Update docx, pptx, and xlsx skills (#1447)

Add support for template formats (.dotx, .potx, .xltx) across validation
and the helper scripts.

Consolidate the shared office helpers into a single module, replacing the
pack/unpack pipeline with explicit zip/unzip steps. Extraction now rejects
symlink and path-traversal archive entries. Move run merging to a
standalone docx script, since it only ever applied to Word documents.

Fix the redlining validator so it compares against the original even when
a document has no tracked changes, which is when an untracked edit would
otherwise go unreported.

Provision a LibreOffice user profile per invocation so conversions work in
sandboxed environments.

Trim the skill docs to the guidance that earns its place.

skills/pptx/scripts/thumbnail.py

@@ -16,6 +16,7 @@ Examples:
 """
 
 import argparse
+import posixpath
 import subprocess
 import sys
 import tempfile
@@ -23,9 +24,12 @@ import zipfile
 from pathlib import Path
 
 import defusedxml.minidom
-from office.soffice import get_soffice_env
+from defusedxml import ElementTree
+from office.helpers import SLIDE_REL_TYPE, opc_target
+from office.soffice import run_soffice
 from PIL import Image, ImageDraw, ImageFont
 
+
 THUMBNAIL_WIDTH = 300
 CONVERSION_DPI = 100
 MAX_COLS = 6
@@ -92,28 +96,45 @@ def main():
         sys.exit(1)
 
 
+def _is_hidden(zf: zipfile.ZipFile, part: str) -> bool:
+    try:
+        with zf.open(part) as f:
+            for _, root in ElementTree.iterparse(f, events=("start",)):
+                return root.get("show") in ("0", "false")  
+    except (KeyError, ElementTree.ParseError):
+        return False
+    return False
+
+
 def get_slide_info(pptx_path: Path) -> list[dict]:
     with zipfile.ZipFile(pptx_path, "r") as zf:
         rels_content = zf.read("ppt/_rels/presentation.xml.rels").decode("utf-8")
         rels_dom = defusedxml.minidom.parseString(rels_content)
 
-        rid_to_slide = {}
+        rid_to_part = {}
         for rel in rels_dom.getElementsByTagName("Relationship"):
-            rid = rel.getAttribute("Id")
-            target = rel.getAttribute("Target")
-            rel_type = rel.getAttribute("Type")
-            if "slide" in rel_type and target.startswith("slides/"):
-                rid_to_slide[rid] = target.replace("slides/", "")
+            if rel.getAttribute("Type") != SLIDE_REL_TYPE:
+                continue
+            part = opc_target(
+                rel.getAttribute("Target"),
+                "ppt/presentation.xml",
+                rel.getAttribute("TargetMode"),
+            )
+            if part is not None:
+                rid_to_part[rel.getAttribute("Id")] = part
 
         pres_content = zf.read("ppt/presentation.xml").decode("utf-8")
         pres_dom = defusedxml.minidom.parseString(pres_content)
 
+        present = set(zf.namelist())
+
         slides = []
         for sld_id in pres_dom.getElementsByTagName("p:sldId"):
-            rid = sld_id.getAttribute("r:id")
-            if rid in rid_to_slide:
-                hidden = sld_id.getAttribute("show") == "0"
-                slides.append({"name": rid_to_slide[rid], "hidden": hidden})
+            part = rid_to_part.get(sld_id.getAttribute("r:id"))
+            if part is not None and part in present:
+                slides.append(
+                    {"name": posixpath.basename(part), "hidden": _is_hidden(zf, part)}
+                )
 
         return slides
 
@@ -123,6 +144,15 @@ def build_slide_list(
     visible_images: list[Path],
     temp_dir: Path,
 ) -> list[tuple[Path, str]]:
+    visible_count = sum(1 for info in slide_info if not info["hidden"])
+    rendered_hidden = len(visible_images) == len(slide_info) != visible_count
+
+    if not rendered_hidden and visible_count != len(visible_images):
+        raise ValueError(
+            f"LibreOffice rendered {len(visible_images)} page(s) for {visible_count} "
+            f"visible slide(s) of {len(slide_info)}; thumbnails would be mislabeled"
+        )
+
     if visible_images:
         with Image.open(visible_images[0]) as img:
             placeholder_size = img.size
@@ -133,15 +163,15 @@ def build_slide_list(
     visible_idx = 0
 
     for info in slide_info:
-        if info["hidden"]:
+        if info["hidden"] and not rendered_hidden:
             placeholder_path = temp_dir / f"hidden-{info['name']}.jpg"
             placeholder_img = create_hidden_placeholder(placeholder_size)
             placeholder_img.save(placeholder_path, "JPEG")
             slides.append((placeholder_path, f"{info['name']} (hidden)"))
         else:
-            if visible_idx < len(visible_images):
-                slides.append((visible_images[visible_idx], info["name"]))
-                visible_idx += 1
+            label = f"{info['name']} (hidden)" if info["hidden"] else info["name"]
+            slides.append((visible_images[visible_idx], label))
+            visible_idx += 1
 
     return slides
 
@@ -158,22 +188,14 @@ def create_hidden_placeholder(size: tuple[int, int]) -> Image.Image:
 def convert_to_images(pptx_path: Path, temp_dir: Path) -> list[Path]:
     pdf_path = temp_dir / f"{pptx_path.stem}.pdf"
 
-    result = subprocess.run(
-        [
-            "soffice",
-            "--headless",
-            "--convert-to",
-            "pdf",
-            "--outdir",
-            str(temp_dir),
-            str(pptx_path),
-        ],
+    result = run_soffice(
+        ["--headless", "--convert-to", "pdf", "--outdir", str(temp_dir), str(pptx_path)],
         capture_output=True,
         text=True,
-        env=get_soffice_env(),
     )
     if result.returncode != 0 or not pdf_path.exists():
-        raise RuntimeError("PDF conversion failed")
+        detail = (result.stderr or result.stdout or "").strip()
+        raise RuntimeError(f"PDF conversion failed: {detail}" if detail else "PDF conversion failed")
 
     result = subprocess.run(
         [