Repurpose MattOS as a local-first Obsidian second brain

[MWolf10925]

Summary

• add the local-first Obsidian MattOS workflow and templates
• add reliable automatic dump watching with SHA256/debounce state
• add Ollama preflight, timeouts, fallback, diagnostics, and optional dedicated mode
• add Windows launchers, startup helpers, health checks, and watcher debugging
• keep personal vault notes, archives, logs, and machine-specific Obsidian state out of Git

Verification

• Python tools compile successfully
• MattOS health check passes
• Ollama qwen3.5:4b strict JSON test passes
• watcher auto-processing and run-lock behavior were tested end-to-end

The original kernel files are intentionally not deleted by this commit.

[MattRyder]

Hi @MWolf10925, just wondered if this was a valid request that you really looked to merge into my OS repo from 2011?

Looks to have been Claude-authored, so I'm assuming not, and if so I'll close this PR.
And some unsolicited advice, if I were starting out in software, I'd probably try to keep AI at arm's length. You can produce a fantastic amount of code with it in seconds, but as with this PR, you can find that you don't really know what it does, and you never get out of advanced beginner mode.

Less code is always better. No code is best. Write what you understand and understand what you've written.