.github/workflows/release.yml (1)

207-211: Fragile header-skip assumption could duplicate or lose content.

tail -n +7 assumes the existing CHANGELOG.md has exactly 6 header lines. If the header format varies (e.g., different number of blank lines, additional metadata), this could either duplicate header content or skip actual changelog entries.

Consider a more robust approach that finds the first version entry dynamically:

♻️ Proposed fix using pattern-based skip

-            # Append previous changelog content (skip the header block — first 6 lines)
-            if [ -f CHANGELOG.md ]; then
-              tail -n +7 CHANGELOG.md
-            fi
+            # Append previous changelog content (skip header, start from first version entry)
+            if [ -f CHANGELOG.md ]; then
+              # Find line number of first "## [" (version entry) and print from there
+              FIRST_VERSION_LINE=$(grep -n '^## \[' CHANGELOG.md | head -1 | cut -d: -f1)
+              if [ -n "$FIRST_VERSION_LINE" ]; then
+                tail -n +"$FIRST_VERSION_LINE" CHANGELOG.md
+              fi
+            fi

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In @.github/workflows/release.yml around lines 207 - 211, The current use of
"tail -n +7 CHANGELOG.md" is fragile; update the block that writes to
CHANGELOG_NEW.md to dynamically find the first version heading instead of
assuming 6 header lines — detect the first changelog entry by matching a version
heading pattern (e.g., a line matching '^## \[' or '^#\s*[0-9]\+' / '^##\s*\\[')
and print the file starting from that matched line; replace the "tail -n +7
CHANGELOG.md" invocation with a pattern-based command (awk/sed/grep) that
searches CHANGELOG.md for the first version header and outputs from there into
CHANGELOG_NEW.md so headers are neither duplicated nor accidentally skipped.