llm: add model temperature mapping and temperature logging

[dwojtaszek]

This PR introduces a simple model->temperature mapping and logs the effective temperature used when calling the GitHub Models API.

Changes:

• Add modelTemperatureMap and getMappedTemperature to internal/llm/client.go.
• Use mapping (when present) to determine the temperature; otherwise use promptConfig.ModelParameters.Temperature.
• Rename local variable to effectiveTemperature for clarity.
• Add a console log that prints selected model and effective temperature before the API call.

Verification:

• go build ./... passes locally.
• Ran the CLI and observed the new log line (e.g. 'Selected model=openai/gpt-5-mini, temperature=1.00').

Notes:

• No behavior changes other than temperature selection precedence and added logging.
• Follow-ups: add unit tests, prefix-based model matching, or a verbose flag for the log.,