fix: use ReportConfig.max_examples instead of hardcoded limits (#114)

openagent_eval/reports/markdown.py

@@ -102,15 +102,15 @@ def generate(self, report: Any) -> str:
             # Error details
             sections.append("### Error Details")
             sections.append("")
-            for i, err in enumerate(result.errors[:10], 1):
+            for i, err in enumerate(result.errors[:config.report.max_examples], 1):
                 sections.append(f"{i}. **{err.get('error_type', 'Unknown')}**: {err.get('error', 'No message')}")
             sections.append("")
 
         # Example Results
         if result.results:
             sections.append("## Sample Results")
             sections.append("")
-            max_examples = min(len(result.results), 10)
+            max_examples = min(len(result.results), config.report.max_examples)
             for i, eval_result in enumerate(result.results[:max_examples], 1):
                 sections.append(f"### Result {i}")
                 sections.append("")

openagent_eval/reports/terminal.py

@@ -181,7 +181,7 @@ def print_report(self, report: Any) -> None:
             example_table.add_column("#", style="dim")
             example_table.add_column("Question", max_width=40)
             example_table.add_column("Metrics", max_width=30)
-            for i, eval_result in enumerate(result.results[:5], 1):
+            for i, eval_result in enumerate(result.results[:config.report.max_examples], 1):
                 metrics_str = ", ".join(
                     f"{k}={v:.2f}" for k, v in eval_result.metrics.items()
                 )