Merge PR #69: run full routing benchmark through Bash router

.github/workflows/ci.yml

@@ -71,17 +71,10 @@ jobs:
         shell: bash
         run: |
           set -euo pipefail
+          bash skills/scripts/test-routing.sh
+
           scratch="$(mktemp -d)"
           trap 'rm -rf "$scratch"' EXIT
-          while IFS='|' read -r hint expected; do
-            out="$scratch/$expected"
-            bash skills/scripts/master-route.sh --hint "$hint" --out-dir "$out"
-            grep -Eq "^- primary: ${expected}$" "$out/route-scope.md"
-          done <<'CASES'
-          apk reverse with jadx|R1
-          js reverse signature|R3
-          case review evidence graph|R40
-          CASES
 
           if bash skills/scripts/case-init.sh \
               --hint "offline apk" \

skills/scripts/test-routing.sh

@@ -0,0 +1,76 @@
+#!/usr/bin/env bash
+# Run the shared routing benchmark through the Bash structured router.
+set -euo pipefail
+
+SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+BENCHMARK="$SCRIPT_DIR/../tests/routing-benchmark.json"
+ROUTER="$SCRIPT_DIR/master-route.sh"
+
+PYTHON=""
+for candidate in python3 python; do
+  if command -v "$candidate" >/dev/null 2>&1; then
+    PYTHON="$candidate"
+    break
+  fi
+done
+if [[ -z "$PYTHON" ]]; then
+  echo 'ERROR: Python 3 is required by the Bash routing benchmark runner.' >&2
+  exit 2
+fi
+
+"$PYTHON" - "$BENCHMARK" "$ROUTER" <<'PY'
+import json
+import pathlib
+import re
+import subprocess
+import sys
+import tempfile
+
+benchmark_path = pathlib.Path(sys.argv[1])
+router_path = pathlib.Path(sys.argv[2])
+benchmark = json.loads(benchmark_path.read_text(encoding="utf-8-sig"))
+cases = benchmark["cases"]
+
+failures = []
+passed = 0
+print(f"=== Bash routing benchmark | {len(cases)} cases ===")
+
+with tempfile.TemporaryDirectory(prefix="rs-routing-bash-") as scratch:
+    root = pathlib.Path(scratch)
+    for index, case in enumerate(cases):
+        hint = case["hint"]
+        expected = case["expect"]
+        out_dir = root / str(index)
+        result = subprocess.run(
+            ["bash", str(router_path), "--hint", hint, "--out-dir", str(out_dir)],
+            capture_output=True,
+            text=True,
+            encoding="utf-8",
+            errors="replace",
+        )
+
+        got = "ERR"
+        scope_path = out_dir / "route-scope.md"
+        if result.returncode == 0 and scope_path.is_file():
+            match = re.search(
+                r"(?m)^- primary:[ \t]*(\S+)[ \t]*$",
+                scope_path.read_text(encoding="utf-8"),
+            )
+            if match:
+                got = match.group(1)
+
+        if got == expected:
+            passed += 1
+        else:
+            failures.append(
+                f"hint={hint!r} expect={expected} got={got} exit={result.returncode}"
+            )
+
+print(f"TOTAL={len(cases)} PASS={passed} FAIL={len(failures)}")
+if failures:
+    for failure in failures:
+        print(f"[FAIL] {failure}", file=sys.stderr)
+    raise SystemExit(1)
+
+print(f"OVERALL: ALL PASS ({passed})")
+PY