Pull Request Title: Add Ollama Service

Description

Adds a Tailscale sidecar configuration for Ollama, a tool for running large language models (LLMs) locally. This lets users access their local models securely from any device on their Tailnet — including phones and remote machines — without exposing the Ollama API to the public internet.

Includes:

• compose.yaml following the ScaleTail sidecar pattern (network_mode: service:tailscale, health checks, depends_on with health condition)
• .env template with SERVICE, IMAGE_URL, TS_AUTHKEY, and an optional OLLAMA_API_KEY
• README.md covering prerequisites, volumes, MagicDNS/HTTPS setup, optional LAN port exposure, and first-run model pull instructions
• Optional yourNetwork external network on the Tailscale container, enabling other containers (e.g. Open WebUI) to reach Ollama via inter-container networking

Related Issues

N/A

Type of Change

• Bug fix
• New feature
• Documentation update
• Refactoring

How Has This Been Tested?

1. Ran docker compose config — no errors or missing variable warnings.
2. Pre-created bind-mount directories (config, ts/state, ollama-data) and started the stack with docker compose up -d.
3. Confirmed both containers reached healthy status via docker compose ps.
4. Verified Tailscale node registration and Tailnet IP with docker exec tailscale-ollama tailscale status.
5. Pulled tinyllama and sent a test generation request via curl to the Tailnet IP on port 11434 — received a valid JSON response.
6. Repeated the curl test from a second device on the same Tailnet to confirm remote access works end-to-end.
7. Verified tailscale-ollama appears in docker network inspect yourNetwork.

Checklist

• I have performed a self-review of my code
• I have added tests that prove my fix or feature works
• I have updated necessary documentation (e.g. frontpage README.md)
• Any dependent changes have been merged and published in downstream modules

Screenshots (if applicable)

N/A — no visual UI changes.

Additional Notes

• The yourNetwork network is optional. Users who don't need inter-container communication can remove the networks: sections from compose.yaml entirely.
• Ollama models can be large (several GB each). The README notes this under the Volumes section so users are aware before first pull.
• OLLAMA_KEEP_ALIVE is set to 24h by default to keep models warm; users can adjust or remove this to suit their hardware.