fix:issue-discovery-fixed-base-cache-miss

[tao-hunter]

Summary

Apply RepositoryConfig.fixed_base_score consistently in mirror issue-discovery solving PR scoring, including the cache-miss path.

• Thread repo_config into _resolve_solving_pr_score in mirror_scan.
• On cache miss, compute token fields as before, but override only base_score when repo_config.fixed_base_score is set.
• Keep token-derived fields unchanged to preserve eligibility/pioneer/reporting behavior.
• Add regression coverage to ensure fixed-base behavior is identical across cache hit and cache miss for the same repo config.

Related Issues

Fixes #1172

Type of Change

• Bug fix
• New feature
• Refactor
• Documentation
• Other

Testing

• Updated unit tests in tests/validator/issue_discovery/test_mirror_scan.py.
• Added test coverage for:
  • fixed-base override applied on cache miss
  • subsequent cache hit returns the same fixed base
  • token score remains token-derived in both paths
• Local run in this environment:
  • pytest -q tests/validator/issue_discovery/test_mirror_scan.py (module skipped due to optional import guards in this environment)

Checklist

• Code follows project style/conventions
• Self-reviewed changes
• Tests added/updated for changed behavior
• No unintended behavior changes to token-derived gates/signals
• Documentation/comments updated where helpful