Storing HTML in db while avoiding persistent xss/sql injection

I'm building a page in asp.net that will use tiny mce to provide a rich text editor on the page. Tiny mc开发者_开发技巧e outputs the rich text as html which I would like to save to a database. Then at a later date, I want to pull the HTML from the database and display it in a page.

I'm concerned about allowing malicious html, js tags into my database that would later be output.

Can someone walk me through at what point in my process I should html encode/decode etc. to prevent a persistent xss attack and or sql injection attack?

---

We use the Microsoft Web Protection Library to scrape out any potentially dangerous HTML on the way in. What I mean by "on the way in" - when the page is posted to the server, we scrub the HTML using MS WPL and take the results of that and throw that into the database. Don't even let any bad data get to your database, and you'll be safer for it. As far as encoding, you won't want to mess with HTML encoding/decoding - just take whatever is in your tinyMCE control, scrub it, and save it. Then on your display page, just write it out like it exists in your database into a literal control or something like that, and you should be good.

I believe Microsoft.Security.Application.Sanitizer.GetSafeHtmlFragment(input) will do exactly what you want here.

---

Are these admins that are using the RTE? If so, I wouldn't worry about it.

If not, then I don't recommend using a WYSIWYIG such as TinyMCE. You'll have to actually look for malicious input, and chances are, you will miss some. Since the RTE outputs plain HTML, which I assume you want, you can't just convert HTML entities. That would kind of eliminate the whole point of using TinyMCE.

Stopping SQL injection is done in the backend when inserting the data into the database. You will want to use a parametrized query or escape the input (not sure how in ASP.NET, I'm a PHP guy.)

---

Couldn't you use a rich text editor that uses BBCode and on the server, escape everything that needs to be escaped and convert BBCode to HTML markup afterwards?

You could also, instead of producing BBCode on the client, convert the HTML markup to BBCode on the server, escape the remaining HTML and convert the result from BBCode back to HTML.

---

There are two approaches, you will probably use the first one

1) you will make a list of permitted tags and escape/strip rest of them. TinyMCE has probably some feature to disallow user to use some tags..(vut this is only client side, you should validate it on server)

2) you will encode permitted tags differently ([b]bold[/b]), than you could save everything to DB and while rendering escape everything and than interpret your special tags

Third approach: if the user is admin (the one who should know whats is he doing), than you can leave everyhing without escaping...he is the responsible one for his own mistakes....