Pass max output tokens into transcript prompts

[Mirochill]

Fixes #258

Summary:

• include max_output_tokens in standard and long-form prompt parameters
• add the configured output-token budget to the composed transcript-generation prompt
• format the long-form instruction with the configured token count
• add focused tests for prompt parameter composition and the composed prompt variables

Validation:

• git diff --check
• git diff --check HEAD~1..HEAD
• git show --check --oneline HEAD -- podcastfy/content_generator.py tests/test_genai_podcast.py
• Static inspection of podcastfy/content_generator.py and tests/test_genai_podcast.py
• Not run locally