Improve markdown metadata workflow

[juansensio]

At this moment, the workflow looks like this:

1. The user creates a README.md with markdown metadata and content
2. The CLI parses de markdown to HTML and send to API
3. The API stores HTML in database
4. The UI retrieves HTML and shows to user
5. In order to edit the metadata 5.1 The UI converts the HTML to markdown for the Markdown editor 5.2 The UI converts markdown to HTML to show updates
6. At update from CLI, the CLI retrieves HTML and converts to Markdown to compare with local data.

We could switch to use Markdown by default everywhere and just convert and render HTML in UI.

This would require to update the database, converting all HTML descriptions to markdown