Fix: Remove hardcoded temperature overrides to use provider defaults

[DanielRyanSmith]

Resolves #437

Overview

Removes the hardcoded 0.01 temperature overrides from the Requirements Extraction and Coverage Audit phases, allowing them to fall back to the provider's default model temperature.

Root Cause / Motivation

Previously, the requirements_extraction and coverage_audit phases explicitly hardcoded their temperature to 0.01 when making LLM requests. This prevented the LLM's default temperature from being used and ignored the tool's intended fallback mechanism where no override allows the default to be respected.

Detailed Changelog

• wptgen/phases/requirements_extraction.py: Removed the temperature keyword argument when calling generate_safe, allowing the API client to use the provider's default temperature.
• wptgen/phases/coverage_audit.py: Removed the temperature keyword argument when calling generate_safe, allowing the API client to use the provider's default temperature.

[DanielRyanSmith]

This was an even smaller change than I had initially anticipated 😅

See the issue for rationale, but mainly this line from the Gemini 3 documentation:

"For all Gemini 3 models, we strongly recommend keeping the temperature parameter at its default value of 1.0.

"While previous models often benefited from tuning temperature to control creativity versus determinism, Gemini 3's reasoning capabilities are optimized for the default setting. Changing the temperature (setting it below 1.0) may lead to unexpected behavior, such as looping or degraded performance, particularly in complex mathematical or reasoning tasks."