sweep 01-06 wrap-up: READING-LIST regen (583 files), generator xref-tally fix, INDEX fixes

- make_rootlist.py: xref items no longer counted into section status tallies
  (they carry the canonical card's marker in parens, which was inflating counts:
  sec11 6/1/0, sec10 5/0/27 now correct)
- sec10 INDEX: #23 Jamison, #33 Saks ❌->✅ (stale after user-oracle rescues)
- sec03 MISSING: #65 stale entry counted out (already corrected 2026-09-24)
- AGENTS.md: section file counts + sec03 40/0/28 + total 583 files
- READING-LIST.md regenerated: 583 files, 768 links

tools/make_rootlist.py

@@ -39,10 +39,11 @@ def parse_index(sec):
         m = re.match(r'^\|\s*(\d{2})\s*\|\s*([^|]*)\|\s*([^|]*)\s*\|\s*([^|]*)\s*\|\s*([^|]+?)\s*\|\s*([^|]+)\|', line)
         if m:
             num, grp, title, author, status, card = m.groups()
+            is_xref = status.strip().startswith("xref")
             em = re.search(r'[✅🔶❌⬜]', status)
             items[num] = dict(title=title.strip(), author=author.strip(),
-                              status=(em.group(0) if em else ("xref" if status.strip().startswith("xref") else status.strip())),
-                              card=card.strip())
+                              status=(em.group(0) if em else status.strip()),
+                              is_xref=is_xref, card=card.strip())
     return name, items
 
 def parse_manifest(sec):
@@ -208,6 +209,8 @@ def main():
         dfiles_n = len([f for f in os.listdir(os.path.join(BASE, "downloads", ddir)) if f != "MANIFEST.md"])
         stat = {e: 0 for e in "✅🔶❌"}
         for it in items.values():
+            if it.get("is_xref"):
+                continue  # xref items are counted in their canonical section
             stat[it["status"]] = stat.get(it["status"], 0) + 1
         out.append(f"## {n}. {name}")
         out.append("")