| name | md2html |
| description | DEPRECATED — moved to mcp-server-document as md2html_convert tool |
| allowed-tools | Bash(python *), Read |
md2html — moved to mcp-server-document
This skill has been promoted to a tool in mcp-server-document on
2026-05-03 so it is reachable from any MCP-aware client without going
through the slash-command skill mechanism.
New invocation
Through MCP (preferred):
- Server:
mcp-server-document
- Tool:
md2html_convert(md_file, output_file=None, title=None)
The tool returns a status string with output path, math-block count,
embedded image count, and per-image embedding warnings.
Source: S:\mcp-server\src\mcp_server_document\md2html\
tools.py — MCP tool wrapper (md2html_convert)
converter.py — pure conversion core (md_to_html)
Legacy CLI (still works)
The original standalone script remains at this folder and can be run
directly when MCP is unavailable:
python md2html.py <input.md> [<output.html>] [<title>]
Why moved
- Single MCP server (
mcp-server-document) already handles
paper_writing / presentation / diagram / ocr / etc., and md2html is
a natural fit for the document family.
- MCP tools are usable from any client (Codex Desktop, Codex,
custom integrations); skills are Codex only.
- Pure-function
converter.md_to_html is now importable as a regular
Python API for other tools.