feat(report): enhance GitHub summary with baseline comparison

Improve GitHub Actions workflow summary output with dedicated baseline
comparison table and human-readable metric formatting.

Changes:
- Add baseline_meta field to Report for storing baseline run metadata
- Generate comparison header showing current vs baseline commits
- Add baseline comparison table with performance deltas
- Format verification expressions with proper unit scaling
- Support numeric literals in expressions (integers, decimals, scientific)
- Add comprehensive test coverage for formatting functions

The verification table now shows resolved metric values (e.g., "rps = 3.38M req/s")
instead of raw numbers, making reports more readable for humans.

Author: farhan-syah

fluxbench-cli/src/executor/report.rs

@@ -166,5 +166,6 @@ pub fn build_report(
         synthetics: Vec::new(),        // Filled by execute_verifications
         verifications: Vec::new(),     // Filled by execute_verifications
         summary,
+        baseline_meta: None,
     }
 }

fluxbench-cli/src/lib.rs

@@ -601,6 +601,9 @@ fn compare_benchmarks(
     let total_duration_ms = start_time.elapsed().as_secs_f64() * 1000.0;
     let mut report = build_report(&results, &stats, &exec_config, total_duration_ms);
 
+    // Store baseline metadata for summary header
+    report.baseline_meta = Some(baseline.meta.clone());
+
     // Add comparison data
     let regression_threshold = cli.threshold.unwrap_or(config.ci.regression_threshold);
     let baseline_map: std::collections::HashMap<_, _> = baseline