fix: Correct coverage calculation to return 0% for untested issues

Previously, coverage analysis was incorrectly using keywords from all
existing tests, causing false positives where untested issues showed
coverage percentages instead of 0%.

Changes:
- Only count tests specifically related to the analyzed issue
- Return 0% coverage when no issue-specific tests exist
- Maintain accurate coverage calculation for tested issues

This ensures that Issue #3 correctly shows 0.0% coverage instead of
33.3%, while Issue #11 still correctly shows 100.0% coverage.

🤖 Generated with [Claude Code](https://claude.ai/code)

Co-Authored-By: Claude <noreply@anthropic.com>

tddai/coverage_analyzer.py

@@ -353,8 +353,15 @@ class CoverageAnalyzer:
 
         # Get all covered keywords
         covered_keywords = set()
+        issue_related_tests = []
         for test in existing_tests:
-            covered_keywords.update(test.coverage_keywords)
+            if test.related_issue:  # Only count tests specifically for this issue
+                covered_keywords.update(test.coverage_keywords)
+                issue_related_tests.append(test)
+
+        # If no issue-specific tests found, coverage should be 0%
+        if not issue_related_tests:
+            return 0.0
 
         # Check coverage for each requirement
         for requirement in requirements: