Add a check for articles formatting

.ci/check_knowledge_markup.py

diff --git a/.ci/check_knowledge_markup.py b/.ci/check_knowledge_markup.py
new file mode 100644
index 0000000..2171c8b
--- /dev/null
+++ b/.ci/check_knowledge_markup.py
@@ -0,0 +1,86 @@
+import re
+import sys
+from pathlib import Path
+
+MARKDOWN_LINK = re.compile(r'(?<!!)\[[^\]]+\]\([^)]+\)')
+MARKDOWN_BOLD = re.compile(r'\*\*[^*\n]+\*\*')
+MARKDOWN_ITALIC = re.compile(r'(?<!\*)\*(?!\*)[^*\n]+\*(?!\*)')
+MARKDOWN_CODE = re.compile(r'`[^`\n]+`')
+MARKDOWN_HEADING = re.compile(r'(?m)^#{1,6}\s')
+
+ALL_PATTERNS = [
+    (MARKDOWN_LINK,    "markdown link [text](url)"),
+    (MARKDOWN_BOLD,    "markdown bold **text**"),
+    (MARKDOWN_ITALIC,  "markdown italic *text*"),
+    (MARKDOWN_CODE,    "markdown code `backtick`"),
+    (MARKDOWN_HEADING, "markdown heading #"),
+]
+
+INLINE_TAG_RE = re.compile(
+    r'<(figcaption|p|li|dt|dd|td|th)[^>]*>[^\n]+</\1>',
+    re.IGNORECASE,
+)
+
+ADJACENT_BLOCK_RE = re.compile(
+    r'<(figure|aside|div|section|blockquote)[^>]*>\n([^\n].*?)\n</\1>',
+    re.DOTALL | re.IGNORECASE,
+)
+
+
+def check_inline_tags(content, md_file, errors):
+    """Catch markdown inside single-line HTML tags like <figcaption>...</figcaption>."""
+    for match in INLINE_TAG_RE.finditer(content):
+        line = match.group(0)
+        line_num = content[:match.start()].count('\n') + 1
+        for pattern, description in ALL_PATTERNS:
+            if pattern.search(line):
+                errors.append(
+                    f"{md_file}:{line_num}: {description} inside inline HTML tag:\n"
+                    f"  {line[:200]!r}"
+                )
+                break
+
+
+def check_adjacent_blocks(content, md_file, errors):
+    """Catch markdown on lines immediately adjacent to block HTML tags (no blank line)."""
+    for block_match in ADJACENT_BLOCK_RE.finditer(content):
+        inner = block_match.group(2)
+        if not inner.strip():
+            continue
+        inner_no_images = re.sub(r'!\[[^\]]*\]\([^)]*\)', '', inner)
+        line_num = content[:block_match.start()].count('\n') + 1
+        for pattern, description in ALL_PATTERNS:
+            if pattern.search(inner_no_images):
+                errors.append(
+                    f"{md_file}:{line_num}: {description} adjacent to HTML block tag "
+                    f"(no blank line separator):\n"
+                    f"  {block_match.group(0)[:200].strip()!r}"
+                )
+                break
+
+
+def main():
+    knowledge_root = Path("cc-by-sa/knowledge")
+    files = sorted(knowledge_root.rglob("*.md"))
+
+    if not files:
+        print("No .md files found under cc-by-sa/knowledge/")
+        sys.exit(1)
+
+    errors = []
+    for md_file in files:
+        content = md_file.read_text(encoding="utf-8")
+        check_inline_tags(content, md_file, errors)
+        check_adjacent_blocks(content, md_file, errors)
+
+    if errors:
+        print(f"❌ Found {len(errors)} markdown-in-HTML error(s):\n")
+        for e in errors:
+            print(f"  {e}\n")
+        sys.exit(1)
+
+    print(f"✅ Checked {len(files)} files — no markdown markup inside HTML blocks.")
+
+
+if __name__ == "__main__":
+    main()
\ No newline at end of file

.github/workflows/check_knowledge.yaml

diff --git a/.github/workflows/check_knowledge.yaml b/.github/workflows/check_knowledge.yaml
new file mode 100644
index 0000000..bee74f2
--- /dev/null
+++ b/.github/workflows/check_knowledge.yaml
@@ -0,0 +1,20 @@
+name: Check Knowledge content markup
+
+on:
+  pull_request:
+    paths:
+      - "cc-by-sa/knowledge/**/*.md"
+
+jobs:
+  check-markdown-in-html:
+    name: No markdown inside HTML blocks
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+
+      - uses: actions/setup-python@v5
+        with:
+          python-version: "3.10"
+
+      - name: Check for markdown markup inside HTML
+        run: python3 .ci/check_knowledge_markup.py
\ No newline at end of file