Add AMD Strix Halo local LLM guide

[hogeheer499-commits]

What

Adds the AMD Strix Halo Local LLM Guide to the existing LLM resources section.

Why

The section already links general local-LLM tools and guides, but not a reproducible hardware-specific path for AMD Ryzen AI MAX+ 395 systems. I maintain and personally use this guide on a retail Strix Halo system. It covers the setup choices that determine whether local inference works in practice, while keeping benchmark claims linked to raw evidence and preserving failed routes.

Validation

• Used the repository's requested [Resource Title](url) — description. format.
• Kept the change to one concise resource entry.
• Verified the destination and description against the current public guide.