docs: highlight new Ollama (local) preset in README - #128

Thanks for taking the time to open this, and for reading the local-model section closely enough to want it clearer.

I'm going to pass on this one, for three reasons:

It's already covered. Two paragraphs above the insertion point, the section walks through exactly this: open the LLM tab, add a provider, pick Ollama (local), and the wizard skips the key screen because a local server has no API key. The preset table right below the new line lists Ollama (local) | http://localhost:11434, and the paragraph after it separates the local preset from Ollama Cloud. The added sentence restates its neighbours rather than adding to them.

The auto-detect claim isn't accurate. The preset does not detect anything. It's a hardcoded default (baseUrl: "http://localhost:11434", flagged local) in src/tui/providers/provider-presets.ts. There's no port probe and no check that a server is running — if ollama serve isn't up, you find out when the model list comes back empty. Shipping "auto-detects your local Ollama server" would set an expectation the code doesn't meet.

We don't keep version badges in the README. There are no other "New in vX" markers in the file, and the preset landed back in v0.2.1 (#122) — we're several releases past that now, so the badge would read as stale the day it merged. Release notes are where that belongs.

If you'd like to improve this section, here's a gap that's actually open: the README never says what happens when the Ollama server isn't running — the model list simply comes back empty, with no hint to go start it. A sentence covering that would be a real addition, and I'd be glad to review it.

Opened a new PR with the requested rework and the relevant updates consolidated into a clean follow-up: #204