From 1c1bf2234611628ab5b132c273adba4d4c201e50 Mon Sep 17 00:00:00 2001
From: santhreal <64453045+santhreal@users.noreply.github.com>
Date: Thu, 23 Jul 2026 22:35:44 -0700
Subject: [PATCH 1/5] fix(visualize): return empty on extract_code_block
 language miss

A mismatched language hint returned the full raw text, so codegen's
`hint or any-fence` fallback never ran and ```json Chart.js fences with
trailing prose failed local validation.
---
 deeptutor/agents/visualize/utils.py               | 4 ++++
 tests/agents/visualize/test_extract_code_block.py | 8 ++++++++
 2 files changed, 12 insertions(+)

diff --git a/deeptutor/agents/visualize/utils.py b/deeptutor/agents/visualize/utils.py
index 0089b41f28..3633b5e4ce 100644
--- a/deeptutor/agents/visualize/utils.py
+++ b/deeptutor/agents/visualize/utils.py
@@ -37,6 +37,7 @@ def extract_code_block(text: str, language: str = "") -> str:
 
     If *language* is given the block must start with that tag;
     otherwise any triple-backtick fence is accepted.
+    A language miss returns ``""`` so callers can fall back with ``or``.
     """
     # Closing fence may sit on the same line as the last content line.
     if language:
@@ -46,6 +47,9 @@ def extract_code_block(text: str, language: str = "") -> str:
     match = re.search(pattern, text or "", re.IGNORECASE)
     if match:
         return match.group(1).strip()
+    # Language miss must be falsy so `hint or any-fence` can fall through.
+    if language:
+        return ""
     return (text or "").strip()
 
 
diff --git a/tests/agents/visualize/test_extract_code_block.py b/tests/agents/visualize/test_extract_code_block.py
index 79d13244fb..c979a243e9 100644
--- a/tests/agents/visualize/test_extract_code_block.py
+++ b/tests/agents/visualize/test_extract_code_block.py
@@ -14,3 +14,11 @@ def test_extract_code_block_closing_fence_without_leading_newline() -> None:
 def test_extract_code_block_normal_fenced_block() -> None:
     raw = "```javascript\nconst x = 1;\n```"
     assert extract_code_block(raw, "javascript") == "const x = 1;"
+
+
+def test_extract_code_block_language_miss_is_empty_for_or_fallback() -> None:
+    cfg = '{"type": "bar", "data": {"labels": ["A"], "datasets": [{"data": [1]}]}}'
+    raw = f"```json\n{cfg}\n```\nThanks."
+    assert extract_code_block(raw, "javascript") == ""
+    extracted = extract_code_block(raw, "javascript") or extract_code_block(raw)
+    assert extracted == cfg

From 6e440690a58b301cc53440ad9347335bb380553a Mon Sep 17 00:00:00 2001
From: santhreal <64453045+santhreal@users.noreply.github.com>
Date: Fri, 24 Jul 2026 23:43:44 -0700
Subject: [PATCH 2/5] chore: ruff-format builtin skill markdown

---
 deeptutor/skills/builtin/docx/SKILL.md | 54 +++++++++-----
 deeptutor/skills/builtin/pdf/SKILL.md  | 98 ++++++++++++++++++--------
 deeptutor/skills/builtin/pptx/SKILL.md | 36 ++++++----
 deeptutor/skills/builtin/xlsx/SKILL.md | 49 ++++++++-----
 4 files changed, 159 insertions(+), 78 deletions(-)

diff --git a/deeptutor/skills/builtin/docx/SKILL.md b/deeptutor/skills/builtin/docx/SKILL.md
index 0608b794c1..ba0402d42d 100644
--- a/deeptutor/skills/builtin/docx/SKILL.md
+++ b/deeptutor/skills/builtin/docx/SKILL.md
@@ -26,9 +26,10 @@ After `exec` completes, use the Generated artifacts URL from the tool result in
 
 ```python
 from docx import Document
+
 doc = Document("in.docx")
-text = "\n".join(p.text for p in doc.paragraphs)      # body paragraphs
-for tbl in doc.tables:                                 # tables
+text = "\n".join(p.text for p in doc.paragraphs)  # body paragraphs
+for tbl in doc.tables:  # tables
     for row in tbl.rows:
         print([c.text for c in row.cells])
 ```
@@ -39,6 +40,7 @@ To read **tracked changes**, parse the XML directly — `python-docx` ignores `<
 
 ```python
 import zipfile, re
+
 xml = zipfile.ZipFile("in.docx").read("word/document.xml").decode("utf-8")
 # inserted text = <w:ins>…<w:t>…  deleted = <w:del>…<w:delText>…
 print(re.findall(r"<w:t[^>]*>(.*?)</w:t>", xml))
@@ -51,20 +53,23 @@ from docx import Document
 from docx.shared import Pt, Inches, RGBColor
 from docx.enum.text import WD_ALIGN_PARAGRAPH
 
-doc = Document()                                  # default template page size
-doc.add_heading("Quarterly Report", level=0)      # 0 = title; 1..9 = H1..H9
+doc = Document()  # default template page size
+doc.add_heading("Quarterly Report", level=0)  # 0 = title; 1..9 = H1..H9
 p = doc.add_paragraph("Intro paragraph. ")
-run = p.add_run("Bold tail."); run.bold = True
-doc.add_paragraph("First item", style="List Bullet")   # real list style, never a "• " literal
+run = p.add_run("Bold tail.")
+run.bold = True
+doc.add_paragraph("First item", style="List Bullet")  # real list style, never a "• " literal
 doc.add_paragraph("Step one", style="List Number")
 
 # Table — header row + data
-tbl = doc.add_table(rows=1, cols=2); tbl.style = "Light Grid Accent 1"
+tbl = doc.add_table(rows=1, cols=2)
+tbl.style = "Light Grid Accent 1"
 tbl.rows[0].cells[0].text, tbl.rows[0].cells[1].text = "Metric", "Value"
 for k, v in [("Revenue", "1.2M"), ("Growth", "15%")]:
-    c = tbl.add_row().cells; c[0].text, c[1].text = k, v
+    c = tbl.add_row().cells
+    c[0].text, c[1].text = k, v
 
-doc.add_picture("chart.png", width=Inches(5))     # image, scaled to width
+doc.add_picture("chart.png", width=Inches(5))  # image, scaled to width
 doc.add_page_break()
 doc.save("out.docx")
 ```
@@ -79,8 +84,9 @@ Rules:
 
 ```python
 from docx.shared import Inches
+
 sec = doc.sections[0]
-sec.page_width, sec.page_height = Inches(8.5), Inches(11)   # Letter
+sec.page_width, sec.page_height = Inches(8.5), Inches(11)  # Letter
 sec.top_margin = sec.bottom_margin = Inches(1)
 sec.header.paragraphs[0].text = "Confidential"
 ```
@@ -90,14 +96,22 @@ Page-number fields aren't in the python-docx API; inject the field XML into a fo
 ```python
 from docx.oxml.ns import qn
 from docx.oxml import OxmlElement
+
+
 def add_page_number(paragraph):
     for t in ("begin", "instr", "end"):
         r = OxmlElement("w:r")
         if t == "instr":
-            fld = OxmlElement("w:instrText"); fld.set(qn("xml:space"), "preserve"); fld.text = "PAGE"
+            fld = OxmlElement("w:instrText")
+            fld.set(qn("xml:space"), "preserve")
+            fld.text = "PAGE"
         else:
-            fld = OxmlElement("w:fldChar"); fld.set(qn("w:fldCharType"), t)
-        r.append(fld); paragraph._p.append(r)
+            fld = OxmlElement("w:fldChar")
+            fld.set(qn("w:fldCharType"), t)
+        r.append(fld)
+        paragraph._p.append(r)
+
+
 add_page_number(doc.sections[0].footer.paragraphs[0])
 ```
 
@@ -136,12 +150,16 @@ Only when python-docx can't express it: **tracked changes, comments, exact-fidel
 
 ```python
 import zipfile
+
 src, dst = "in.docx", "out.docx"
-with zipfile.ZipFile(src) as z: xml = z.read("word/document.xml").decode("utf-8")
-xml = xml.replace("OLD", "NEW")          # or splice tracked-change elements (below)
+with zipfile.ZipFile(src) as z:
+    xml = z.read("word/document.xml").decode("utf-8")
+xml = xml.replace("OLD", "NEW")  # or splice tracked-change elements (below)
 with zipfile.ZipFile(src) as zin, zipfile.ZipFile(dst, "w", zipfile.ZIP_DEFLATED) as zout:
     for item in zin.infolist():
-        data = xml.encode("utf-8") if item.filename == "word/document.xml" else zin.read(item.filename)
+        data = (
+            xml.encode("utf-8") if item.filename == "word/document.xml" else zin.read(item.filename)
+        )
         zout.writestr(item, data)
 ```
 
@@ -180,7 +198,9 @@ Always confirm the file reopens cleanly — a silent corruption is the most comm
 
 ```python
 from docx import Document
-d = Document("out.docx"); print(len(d.paragraphs), "paragraphs OK")
+
+d = Document("out.docx")
+print(len(d.paragraphs), "paragraphs OK")
 ```
 
 For raw-OOXML edits also run `python -c "import zipfile; zipfile.ZipFile('out.docx').testzip()"` and well-formedness-check each edited XML part with `lxml.etree.parse`.
diff --git a/deeptutor/skills/builtin/pdf/SKILL.md b/deeptutor/skills/builtin/pdf/SKILL.md
index 777e4ac837..3d0ee7b500 100644
--- a/deeptutor/skills/builtin/pdf/SKILL.md
+++ b/deeptutor/skills/builtin/pdf/SKILL.md
@@ -26,11 +26,12 @@ After `exec` completes, use the Generated artifacts URL from the tool result in
 
 ```python
 import pdfplumber
+
 with pdfplumber.open("in.pdf") as pdf:
     for i, page in enumerate(pdf.pages, 1):
         print(f"--- page {i} ---")
-        print(page.extract_text() or "")        # layout-aware text
-        for t in page.extract_tables():          # list of tables; each is list[row]
+        print(page.extract_text() or "")  # layout-aware text
+        for t in page.extract_tables():  # list of tables; each is list[row]
             for row in t:
                 print(row)
 ```
@@ -38,6 +39,7 @@ with pdfplumber.open("in.pdf") as pdf:
 Tables → DataFrame/Excel:
 ```python
 import pdfplumber, pandas as pd
+
 frames = []
 with pdfplumber.open("in.pdf") as pdf:
     for page in pdf.pages:
@@ -50,8 +52,12 @@ if frames:
 
 Messy tables: pass strategies, or crop a region with `page.within_bbox((x0, top, x1, bottom))` first:
 ```python
-ts = {"vertical_strategy": "lines", "horizontal_strategy": "lines",
-      "snap_tolerance": 3, "intersection_tolerance": 15}
+ts = {
+    "vertical_strategy": "lines",
+    "horizontal_strategy": "lines",
+    "snap_tolerance": 3,
+    "intersection_tolerance": 15,
+}
 page.extract_tables(ts)
 ```
 
@@ -76,11 +82,16 @@ w.write("merged.pdf")
 # Split: one file per page
 r = PdfReader("in.pdf")
 for i, p in enumerate(r.pages, 1):
-    w = PdfWriter(); w.add_page(p); w.write(f"page_{i}.pdf")
+    w = PdfWriter()
+    w.add_page(p)
+    w.write(f"page_{i}.pdf")
 
 # Rotate page 0 by 90 degrees clockwise
-r = PdfReader("in.pdf"); w = PdfWriter()
-r.pages[0].rotate(90); w.add_page(r.pages[0]); w.write("rotated.pdf")
+r = PdfReader("in.pdf")
+w = PdfWriter()
+r.pages[0].rotate(90)
+w.add_page(r.pages[0])
+w.write("rotated.pdf")
 ```
 
 - **Metadata**: `PdfReader("in.pdf").metadata` (`.title`, `.author`, ...).
@@ -91,10 +102,13 @@ r.pages[0].rotate(90); w.add_page(r.pages[0]); w.write("rotated.pdf")
 Watermark (stamp one page over every page):
 ```python
 from pypdf import PdfReader, PdfWriter
+
 wm = PdfReader("stamp.pdf").pages[0]
-r = PdfReader("in.pdf"); w = PdfWriter()
+r = PdfReader("in.pdf")
+w = PdfWriter()
 for p in r.pages:
-    p.merge_page(wm); w.add_page(p)
+    p.merge_page(wm)
+    w.add_page(p)
 w.write("stamped.pdf")
 ```
 
@@ -108,15 +122,20 @@ from reportlab.lib import colors
 from reportlab.platypus import SimpleDocTemplate, Paragraph, Spacer, Table, TableStyle
 
 styles = getSampleStyleSheet()
-story = [Paragraph("Report Title", styles["Title"]), Spacer(1, 12),
-         Paragraph("Body text. " * 20, styles["Normal"])]
+story = [
+    Paragraph("Report Title", styles["Title"]),
+    Spacer(1, 12),
+    Paragraph("Body text. " * 20, styles["Normal"]),
+]
 data = [["Product", "Q1", "Q2"], ["Widgets", "120", "135"]]
 tbl = Table(data)
-tbl.setStyle(TableStyle([
-    ("BACKGROUND", (0, 0), (-1, 0), colors.grey),
-    ("TEXTCOLOR", (0, 0), (-1, 0), colors.whitesmoke),
-    ("GRID", (0, 0), (-1, -1), 0.5, colors.black),
-]))
+tbl.setStyle(
+    TableStyle([
+        ("BACKGROUND", (0, 0), (-1, 0), colors.grey),
+        ("TEXTCOLOR", (0, 0), (-1, 0), colors.whitesmoke),
+        ("GRID", (0, 0), (-1, -1), 0.5, colors.black),
+    ])
+)
 story += [Spacer(1, 12), tbl]
 SimpleDocTemplate("out.pdf", pagesize=letter).build(story)
 ```
@@ -132,17 +151,18 @@ import os
 from reportlab.pdfbase import pdfmetrics
 from reportlab.pdfbase.ttfonts import TTFont
 
+
 def register_cjk_font(name="CJK"):
     # TrueType ONLY — reportlab cannot embed CFF/OpenType outlines, so a .otf
     # like Noto Sans CJK fails with "postscript outlines are not supported".
     for path in [
-        "/usr/share/fonts/truetype/wqy/wqy-zenhei.ttc",        # Linux sandbox (fonts-wqy-zenhei)
+        "/usr/share/fonts/truetype/wqy/wqy-zenhei.ttc",  # Linux sandbox (fonts-wqy-zenhei)
         "/usr/share/fonts/truetype/wqy/wqy-microhei.ttc",
-        "/System/Library/Fonts/STHeiti Light.ttc",             # macOS
+        "/System/Library/Fonts/STHeiti Light.ttc",  # macOS
         "/System/Library/Fonts/Hiragino Sans GB.ttc",
         "/System/Library/Fonts/Supplemental/Songti.ttc",
         "/System/Library/Fonts/Supplemental/Arial Unicode.ttf",
-        "C:/Windows/Fonts/msyh.ttc",                           # Windows
+        "C:/Windows/Fonts/msyh.ttc",  # Windows
     ]:
         if os.path.exists(path):
             try:
@@ -152,9 +172,10 @@ def register_cjk_font(name="CJK"):
                 continue
     raise RuntimeError("No CJK-capable TrueType font found — do not emit tofu; say so.")
 
+
 font = register_cjk_font()
 styles = getSampleStyleSheet()
-for s in styles.byName.values():                 # make the CJK font the default everywhere
+for s in styles.byName.values():  # make the CJK font the default everywhere
     s.fontName = font
 # Tables don't read the stylesheet — set the font in the TableStyle too:
 #   ("FONTNAME", (0, 0), (-1, -1), font)
@@ -172,6 +193,7 @@ Markdown/HTML → PDF needs an external converter (`soffice`/`pandoc`) that is u
 First detect whether the PDF has real fillable (AcroForm) fields:
 ```python
 from pypdf import PdfReader
+
 fields = PdfReader("form.pdf").get_fields()
 print("fillable" if fields else "flat (no fields)")
 ```
@@ -179,15 +201,16 @@ print("fillable" if fields else "flat (no fields)")
 **Fillable** — inspect field names/types, then fill and write:
 ```python
 from pypdf import PdfReader, PdfWriter
+
 r = PdfReader("form.pdf")
 for name, f in r.get_fields().items():
-    print(name, f.get("/FT"), f.get("/_States_"))   # /Tx text, /Btn checkbox/radio, /Ch choice
+    print(name, f.get("/FT"), f.get("/_States_"))  # /Tx text, /Btn checkbox/radio, /Ch choice
 
 w = PdfWriter(clone_from=r)
-values = {"first_name": "Bart", "agree": "/Yes"}     # checkbox/radio: use its on-state, NOT True/False
+values = {"first_name": "Bart", "agree": "/Yes"}  # checkbox/radio: use its on-state, NOT True/False
 for page in w.pages:
     w.update_page_form_field_values(page, values, auto_regenerate=False)
-w.set_need_appearances_writer(True)                  # force viewers to render the values
+w.set_need_appearances_writer(True)  # force viewers to render the values
 w.write("filled.pdf")
 ```
 Checkbox/radio values are on-state strings, not booleans — read the field's `/_States_` (e.g. `/Yes`, `/On`); `/Off` clears it.
@@ -195,24 +218,36 @@ Checkbox/radio values are on-state strings, not booleans — read the field's `/
 **Flat form (no fields)** — overlay text with `FreeText` annotations at PDF coordinates. Get real coordinates from the layout with pdfplumber instead of guessing:
 ```python
 import pdfplumber
+
 with pdfplumber.open("form.pdf") as pdf:
     pg = pdf.pages[0]
-    for wd in pg.extract_words():     # each has x0, top, x1, bottom (TOP-left origin!)
+    for wd in pg.extract_words():  # each has x0, top, x1, bottom (TOP-left origin!)
         print(wd["text"], wd["x0"], wd["top"])
-    for rc in pg.rects:               # small squares are likely checkboxes
+    for rc in pg.rects:  # small squares are likely checkboxes
         print("rect", rc["x0"], rc["top"], rc["x1"], rc["bottom"])
 ```
 pdfplumber `top` is measured from the page top; pypdf rects are bottom-left, so convert: `pdf_y = page_height - top`. Place text just right of the matching label:
 ```python
 from pypdf import PdfReader, PdfWriter
 from pypdf.annotations import FreeText
-r = PdfReader("form.pdf"); w = PdfWriter(); w.append(r)
+
+r = PdfReader("form.pdf")
+w = PdfWriter()
+w.append(r)
 h = float(r.pages[0].mediabox.height)
 top = 700  # pdfplumber 'top' of the label's row
-w.add_annotation(page_number=0, annotation=FreeText(
-    text="Smith", rect=(255, h - top - 14, 720, h - top),   # (x0, y0, x1, y1)
-    font="Helvetica", font_size="10pt", font_color="000000",
-    border_color=None, background_color=None))
+w.add_annotation(
+    page_number=0,
+    annotation=FreeText(
+        text="Smith",
+        rect=(255, h - top - 14, 720, h - top),  # (x0, y0, x1, y1)
+        font="Helvetica",
+        font_size="10pt",
+        font_color="000000",
+        border_color=None,
+        background_color=None,
+    ),
+)
 w.write("filled.pdf")
 ```
 Verify: re-open the output and re-read `get_fields()` values (fillable) or re-extract text (overlay) to confirm the values landed.
@@ -223,9 +258,10 @@ Verify: re-open the output and re-read `get_fields()` values (fillable) or re-ex
 
 ```python
 import fitz  # PyMuPDF
+
 doc = fitz.open("in.pdf")
 for i, page in enumerate(doc, 1):
-    page.get_pixmap(dpi=150).save(f"page_{i}.png")   # higher dpi = sharper + larger
+    page.get_pixmap(dpi=150).save(f"page_{i}.png")  # higher dpi = sharper + larger
 ```
 
 `fitz` also extracts text (`page.get_text()`) and can render a sub-region via `page.get_pixmap(clip=fitz.Rect(x0, y0, x1, y1))`. It does **not** OCR — a rendered scanned page is still just pixels (see Scanned PDFs above).
diff --git a/deeptutor/skills/builtin/pptx/SKILL.md b/deeptutor/skills/builtin/pptx/SKILL.md
index 25fd422d34..ab835910ab 100644
--- a/deeptutor/skills/builtin/pptx/SKILL.md
+++ b/deeptutor/skills/builtin/pptx/SKILL.md
@@ -34,6 +34,7 @@ the final answer so the user can download the deck.
 ## Read / extract
 ```python
 from pptx import Presentation
+
 prs = Presentation("deck.pptx")
 print(len(prs.slides), prs.slide_width, prs.slide_height)  # EMU dims
 
@@ -41,7 +42,7 @@ for i, slide in enumerate(prs.slides, 1):
     print(f"--- slide {i} (layout: {slide.slide_layout.name}) ---")
     for shape in slide.shapes:
         if shape.has_text_frame:
-            print(shape.text_frame.text)          # \n-joined paragraphs
+            print(shape.text_frame.text)  # \n-joined paragraphs
         elif shape.has_table:
             for row in shape.table.rows:
                 print([c.text for c in row.cells])
@@ -67,15 +68,17 @@ for idx, lay in enumerate(prs.slide_layouts):
 # Title slide
 s = prs.slides.add_slide(prs.slide_layouts[0])
 s.shapes.title.text = "My Deck"
-s.placeholders[1].text = "Subtitle"      # idx from the listing above
+s.placeholders[1].text = "Subtitle"  # idx from the listing above
 
 # Title + bullets
 s = prs.slides.add_slide(prs.slide_layouts[1])
 s.shapes.title.text = "Agenda"
 tf = s.placeholders[1].text_frame
-tf.text = "First point"                    # first paragraph
+tf.text = "First point"  # first paragraph
 for line, lvl in [("Second", 0), ("Sub-point", 1)]:
-    p = tf.add_paragraph(); p.text = line; p.level = lvl
+    p = tf.add_paragraph()
+    p.text = line
+    p.level = lvl
 
 prs.save("out.pptx")
 ```
@@ -85,7 +88,10 @@ indentation/bullets come from the layout via `paragraph.level`.
 Add a free text box or picture on any slide:
 ```python
 tb = s.shapes.add_textbox(Inches(1), Inches(1), Inches(8), Inches(1))
-r = tb.text_frame.paragraphs[0].add_run(); r.text = "Hi"; r.font.size = Pt(28); r.font.bold = True
+r = tb.text_frame.paragraphs[0].add_run()
+r.text = "Hi"
+r.font.size = Pt(28)
+r.font.bold = True
 s.shapes.add_picture("logo.png", Inches(0.5), Inches(0.5), height=Inches(1))  # omit w to keep ratio
 ```
 
@@ -95,7 +101,8 @@ Edit at the **run** level to preserve a run's formatting; rewriting
 ```python
 for slide in prs.slides:
     for shape in slide.shapes:
-        if not shape.has_text_frame: continue
+        if not shape.has_text_frame:
+            continue
         for para in shape.text_frame.paragraphs:
             for run in para.runs:
                 if "{{NAME}}" in run.text:
@@ -110,6 +117,7 @@ python-pptx has no direct setter; swap the bytes of the related image part. Read
 the picture's `r:embed` rId off its `<a:blip>`, then overwrite the part's blob.
 ```python
 from pptx.oxml.ns import qn
+
 for shape in slide.shapes:
     if shape.shape_type == 13:  # MSO_SHAPE_TYPE.PICTURE
         blip = shape._element.find(".//" + qn("a:blip"))
@@ -121,16 +129,19 @@ for shape in slide.shapes:
 ### Tables and charts
 ```python
 from pptx.util import Inches
-tbl = s.shapes.add_table(rows=2, cols=2, left=Inches(1), top=Inches(1),
-                         width=Inches(6), height=Inches(2)).table
-tbl.cell(0,0).text = "Header"
+
+tbl = s.shapes.add_table(
+    rows=2, cols=2, left=Inches(1), top=Inches(1), width=Inches(6), height=Inches(2)
+).table
+tbl.cell(0, 0).text = "Header"
 
 from pptx.chart.data import CategoryChartData
 from pptx.enum.chart import XL_CHART_TYPE
-cd = CategoryChartData(); cd.categories = ["Q1","Q2","Q3"]
+
+cd = CategoryChartData()
+cd.categories = ["Q1", "Q2", "Q3"]
 cd.add_series("Sales", (4.5, 5.5, 6.2))
-s.shapes.add_chart(XL_CHART_TYPE.COLUMN_CLUSTERED, Inches(1), Inches(1),
-                   Inches(8), Inches(4.5), cd)
+s.shapes.add_chart(XL_CHART_TYPE.COLUMN_CLUSTERED, Inches(1), Inches(1), Inches(8), Inches(4.5), cd)
 ```
 
 ## Design (only when the user wants a polished deck, not a data dump)
@@ -193,6 +204,7 @@ Minimal text edit by zip surgery (zip members can't be overwritten in place —
 rebuild the archive, swapping the one part):
 ```python
 import zipfile
+
 target = "ppt/slides/slide1.xml"
 with zipfile.ZipFile("in.pptx") as zin:
     xml = zin.read(target).decode().replace("Old title", "New title")
diff --git a/deeptutor/skills/builtin/xlsx/SKILL.md b/deeptutor/skills/builtin/xlsx/SKILL.md
index e9e284c257..0a37f02e46 100644
--- a/deeptutor/skills/builtin/xlsx/SKILL.md
+++ b/deeptutor/skills/builtin/xlsx/SKILL.md
@@ -57,9 +57,10 @@ Pick by what the deliverable needs:
 
 ```python
 import pandas as pd
-df = pd.read_excel("in.xlsx")                          # first sheet
-sheets = pd.read_excel("in.xlsx", sheet_name=None)     # dict of all sheets
-df = pd.read_excel("in.xlsx", dtype={"id": str})       # stop id->float coercion
+
+df = pd.read_excel("in.xlsx")  # first sheet
+sheets = pd.read_excel("in.xlsx", sheet_name=None)  # dict of all sheets
+df = pd.read_excel("in.xlsx", dtype={"id": str})  # stop id->float coercion
 ```
 
 To read **computed results** of formulas (not the formula text), use openpyxl
@@ -67,8 +68,9 @@ with `data_only=True` — returns the value Excel last cached:
 
 ```python
 from openpyxl import load_workbook
+
 wb = load_workbook("in.xlsx", data_only=True)
-val = wb["Sheet1"]["B10"].value     # None if Excel never opened/saved the file
+val = wb["Sheet1"]["B10"].value  # None if Excel never opened/saved the file
 ```
 
 Gotcha: never `save()` a workbook loaded with `data_only=True` — that discards
@@ -83,19 +85,21 @@ Large file: `load_workbook(path, read_only=True)` streams rows cheaply.
 from openpyxl import Workbook
 from openpyxl.styles import Font, PatternFill, Alignment
 
-wb = Workbook(); ws = wb.active; ws.title = "Summary"
-ws.append(["Region", "Sales"])              # header row
+wb = Workbook()
+ws = wb.active
+ws.title = "Summary"
+ws.append(["Region", "Sales"])  # header row
 for r in [("West", 120), ("East", 95)]:
     ws.append(r)
-ws["B4"] = "=SUM(B2:B3)"                     # see formula gotcha above
+ws["B4"] = "=SUM(B2:B3)"  # see formula gotcha above
 
 ws["A1"].font = Font(bold=True)
 ws["A1"].fill = PatternFill("solid", fgColor="DDDDDD")
 ws["A1"].alignment = Alignment(horizontal="center")
-ws["B2"].number_format = "#,##0"            # thousands separator
+ws["B2"].number_format = "#,##0"  # thousands separator
 ws.column_dimensions["A"].width = 18
-ws.freeze_panes = "A2"                       # freeze header
-wb.create_sheet("Detail")                    # second sheet
+ws.freeze_panes = "A2"  # freeze header
+wb.create_sheet("Detail")  # second sheet
 wb.save("out.xlsx")
 ```
 
@@ -112,7 +116,8 @@ rewrites the whole sheet, losing styles).
 
 ```python
 from openpyxl import load_workbook
-wb = load_workbook("in.xlsx")               # keep formulas (data_only=False)
+
+wb = load_workbook("in.xlsx")  # keep formulas (data_only=False)
 ws = wb["Sheet1"]
 ws["C2"] = "Updated"
 wb.save("in.xlsx")
@@ -129,10 +134,13 @@ affected formulas yourself, or avoid structural shifts in formula-heavy sheets.
 
 ```python
 from openpyxl.chart import BarChart, Reference
-ch = BarChart(); ch.title = "Sales"
-data = Reference(ws, min_col=2, min_row=1, max_row=3)   # include header for title
+
+ch = BarChart()
+ch.title = "Sales"
+data = Reference(ws, min_col=2, min_row=1, max_row=3)  # include header for title
 cats = Reference(ws, min_col=1, min_row=2, max_row=3)
-ch.add_data(data, titles_from_data=True); ch.set_categories(cats)
+ch.add_data(data, titles_from_data=True)
+ch.set_categories(cats)
 ws.add_chart(ch, "E2")
 ```
 LineChart / PieChart / ScatterChart follow the same shape.
@@ -145,10 +153,15 @@ that recalc surfaced (`#REF!` bad reference, `#DIV/0!` zero denominator,
 
 ```python
 from openpyxl import load_workbook
+
 wb = load_workbook("out.xlsx", data_only=True)
-errs = [f"{s}!{c.coordinate}={c.value}"
-        for s in wb.sheetnames for row in wb[s].iter_rows() for c in row
-        if isinstance(c.value, str) and c.value.startswith("#")]
+errs = [
+    f"{s}!{c.coordinate}={c.value}"
+    for s in wb.sheetnames
+    for row in wb[s].iter_rows()
+    for c in row
+    if isinstance(c.value, str) and c.value.startswith("#")
+]
 print(errs or "clean")
 ```
 This only catches errors in *cached* values. If you wrote formulas and couldn't
@@ -159,7 +172,7 @@ sidesteps this.
 ## CSV / TSV
 
 ```python
-df = pd.read_csv("in.csv")                   # sep="\t" for TSV
+df = pd.read_csv("in.csv")  # sep="\t" for TSV
 df.to_csv("out.csv", index=False)
 ```
 For messy input (junk rows, header not on row 1, ragged columns): inspect raw

From e73531d6850c98fc253812ab30cf3c451eefb647 Mon Sep 17 00:00:00 2001
From: santhreal <64453045+santhreal@users.noreply.github.com>
Date: Fri, 24 Jul 2026 23:46:37 -0700
Subject: [PATCH 3/5] chore: trigger CI after skill format

---
 tests/agents/visualize/test_extract_code_block.py | 1 +
 1 file changed, 1 insertion(+)

diff --git a/tests/agents/visualize/test_extract_code_block.py b/tests/agents/visualize/test_extract_code_block.py
index c979a243e9..5145daf4db 100644
--- a/tests/agents/visualize/test_extract_code_block.py
+++ b/tests/agents/visualize/test_extract_code_block.py
@@ -22,3 +22,4 @@ def test_extract_code_block_language_miss_is_empty_for_or_fallback() -> None:
     assert extract_code_block(raw, "javascript") == ""
     extracted = extract_code_block(raw, "javascript") or extract_code_block(raw)
     assert extracted == cfg
+

From f7f849d866d815a59f249ad2a85f8e680fca7353 Mon Sep 17 00:00:00 2001
From: santhreal <64453045+santhreal@users.noreply.github.com>
Date: Thu, 30 Jul 2026 10:33:24 -0700
Subject: [PATCH 4/5] test: document extract code block contracts

---
 tests/agents/visualize/test_extract_code_block.py | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/tests/agents/visualize/test_extract_code_block.py b/tests/agents/visualize/test_extract_code_block.py
index 5145daf4db..0402c235b8 100644
--- a/tests/agents/visualize/test_extract_code_block.py
+++ b/tests/agents/visualize/test_extract_code_block.py
@@ -6,17 +6,20 @@
 
 
 def test_extract_code_block_closing_fence_without_leading_newline() -> None:
+    """Extract a closing fence attached to the final content line."""
     raw = "```mermaid\ngraph TD\n  A-->B```"
     assert extract_code_block(raw, "mermaid") == "graph TD\n  A-->B"
     assert extract_code_block(raw) == "graph TD\n  A-->B"
 
 
 def test_extract_code_block_normal_fenced_block() -> None:
+    """Extract a standard fenced block with its requested language."""
     raw = "```javascript\nconst x = 1;\n```"
     assert extract_code_block(raw, "javascript") == "const x = 1;"
 
 
 def test_extract_code_block_language_miss_is_empty_for_or_fallback() -> None:
+    """Return empty on a language miss so generic extraction can run."""
     cfg = '{"type": "bar", "data": {"labels": ["A"], "datasets": [{"data": [1]}]}}'
     raw = f"```json\n{cfg}\n```\nThanks."
     assert extract_code_block(raw, "javascript") == ""

From 288dcf1571528f47fa44e8a08dd61bf2f3cfcbd8 Mon Sep 17 00:00:00 2001
From: santhreal <64453045+santhreal@users.noreply.github.com>
Date: Thu, 30 Jul 2026 10:45:04 -0700
Subject: [PATCH 5/5] style: format extract code block tests

---
 tests/agents/visualize/test_extract_code_block.py | 1 -
 1 file changed, 1 deletion(-)

diff --git a/tests/agents/visualize/test_extract_code_block.py b/tests/agents/visualize/test_extract_code_block.py
index 0402c235b8..bdbee712e4 100644
--- a/tests/agents/visualize/test_extract_code_block.py
+++ b/tests/agents/visualize/test_extract_code_block.py
@@ -25,4 +25,3 @@ def test_extract_code_block_language_miss_is_empty_for_or_fallback() -> None:
     assert extract_code_block(raw, "javascript") == ""
     extracted = extract_code_block(raw, "javascript") or extract_code_block(raw)
     assert extracted == cfg
-