?? --status shows the loaded model name; README pick-your-shell

[FrancoisChastel]

Summary

Two small ergonomic improvements that work well together:

`?? --status` shows the running model

Before:
```
up fast → http://127.0.0.1:8091
up balanced → http://127.0.0.1:8080
up smart → http://127.0.0.1:8093
```

After:
```
up fast → Qwen3.6-35B-A3B-UD-Q4_K_M · http://127.0.0.1:8091
up balanced → Qwen3.6-27B-Q4_K_M · http://127.0.0.1:8080
up smart → Qwen3-Coder-Next-Q4_K_M · http://127.0.0.1:8093
```

The model name is read from llama-server's `/v1/models` endpoint (not from the GGUF in cache by tier name), so explicit `?? --model PATH` starts are reflected accurately and the display matches what the model would identify itself as in API responses.

README "Pick your shell" table

Right after the install steps, a compact table tells bash and fish users they're covered without having to spelunk through CONFIGURATION.md:

Shell	One-line setup	Get
zsh	source `shellllm.zsh`	the five glyphs: `,` `,,` `?` `??` `???`
bash	source `shellllm.bash`	`llmc` / `llmf` / `llma` / `llmm` aliases
fish / POSIX	nothing — call the CLIs	`shellllm-comma` / `shellllm-ask` / `shellllm-recall`

Test plan

• 264 tests pass (`pytest -q`)
• `ruff check` and `ruff format --check` clean
• `zsh -n` syntax check passes
• Verified locally — three tiers running side by side, `?? --status` shows each one's loaded GGUF name with a separator and the URL
• Manual: install fresh from the formula, verify the README table renders correctly on GitHub