feat: models have max_tokens, temperature, closes #330

[nevercast]

• Added max_tokens and temperature options to each of the models
• Prompter gets max_tokens and temperature from profile.
• Prints these settings at startup if set.
• Throws if temperature is set and out of range

Worth considering in future (another issue/PR):

Expose other options for the LLMs, perhaps having "max_tokens" and "temperature" as top level settings is not ideal, I would suggest:

• we have model.options, and we splat those into the API call allowing setting of any options. Appropriate transformations for API that do not follow OpenAI conventions, such as Gemini.
• or, we have a known list of options in model.* such as temperature and max_tokens. Becomes max_completion_tokens for OpenAI specifically, necessary to be o1 compatible.

OpenAI and Anthropic tested as I have accounts with these providers. Referred to official documentation for all other changes. Profiles have not been changed nor has any documentation (README).

Happy to receive criticism.