)

Checklist

• Entry added to contents.json only (no README.md changes)
• Short, plain description starting with capital and ending with period
• Fits AI category (LLM inference engine)
• MIT license (allowed)
• Swift 6.2+
• 15+ GitHub stars
• English README

[vpeschenkov]

@christopherkarani thanks!