fix: add missing LLM_PROVIDER and MODEL_NAME to pydantic-ai .env.example

[brone1323]

Summary

pydantic-ai/main.py reads two env vars that were absent from .env.example:

• LLM_PROVIDER — selects the backend (openai / anthropic / gemini); defaults to "openai" silently
• MODEL_NAME — optional per-provider model override

Without these in the template, a user copying .env.example to .env would:

1. Have no idea they can switch providers without editing source code
2. Not know which default model each provider falls back to

Changes

• Added LLM_PROVIDER and MODEL_NAME with inline comments explaining accepted values and defaults
• Added a comment to the key block clarifying only the key for the chosen provider is needed
• Added a comment to GITHUB_TOKEN explaining why it is required

Test plan

• Copy .env.example to .env, fill in one API key, run python main.py -- should work without reading main.py source
• Set LLM_PROVIDER=anthropic, fill ANTHROPIC_API_KEY, run -- should use Anthropic without any code changes

Generated with Claude Code

[CLAassistant]

Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution.
_{You have signed the CLA already but the status is still pending? Let us recheck it.}