Generated banner for Markdown To Word Mermaid
Generated logo for Markdown To Word Mermaid

Markdown To Word Mermaid

GitHubOfficialFirst committed by furukawa658Distributed by GitHub

  • Skill
  • Tool integration
  • Automation and integrations
  • MIT
  • Low risk
  • Has scripts

From the author

Export Markdown containing Mermaid code fences to Word (.docx) with embedded SVG diagrams and PNG fallbacks. Use when diagram sharpness, Office-compatible labels, or page-aware placement matters. Not for Word-to-Markdown conversion, editable Office shapes, or general slide creation.
— the description by furukawa658

AutoGineer summary

Written by AutoGineer from the skill’s own files, not by the author

This skill automates the conversion of Markdown files containing Mermaid code fences into Microsoft Word (.docx) documents. It renders diagrams locally using the Mermaid CLI, embedding them as vector SVGs with 3x PNG fallbacks to ensure compatibility and visual clarity. The process handles document structure, including headings and tables, while applying specific logic for diagram scaling and page-aware placement to prevent diagrams from splitting across pages.

What it does

  • Renders Mermaid code fences into SVG and PNG formats locally.
  • Embeds vector SVGs into Word documents using Office OOXML relationships.
  • Applies page-aware placement to keep tall diagrams together.
  • Scales oversized diagrams to fit the page body height.
  • Preserves Markdown document structure like headings, lists, and tables.
  • Includes 3x PNG fallbacks for diagrams within the generated DOCX.

When to use it

  • When you need to export technical documentation containing Mermaid diagrams to Word.
  • When diagram labels must remain sharp and readable in a printed or zoomed Word document.
  • When you require diagrams to be kept together on a single page rather than split across pages.
  • When you need to maintain consistent font styling across the document and diagrams.

Advantages

  • Provides vector-based diagram rendering for improved clarity.
  • Ensures diagrams are not split across page breaks.
  • Does not require Microsoft Word to be installed to generate the file.
  • Uses atomic output staging to prevent partial or corrupted files.

Usage examples

“Convert my design_doc.md to a Word document named design_doc.docx.”
The agent runs the conversion script, rendering all Mermaid diagrams into the document with SVG and PNG fallbacks.
“Export technical_report.md to Word using the Noto Sans JP font.”
The agent generates the DOCX file, applying the specified font to both the document text and the rendered Mermaid diagrams.

Install

  1. Download the skill as a ZIP:

    https://autogineer.com/agent-skills/github/markdown-to-word-mermaid/download
  2. Unzip it into the cross-client skills folder of your project:

    mkdir -p .agents/skills && unzip markdown-to-word-mermaid.zip -d .agents/skills

Documentation by furukawa658

Markdown to Word with Mermaid

Export technical documents without manually replacing Mermaid fences with screenshots. The bundled converter renders each diagram locally, embeds the SVG through Office OOXML relationships, and retains a 3x PNG fallback. Word is not required to generate the file; an SVG-capable Word version is needed to inspect vector rendering.

When to use

  • A design document, runbook, or report contains Mermaid code fences and must be delivered as DOCX.

  • Diagram labels must remain sharp when zoomed or printed.

  • A tall diagram needs page-aware placement rather than being split across pages.

For general Markdown-to-DOCX conversion with existing PNG images, a general converter such as md-to-docx may be simpler. This skill focuses on rendering and packaging Mermaid diagrams. SVG embedding does not create editable Word shapes, connectors, or text boxes.

Requirements and setup

  • Python 3.10+ and the packages in scripts/requirements.txt.

  • Node.js compatible with the chosen Mermaid CLI release, plus @mermaid-js/mermaid-cli.

  • A local browser usable by Puppeteer and an installed font covering the document's characters.

  • Default font: Noto Sans JP. Use --font-family to select another installed family.

Resolve paths relative to this skill's installation directory, not an assumed repository name. The examples below run from the skill directory. Use the user's selected Python environment (python, python3, or its absolute path).

Before installing dependencies, confirm the environment and obtain permission if required. Reuse an existing Mermaid CLI when available. Otherwise install it into a user-approved local tools directory, not globally:

python -m pip install -r scripts/requirements.txt
npm install --prefix ./run/mermaid-cli @mermaid-js/mermaid-cli

The local executable is ./run/mermaid-cli/node_modules/.bin/mmdc on POSIX and ./run/mermaid-cli/node_modules/.bin/mmdc.cmd on Windows. Keep installed dependencies out of the submitted skill package. Lock the selected dependency versions in the consuming project for reproducible installations.

On Windows the converter detects standard Chrome and Edge paths. For another browser location, set MERMAID_BROWSER_PATH to its absolute executable path:

$env:MERMAID_BROWSER_PATH = "C:\Program Files\Google\Chrome\Application\chrome.exe"

On POSIX, use export MERMAID_BROWSER_PATH=/absolute/path/to/browser, or use Puppeteer's installed browser. Do not disable the browser sandbox just to make a conversion pass.

Workflow

  1. Confirm the input Markdown file, output destination, and font. If no destination is provided, use the input path with a .docx extension.

  2. Check that the output does not exist. Never delete or overwrite an existing document; ask for a different destination.

  3. Run scripts/markdown_to_word.py. Preserve the user's diagram source and meaning. Do not silently repair invalid Mermaid.

  4. Check the generated package and, when Word is available, inspect the document visually. Report package checks and visual checks separately.

With mmdc on PATH:

python scripts/markdown_to_word.py input.md --output output.docx

Using a workspace-local CLI on Windows:

python scripts/markdown_to_word.py input.md --output output.docx --mermaid-cli ./run/mermaid-cli/node_modules/.bin/mmdc.cmd

Selecting an installed font:

python scripts/markdown_to_word.py input.md --output output.docx --font-family "Noto Sans JP" --timeout 120

For a reproducible example, convert assets/example.md to a new output path. It includes English and Japanese labels, multiple diagrams, and a tall diagram.

Conversion behavior

  • Headings, paragraphs, emphasis, lists (preserving ordered list start numbers and separate list boundaries), blockquotes, tables, and code blocks become Word content.

  • Each Mermaid fence renders to SVG and a 3x PNG with a per-command timeout (default: 120 seconds). Office SVG extensions reference the SVG; the normal image reference retains the PNG fallback, and drawing metadata (title and descr) includes the section heading or Mermaid accTitle/accDescr plus the diagram type.

  • Output DOCX files are staged and SVG-injected inside a temporary directory before moving to the destination path so failed runs never leave partial files behind.

  • The selected font is applied to Word paragraph styles and Mermaid's font configuration. Code runs use Consolas. Fonts are not embedded; unavailable fonts may be substituted.

  • htmlLabels: false is set globally and for flowcharts to avoid HTML foreignObject labels that Office may not display. Check actual SVG output for unsupported constructs.

  • Diagrams are centered, keep their aspect ratio, and use a non-splitting paragraph. Diagrams occupying at least 70% of the page body height start on a new page. Oversized diagrams are scaled to the page body height. Large diagrams generate a readability warning.

  • A simple single-line frontmatter title becomes the document title; other frontmatter fields are removed from the body. This is not a full YAML metadata parser.

  • Ordinary Markdown image links remain alt text and a source path or URL; they are not embedded. Raw HTML is reduced to text, not executed.

Stop conditions and limitations

  • Missing input, wrong extensions, an existing output, missing CLI, invalid Mermaid, rendering failure, or unusable SVG dimensions stop the conversion with an error.

  • The converter does not call an external rendering service or a separate AI API. Dependency installation and browser execution are not an offline sandbox. Review untrusted diagram content, including external resources, before rendering it.

  • Diagrams are not semantically editable Office objects. Word rendering, font substitution, and final pagination depend on the viewer and must be checked there.

  • A successfully created DOCX is not proof that every diagram label is readable. Split large diagrams in the source only with the user's approval.

Verification

Run the test suite from the skill directory:

python -m unittest discover -s tests -v

The ordinary suite mocks Mermaid rendering. To run the real-renderer test as well, set MERMAID_CLI to the CLI executable's absolute path, then run the command again. On Windows PowerShell:

$env:MERMAID_CLI = (Join-Path $env:APPDATA "npm\mmdc.cmd")
python -m unittest discover -s tests -v

That test renders the bundled sample and checks the resulting DOCX's SVG references, PNG fallbacks, labels, and drawing dimensions. It does not automate Word visual inspection.

Before reporting completion:

  • Count Mermaid fences and verify the same number of SVG parts and SVG image references in the DOCX.

  • Confirm PNG fallback references exist and no label foreignObject remains in the sample's SVGs.

  • In Word, inspect node names, branch labels, font, diagram placement, and page boundaries; check zoomed or printed output when sharpness matters.

  • If Word or the actual renderer was unavailable, state which checks were not performed.

For contribution scope, evaluation cases, and a draft submission description, see references/publication.md. That file is maintainer guidance, not permission to publish.

Gotchas

  • mmdc.cmd needs Windows command handling; pass its path with --mermaid-cli rather than rewriting the script.

  • A missing font can affect both PNG rendering and SVG display. Use a family installed on the rendering and viewing machines.

  • Keep the SVG and PNG paired. Replacing SVG with a screenshot defeats the vector-preserving workflow.

  • Package validation, renderer tests, agent-selection evaluation, and human visual acceptance are different checks.

Files

scripts/markdown_to_word.py

from __future__ import annotations

import argparse
import json
import os
import re
import shutil
import subprocess
import sys
import tempfile
import zipfile
from pathlib import Path
from typing import Any

from docx import Document
from docx.enum.style import WD_STYLE_TYPE
from docx.enum.text import WD_ALIGN_PARAGRAPH
from docx.oxml import OxmlElement
from docx.oxml.ns import qn
from docx.shared import Emu, Pt
from lxml import etree
from markdown_it import MarkdownIt


SVG_NAMESPACE = "http://schemas.microsoft.com/office/drawing/2016/SVG/main"
SVG_EXTENSION_URI = "{96DAC541-7B7A-43D3-8B79-37D633B846F1}"
IMAGE_RELATIONSHIP = "http://schemas.openxmlformats.org/officeDocument/2006/relationships/image"
RELATIONSHIPS_NAMESPACE = "http://schemas.openxmlformats.org/package/2006/relationships"
CONTENT_TYPES_NAMESPACE = "http://schemas.openxmlformats.org/package/2006/content-types"
WORD_NAMESPACE = "http://schemas.openxmlformats.org/wordprocessingml/2006/main"
DRAWING_NAMESPACE = "http://schemas.openxmlformats.org/drawingml/2006/main"
WORD_DRAWING_NAMESPACE = "http://schemas.openxmlformats.org/drawingml/2006/wordprocessingDrawing"
RELATIONSHIP_ATTRIBUTE = "{http://schemas.openxmlformats.org/officeDocument/2006/relationships}embed"
SVG_CONFIG = {
    "theme": "default",
    "htmlLabels": False,
    "themeVariables": {"fontFamily": "Noto Sans JP, sans-serif"},
    "flowchart": {"htmlLabels": False},
}
PNG_FALLBACK = (
    b"\x89PNG\r\n\x1a\n\x00\x00\x00\rIHDR\x00\x00\x00\x01\x00\x00\x00\x01"
    b"\x08\x06\x00\x00\x00\x1f\x15\xc4\x89\x00\x00\x00\x0bIDAT"
    b"x\x9cc\x00\x02\x00\x00\x05\x00\x01\xa5\xf6E\x8c\x00\x00\x00\x00IEND\xaeB`\x82"
)


def _resolve_cli(cli_name: str) -> str:
    resolved = shutil.which(cli_name)
    if resolved:
        return resolved
    candidate = Path(cli_name)
    if candidate.is_file():
        return str(candidate.resolve())
    raise FileNotFoundError(
        f"Mermaid CLI '{cli_name}' was not found. Install @mermaid-js/mermaid-cli or pass --mermaid-cli."
    )


def _command_prefix(cli: str) -> list[str]:
    cli_path = Path(cli)
    if os.name != "nt" or cli_path.suffix.lower() not in {".cmd", ".bat"}:
        return [cli]
    node = shutil.which("node")
    if node is None:
        raise FileNotFoundError("Node.js is required to invoke Mermaid CLI without a shell on Windows.")
    for directory in cli_path.resolve().parents:
        candidates = (
            directory / "@mermaid-js/mermaid-cli/src/cli.js",
            directory / "node_modules/@mermaid-js/mermaid-cli/src/cli.js",
        )
        entry_point = next((candidate for candidate in candidates if candidate.is_file()), None)
        if entry_point is not None:
            return [node, str(entry_point)]
    raise FileNotFoundError(
        "Could not locate the Mermaid CLI JavaScript entry point beside the Windows command shim."
    )


def _puppeteer_config(directory: Path) -> list[str]:
    browser = os.environ.get("MERMAID_BROWSER_PATH")
    candidates = [Path(browser)] if browser else []
    if os.name == "nt":
        candidates.extend(
            [
                Path(os.environ.get("PROGRAMFILES", "C:/Program Files"))
                / "Google/Chrome/Application/chrome.exe",
                Path(os.environ.get("PROGRAMFILES(X86)", "C:/Program Files (x86)"))
                / "Microsoft/Edge/Application/msedge.exe",
            ]
        )
    browser_path = next((path.resolve() for path in candidates if path.is_file()), None)
    if browser_path is None:
        return []
    config_path = directory / "puppeteer-config.json"
    config_path.write_text(json.dumps({"executablePath": str(browser_path)}), encoding="utf-8")
    return ["-p", str(config_path)]


def _run_command(command: list[str], *, timeout: float = 120.0) -> None:
    try:
        result = subprocess.run(
            command,
            capture_output=True,
            text=True,
            encoding="utf-8",
            errors="replace",
            timeout=timeout,
        )
    except subprocess.TimeoutExpired as error:
        raise RuntimeError(f"Mermaid rendering timed out after {timeout:g}s.") from error
    if result.returncode:
        details = result.stderr.strip() or result.stdout.strip()
        raise RuntimeError(f"Mermaid rendering failed (exit {result.returncode}): {details}")


def _render_mermaid(
    code: str, diagram_number: int, temp_root: Path, cli: str,
    font_family: str = "Noto Sans JP",
    *,
    timeout: float = 120.0,
) -> tuple[Path, Path, float, float]:
    diagram_dir = temp_root / f"diagram-{diagram_number}"
    diagram_dir.mkdir()
    source_path = diagram_dir / "diagram.mmd"
    config_path = diagram_dir / "mermaid-config.json"
    svg_path = diagram_dir / "diagram.svg"
    png_path = diagram_dir / "diagram.png"
    source_path.write_text(code.rstrip() + "\n", encoding="utf-8")
    config = {**SVG_CONFIG, "themeVariables": {"fontFamily": f"{font_family}, sans-serif"}}
    config_path.write_text(json.dumps(config, ensure_ascii=False), encoding="utf-8")

    command_prefix = _command_prefix(cli)
    common = ["-i", str(source_path), "-c", str(config_path), *_puppeteer_config(diagram_dir)]
    _run_command([*command_prefix, *common, "-o", str(svg_path), "-e", "svg"], timeout=timeout)
    _run_command([*command_prefix, *common, "-o", str(png_path), "-e", "png", "-s", "3"], timeout=timeout)

    if not svg_path.is_file() or not png_path.is_file():
        raise RuntimeError("Mermaid CLI completed without producing both SVG and PNG outputs.")
    width, height = _validate_svg(svg_path)
    return svg_path, png_path, width, height


def _svg_length(value: str | None) -> float | None:
    if not value:
        return None
    match = re.fullmatch(r"\s*([0-9]*\.?[0-9]+)(?:px)?\s*", value)
    return float(match.group(1)) if match else None


def _validate_svg(svg_path: Path) -> tuple[float, float]:
    parser = etree.XMLParser(resolve_entities=False, no_network=True, recover=False)
    try:
        root = etree.parse(str(svg_path), parser).getroot()
    except etree.XMLSyntaxError as error:
        raise RuntimeError(f"Mermaid CLI produced invalid SVG: {error}") from error

    if etree.QName(root).localname != "svg":
        raise RuntimeError("Mermaid CLI output does not have an SVG root element.")

    view_box = root.get("viewBox")
    if view_box:
        values = [float(value) for value in re.split(r"[ ,]+", view_box.strip())]
        if len(values) == 4 and values[2] > 0 and values[3] > 0:
            return values[2], values[3]

    width = _svg_length(root.get("width"))
    height = _svg_length(root.get("height"))
    if width and height:
        return width, height
    raise RuntimeError("Mermaid SVG has no usable viewBox or width/height dimensions.")


def _add_inline_content(paragraph: Any, tokens: list[Any], *, header: bool = False) -> None:
    emphasis: list[str] = []
    link_stack: list[str] = []
    for token in tokens:
        if token.type in {"strong_open", "em_open", "s_open"}:
            emphasis.append(token.type.removesuffix("_open"))
        elif token.type in {"strong_close", "em_close", "s_close"}:
            if emphasis:
                emphasis.pop()
        elif token.type == "link_open":
            link_stack.append(token.attrGet("href") or "")
        elif token.type == "link_close":
            href = link_stack.pop() if link_stack else ""
            if href:
                paragraph.add_run(f" ({href})")
        elif token.type in {"text", "code_inline", "html_inline"}:
            text = token.content
            if token.type == "html_inline":
                text = re.sub(r"<[^>]*>", "", text)
            run = paragraph.add_run(text)
            run.bold = header or "strong" in emphasis
            run.italic = "em" in emphasis
            run.font.strike = "s" in emphasis
            if token.type == "code_inline":
                run.font.name = "Consolas"
                run.font.size = Pt(9)
        elif token.type in {"softbreak", "hardbreak"}:
            paragraph.add_run().add_break()
        elif token.type == "image":
            paragraph.add_run(token.content or "[image]")
            source = token.attrGet("src")
            if source:
                paragraph.add_run(f" ({source})")


def _new_paragraph(document: Any, list_stack: list[dict[str, Any]], quote_depth: int) -> Any:
    if quote_depth:
        return document.add_paragraph(style="Quote")
    if list_stack:
        depth = min(len(list_stack), 3)
        suffix = f" {depth}" if depth > 1 else ""
        current_list = list_stack[-1]
        if current_list.get("item_continuation", False):
            return document.add_paragraph(style=f"List Continue{suffix}")
        current_list["item_continuation"] = True
        if current_list["ordered"]:
            paragraph = document.add_paragraph(style=f"List Continue{suffix}")
            run = paragraph.add_run(f"{current_list['next_number']}. ")
            run.bold = False
            current_list["next_number"] += 1
            return paragraph
        return document.add_paragraph(style=f"List Bullet{suffix}")
    return document.add_paragraph()


def _diagram_description(code: str, diagram_number: int, current_heading: str | None) -> tuple[str, str]:
    acc_title = re.search(r"(?m)^\s*accTitle\s*:\s*(.+?)\s*$", code)
    acc_descr = re.search(r"(?m)^\s*accDescr\s*:\s*(.+?)\s*$", code)
    first_statement = next(
        (
            line.strip()
            for line in code.splitlines()
            if line.strip() and not line.strip().startswith("%%")
        ),
        "diagram",
    )
    diagram_kind = first_statement.split()[0]
    title = (
        acc_title.group(1)
        if acc_title
        else f"{current_heading} — Mermaid diagram {diagram_number}"
        if current_heading
        else f"Mermaid diagram {diagram_number}"
    )
    if acc_descr:
        description = f"{title}: {acc_descr.group(1)} ({diagram_kind}; vector SVG)."
    else:
        description = f"{title} ({diagram_kind}; vector SVG)."
    return title, description


def _add_code_block(document: Any, content: str) -> None:
    paragraph = document.add_paragraph(style="No Spacing")
    paragraph.paragraph_format.keep_together = True
    run = paragraph.add_run(content.rstrip("\n"))
    run.font.name = "Consolas"
    run.font.size = Pt(9)


def _add_diagram(
    document: Any,
    svg_path: Path,
    png_path: Path,
    diagram_number: int,
    svg_width: float,
    svg_height: float,
    *,
    code: str = "",
    current_heading: str | None = None,
) -> None:
    section = document.sections[-1]
    available_width = section.page_width - section.left_margin - section.right_margin
    available_height = section.page_height - section.top_margin - section.bottom_margin
    width = float(available_width)
    height = width * svg_height / svg_width
    if height >= available_height * 0.70:
        scale = min(1.0, float(available_height) / height)
        width *= scale
        height *= scale
    if height >= available_height * 0.70:
        print(
            f"Warning: Mermaid diagram {diagram_number} uses most of a page; verify label readability in Word.",
            file=sys.stderr,
        )

    paragraph = document.add_paragraph()
    paragraph.alignment = WD_ALIGN_PARAGRAPH.CENTER
    paragraph.paragraph_format.keep_together = True
    if height >= available_height * 0.70:
        paragraph.paragraph_format.page_break_before = True
    run = paragraph.add_run()
    picture = run.add_picture(str(png_path), width=Emu(round(width)), height=Emu(round(height)))
    title, description = _diagram_description(code, diagram_number, current_heading)
    picture._inline.docPr.set("name", f"mermaid-svg-{diagram_number}")
    picture._inline.docPr.set("title", title)
    picture._inline.docPr.set("descr", description)


def _render_table(document: Any, tokens: list[Any], start: int) -> int:
    rows: list[tuple[list[Any], bool]] = []
    current_cells: list[Any] | None = None
    in_header = False
    index = start + 1
    while index < len(tokens) and tokens[index].type != "table_close":
        token = tokens[index]
        if token.type == "thead_open":
            in_header = True
        elif token.type == "tbody_open":
            in_header = False
        elif token.type == "tr_open":
            current_cells = []
        elif token.type == "inline" and current_cells is not None:
            current_cells.append(token)
        elif token.type == "tr_close" and current_cells is not None:
            rows.append((current_cells, in_header))
            current_cells = None
        index += 1

    column_count = max((len(cells) for cells, _ in rows), default=0)
    if column_count:
        table = document.add_table(rows=len(rows), cols=column_count)
        try:
            table.style = "Table Grid"
        except KeyError:
            pass
        for row_index, (cells, is_header) in enumerate(rows):
            for column_index, inline_token in enumerate(cells):
                cell = table.cell(row_index, column_index)
                _add_inline_content(cell.paragraphs[0], inline_token.children or [], header=is_header)
    return index + 1


def _convert_blocks(
    document: Any, markdown_text: str, temp_root: Path, cli_name: str,
    font_family: str = "Noto Sans JP",
    *,
    timeout: float = 120.0,
) -> int:
    parser = MarkdownIt("commonmark").enable("table")
    tokens = parser.parse(markdown_text)
    list_stack: list[dict[str, Any]] = []
    quote_depth = 0
    current_paragraph = None
    in_heading = False
    current_heading: str | None = None
    diagram_number = 0
    cli: str | None = None
    index = 0

    while index < len(tokens):
        token = tokens[index]
        token_type = token.type

        if token_type == "table_open":
            index = _render_table(document, tokens, index)
            current_paragraph = None
            continue
        if token_type in {"bullet_list_open", "ordered_list_open"}:
            start_attr = token.attrGet("start") if token_type == "ordered_list_open" else None
            start_number = int(start_attr) if start_attr and str(start_attr).isdigit() else 1
            list_stack.append(
                {
                    "ordered": token_type == "ordered_list_open",
                    "next_number": start_number,
                    "item_continuation": False,
                }
            )
        elif token_type == "list_item_open":
            if list_stack:
                list_stack[-1]["item_continuation"] = False
        elif token_type == "list_item_close":
            if list_stack:
                list_stack[-1]["item_continuation"] = False
        elif token_type in {"bullet_list_close", "ordered_list_close"}:
            if list_stack:
                list_stack.pop()
        elif token_type == "blockquote_open":
            quote_depth += 1
        elif token_type == "blockquote_close":
            quote_depth = max(0, quote_depth - 1)
        elif token_type == "paragraph_open":
            current_paragraph = _new_paragraph(document, list_stack, quote_depth)
        elif token_type == "heading_open":
            level = int(token.tag[1])
            current_paragraph = document.add_paragraph(style=f"Heading {level}")
            current_paragraph.paragraph_format.keep_with_next = True
            in_heading = True
        elif token_type == "inline":
            if current_paragraph is None:
                current_paragraph = _new_paragraph(document, list_stack, quote_depth)
            _add_inline_content(current_paragraph, token.children or [])
            if in_heading and token.content.strip():
                current_heading = token.content.strip()
        elif token_type == "paragraph_close":
            current_paragraph = None
        elif token_type == "heading_close":
            current_paragraph = None
            in_heading = False
        elif token_type == "fence":
            if token.info.strip().split(maxsplit=1)[0:1] == ["mermaid"]:
                diagram_number += 1
                if cli is None:
                    cli = _resolve_cli(cli_name)
                svg_path, png_path, svg_width, svg_height = _render_mermaid(
                    token.content, diagram_number, temp_root, cli, font_family, timeout=timeout
                )
                _add_diagram(
                    document,
                    svg_path,
                    png_path,
                    diagram_number,
                    svg_width,
                    svg_height,
                    code=token.content,
                    current_heading=current_heading,
                )
            else:
                _add_code_block(document, token.content)
            current_paragraph = None
        elif token_type == "code_block":
            _add_code_block(document, token.content)
            current_paragraph = None
        elif token_type == "html_block":
            plain_text = re.sub(r"<[^>]*>", "", token.content).strip()
            if plain_text:
                document.add_paragraph(plain_text)
            current_paragraph = None
        elif token_type == "hr":
            document.add_paragraph()
            current_paragraph = None
        index += 1

    return diagram_number


def _svg_relationship_id(rels_root: Any) -> str:
    numbers = [
        int(match.group(1))
        for relationship in rels_root.findall(f"{{{RELATIONSHIPS_NAMESPACE}}}Relationship")
        if (match := re.fullmatch(r"rId([0-9]+)", relationship.get("Id", "")))
    ]
    return f"rId{max(numbers, default=0) + 1}"


def _inject_svg_parts(
    docx_path: Path, svg_paths: list[Path], *, temp_root: Path | None = None
) -> None:
    if not svg_paths:
        return

    ns = {
        "a": DRAWING_NAMESPACE,
        "asvg": SVG_NAMESPACE,
        "wp": WORD_DRAWING_NAMESPACE,
    }
    temporary_docx = (
        temp_root / "injected-output.docx"
        if temp_root is not None
        else docx_path.with_name(f"{docx_path.stem}.tmp{docx_path.suffix}")
    )
    with zipfile.ZipFile(docx_path, "r") as source, zipfile.ZipFile(
        temporary_docx, "w", compression=zipfile.ZIP_DEFLATED
    ) as target:
        document_xml = etree.fromstring(source.read("word/document.xml"))
        relationships_xml = etree.fromstring(source.read("word/_rels/document.xml.rels"))
        content_types_xml = etree.fromstring(source.read("[Content_Types].xml"))
        svg_default = content_types_xml.find(f"{{{CONTENT_TYPES_NAMESPACE}}}Default[@Extension='svg']")
        if svg_default is None:
            svg_default = etree.SubElement(content_types_xml, f"{{{CONTENT_TYPES_NAMESPACE}}}Default")
            svg_default.set("Extension", "svg")
            svg_default.set("ContentType", "image/svg+xml")

        replacements = {
            "word/document.xml": None,
            "word/_rels/document.xml.rels": None,
            "[Content_Types].xml": None,
        }
        for number, svg_path in enumerate(svg_paths, start=1):
            marker = f"mermaid-svg-{number}"
            inline = document_xml.xpath(
                f".//wp:inline[wp:docPr[@name='{marker}']]", namespaces=ns
            )
            if len(inline) != 1:
                raise RuntimeError(f"Could not locate the Word drawing for Mermaid diagram {number}.")
            blip = inline[0].find(".//a:blip", namespaces=ns)
            if blip is None:
                raise RuntimeError(f"Word drawing for Mermaid diagram {number} has no image reference.")

            relationship_id = _svg_relationship_id(relationships_xml)
            relationship = etree.SubElement(
                relationships_xml, f"{{{RELATIONSHIPS_NAMESPACE}}}Relationship"
            )
            relationship.set("Id", relationship_id)
            relationship.set("Type", IMAGE_RELATIONSHIP)
            relationship.set("Target", f"media/mermaid-{number}.svg")

            extension_list = blip.find("a:extLst", namespaces=ns)
            if extension_list is None:
                extension_list = etree.SubElement(blip, f"{{{DRAWING_NAMESPACE}}}extLst")
            extension = etree.SubElement(extension_list, f"{{{DRAWING_NAMESPACE}}}ext")
            extension.set("uri", SVG_EXTENSION_URI)
            svg_blip = etree.SubElement(extension, f"{{{SVG_NAMESPACE}}}svgBlip")
            svg_blip.set(RELATIONSHIP_ATTRIBUTE, relationship_id)
            replacements[f"word/media/mermaid-{number}.svg"] = svg_path.read_bytes()

        replacements["word/document.xml"] = etree.tostring(
            document_xml, xml_declaration=True, encoding="UTF-8", standalone=True
        )
        replacements["word/_rels/document.xml.rels"] = etree.tostring(
            relationships_xml, xml_declaration=True, encoding="UTF-8", standalone=True
        )
        replacements["[Content_Types].xml"] = etree.tostring(
            content_types_xml, xml_declaration=True, encoding="UTF-8", standalone=True
        )

        for item in source.infolist():
            if item.filename not in replacements:
                target.writestr(item, source.read(item.filename))
        for name, data in replacements.items():
            if data is not None:
                target.writestr(name, data)
    temporary_docx.replace(docx_path)


def _extract_frontmatter(markdown_text: str) -> tuple[str | None, str]:
    lines = markdown_text.splitlines()
    if not lines or lines[0].strip() != "---":
        return None, markdown_text
    closing_index = next(
        (index for index in range(1, len(lines)) if lines[index].strip() in {"---", "..."}),
        None,
    )
    if closing_index is None:
        return None, markdown_text

    metadata_lines = lines[1:closing_index]
    first_metadata_line = next((line.strip() for line in metadata_lines if line.strip()), "")
    if not re.match(r"^[A-Za-z_][A-Za-z0-9_-]*\s*:", first_metadata_line):
        return None, markdown_text

    frontmatter = "\n".join(metadata_lines)
    title_match = re.search(r"(?m)^title:\s*(.*?)\s*$", frontmatter)
    title = title_match.group(1).strip("\"'") if title_match else None
    body = "\n".join(lines[closing_index + 1 :])
    if body:
        body += "\n"
    return title or None, body


def convert_markdown(
    input_path: Path, output_path: Path, cli_name: str,
    *, font_family: str = "Noto Sans JP", timeout: float = 120.0,
) -> None:
    input_path = input_path.resolve()
    output_path = output_path.resolve()
    if input_path == output_path:
        raise ValueError("Input Markdown and output DOCX paths must be different.")
    if input_path.suffix.lower() not in {".md", ".markdown"}:
        raise ValueError("Input file must use the .md or .markdown extension.")
    if output_path.suffix.lower() != ".docx":
        raise ValueError("Output file must use the .docx extension.")
    if output_path.exists():
        raise FileExistsError(f"Output file already exists; refusing to overwrite it: {output_path}")
    font_family = font_family.strip()
    if not font_family:
        raise ValueError("Font family must not be empty.")
    if timeout <= 0:
        raise ValueError("Timeout must be greater than zero.")

    title, markdown_text = _extract_frontmatter(input_path.read_text(encoding="utf-8-sig"))
    document = Document()
    for style in document.styles:
        if style.type == WD_STYLE_TYPE.PARAGRAPH:
            style.font.name = font_family
            style.element.get_or_add_rPr().rFonts.set(qn("w:eastAsia"), font_family)
    document.core_properties.title = title or input_path.stem
    first_heading = re.search(r"(?m)^#\s+(.+?)\s*$", markdown_text)
    if title and (first_heading is None or first_heading.group(1) != title):
        document.add_heading(title, level=0)
    output_path.parent.mkdir(parents=True, exist_ok=True)

    with tempfile.TemporaryDirectory(
        prefix=".markdown-to-word-mermaid-", dir=output_path.parent
    ) as temporary_directory:
        temp_root = Path(temporary_directory)
        staged_docx = temp_root / "staged-output.docx"
        diagram_count = _convert_blocks(
            document, markdown_text, temp_root, cli_name, font_family, timeout=timeout
        )
        document.save(staged_docx)
        svg_paths = [temp_root / f"diagram-{number}" / "diagram.svg" for number in range(1, diagram_count + 1)]
        _inject_svg_parts(staged_docx, svg_paths, temp_root=temp_root)
        _publish_docx_exclusively(staged_docx, output_path)


def _publish_docx_exclusively(staged_docx: Path, output_path: Path) -> None:
    try:
        os.link(staged_docx, output_path)
    except FileExistsError as error:
        raise FileExistsError(
            f"Output file already exists; refusing to overwrite it: {output_path}"
        ) from error
    except OSError as error:
        if os.name != "nt":
            raise OSError(
                f"Cannot publish DOCX atomically because hard links are unavailable: {output_path}"
            ) from error
        try:
            os.rename(staged_docx, output_path)
        except FileExistsError as conflict:
            raise FileExistsError(
                f"Output file already exists; refusing to overwrite it: {output_path}"
            ) from conflict
    else:
        staged_docx.unlink(missing_ok=True)


def main() -> int:
    parser = argparse.ArgumentParser(
        description="Convert Markdown to DOCX and embed Mermaid diagrams as SVG vectors."
    )
    parser.add_argument("input", type=Path, help="Input Markdown file")
    parser.add_argument("--output", type=Path, help="Output DOCX path")
    parser.add_argument("--mermaid-cli", default="mmdc", help="Mermaid CLI executable (default: mmdc)")
    parser.add_argument("--font-family", default="Noto Sans JP", help="Installed font for text and diagrams (default: Noto Sans JP)")
    parser.add_argument("--timeout", type=float, default=120.0, help="Per-command Mermaid CLI timeout in seconds (default: 120)")
    args = parser.parse_args()
    output_path = args.output or args.input.with_suffix(".docx")

    try:
        convert_markdown(
            args.input,
            output_path,
            args.mermaid_cli,
            font_family=args.font_family,
            timeout=args.timeout,
        )
    except (FileNotFoundError, OSError, RuntimeError, ValueError) as error:
        print(f"Error: {error}", file=sys.stderr)
        return 1
    print(output_path.resolve())
    return 0


if __name__ == "__main__":
    raise SystemExit(main())

Requirements and compatibility

Written for: GitHub Copilot, Claude Code

Works with: Claude Code, Claude apps, OpenAI Codex, GitHub Copilot, Cursor, Gemini CLI, Kiro, OpenCode, Goose, Amp, Roo Code, JetBrains Junie, Factory Droid, Devin (ex-Windsurf), Mistral Vibe, OpenHands, Letta

“Works with” means the tool reads the open Agent Skills format; a skill written for one tool may use features others lack.

Needs: Claude Code frontmatter, Needs Python, Needs Node.js

Security

What our scanner noticed· Low risk

  • Instructions that use the network
  • Scripts that your agent may run

This skill includes scripts that your agent may run. Read them before you use it.

A static scan is not a guarantee. Skills are third-party content: review a skill before you let an agent use it.

Attribution and licence

First committed by
GitHub avatar of furukawa658furukawa658
Publisher
GitHub (Official)
Distributed by
GitHub
Source
github/awesome-copilot · skills/markdown-to-word-mermaid · commit 797337f
Licence
MIT License (redistribution allowed)
  • LICENSE (added by AutoGineer from LICENSE)
  • ATTRIBUTION.md (written by AutoGineer)
Synced
Contributors
furukawa658

AutoGineer is not affiliated with or endorsed by furukawa658 or GitHub.

Download ZIP

The ZIP holds the skill’s files, its licence files and an ATTRIBUTION.md. Download markdown-to-word-mermaid.zip