feat(audit): wire cross_audit_findings aggregator into synthesize

Loops over audit_weak_types + audit_exception_handling from
the 6 audit_inputs, calls aggregate_cross_audit_findings per
audit, sums the buckets per profile.

Cross-audit aggregation is per-aggregate-flat (all findings go
into 1 bucket per audit). The 3-tier finding-to-aggregate
mapping (find_enclosing_function + type registry + file
heuristic) is the next gap - requires per-finding site
classification.

src/code_path_audit.py

@@ -1082,6 +1082,25 @@ def synthesize_aggregate_profile(
   consumers, aggregate, type_registry, "src"
  )
  tac = compute_real_type_alias_coverage(aggregate, producers, consumers, type_registry, "src")
+ cross_findings = CrossAuditFindings((), (), (), (), ())
+ for audit_name in ("audit_weak_types", "audit_exception_handling"):
+  if audit_name in audit_inputs:
+   findings = audit_inputs[audit_name].get("findings", [])
+   example_file = findings[0].get("file", "") if findings else ""
+   example_line = findings[0].get("line", 0) if findings else 0
+   matched = aggregate_cross_audit_findings(
+    audit_name=audit_name,
+    findings=findings,
+    example_file=example_file,
+    example_line=example_line,
+   )
+   cross_findings = CrossAuditFindings(
+    weak_types=cross_findings.weak_types + matched.weak_types,
+    exception_handling=cross_findings.exception_handling + matched.exception_handling,
+    optional_in_baseline=cross_findings.optional_in_baseline + matched.optional_in_baseline,
+    config_io_ownership=cross_findings.config_io_ownership + matched.config_io_ownership,
+    import_graph=cross_findings.import_graph + matched.import_graph,
+   )
  producer_count = len({f.fqname for f in producers})
  consumer_count = len({f.fqname for f in consumers})
  branches_on_errors = set()
@@ -1112,7 +1131,7 @@ def synthesize_aggregate_profile(
   frequency_evidence=freq_evidence,
   result_coverage=rc,
   type_alias_coverage=tac,
-  cross_audit_findings=CrossAuditFindings((), (), (), (), ()),
+  cross_audit_findings=cross_findings,
   decomposition_cost=dc,
   optimization_candidates=candidates,
   is_candidate=False,