Retina, 675 bytes

i`\[code\](.*)\[/code]
==$1==
i`(?<!==.*)\[b](.*)\[/b](?!.*==)
<strong>$1</strong>
i`(?<!==.*)\[i](.*)\[/i](?!.*==)
<em>$1</em>
i`(?<!==.*)\[(.)](.*)\[/\1](?!.*==)
<$1>$2</$1>
i`(?<!==.*)\[color=(.*)](.*)\[/color](?!.*==)
<span style="color:$1">$2</span>
i`(?<!==.*)\[size=(.*)](.*)\[/size](?!.*==)
<span style="font-size:$1">$2</span>
i`(?<!==.*)\[url](.*)\[/url](?!.*==)
[url=$1]$1[/url]
i`(?<!==.*)\[url=(.*)](.*)\[/url](?!.*==)
<a href="$1">$2</a>
i`(?<!==.*)\[img](.*)\[/img](?!.*==)
<img src="$1">
i`(?<!==.*)\[quote=(.*)](.*)\[/quote](?!.*==)
[quote]<cite>$1</cite>$2[/quote]
i`(?<!==.*)\[quote](.*)\[/quote](?!.*==)
<blockquote>$1<blockquote>
==(.*)==
<code>$1</code>

Try it online!

-20 bytes: I discovered by accident that ] matches that character and it doesn't need to be escaped when it doesn't close a character class

a big mess but by the first 5 minutes i already wanted to be done writing regex.

Explanation

• all exressions are case-insensitive
• First, replace all code tags with == which is guaranteed not to appear elsewhere
• I use negative look(ahead|behind) to assert the things i'm matching aren't enclosed in == in all subsequent expressions, properly escaping whatever is in code tags (this is shorter than checking for the full [code] tag everytime)
• i replace b and i tags with their respective html
• i replace the remaining 1-letter tags with the same tag as html
• i replace color and size tags (an optimization here could allow first replacing size with font-size then matching and replacing both but i couldn't find a way where that's shorter)
• I replace urls without an = with one that has
• I replace all urls properly
• i replace all images
• i replace quotes with an = to one without, that just contains the literal <cite> tag
• I replace quote tags properly
• finally, i replace my custom == tags with their contents in a <code> tag.

Retina, 675 bytes

i`\[code](.*)\[/code]
==$1==
i`(?<!==.*)\[b](.*)\[/b](?!.*==)
<strong>$1</strong>
i`(?<!==.*)\[i](.*)\[/i](?!.*==)
<em>$1</em>
i`(?<!==.*)\[(.)](.*)\[/\1](?!.*==)
<$1>$2</$1>
i`(?<!==.*)\[color=(.*)](.*)\[/color](?!.*==)
<span style="color:$1">$2</span>
i`(?<!==.*)\[size=(.*)](.*)\[/size](?!.*==)
<span style="font-size:$1">$2</span>
i`(?<!==.*)\[url](.*)\[/url](?!.*==)
[url=$1]$1[/url]
i`(?<!==.*)\[url=(.*)](.*)\[/url](?!.*==)
<a href="$1">$2</a>
i`(?<!==.*)\[img](.*)\[/img](?!.*==)
<img src="$1">
i`(?<!==.*)\[quote=(.*)](.*)\[/quote](?!.*==)
[quote]<cite>$1</cite>$2[/quote]
i`(?<!==.*)\[quote](.*)\[/quote](?!.*==)
<blockquote>$1</blockquote>
==(.*)==
<code>$1</code>

Try it online!

-20 bytes: I discovered by accident that ] matches that character and it doesn't need to be escaped when it doesn't close a character class

+0 byte: fixing a typo, saved another ]

a big mess but by the first 5 minutes i already wanted to be done writing regex.

Explanation

• all exressions are case-insensitive
• First, replace all code tags with == which is guaranteed not to appear elsewhere
• I use negative look(ahead|behind) to assert the things i'm matching aren't enclosed in == in all subsequent expressions, properly escaping whatever is in code tags (this is shorter than checking for the full [code] tag everytime)
• i replace b and i tags with their respective html
• i replace the remaining 1-letter tags with the same tag as html
• i replace color and size tags (an optimization here could allow first replacing size with font-size then matching and replacing both but i couldn't find a way where that's shorter)
• I replace urls without an = with one that has
• I replace all urls properly
• i replace all images
• i replace quotes with an = to one without, that just contains the literal <cite> tag
• I replace quote tags properly
• finally, i replace my custom == tags with their contents in a <code> tag.

Retina, 695 bytes

i`\[code\](.*)\[/code\]
==$1==
i`(?<!==.*)\[b\](.*)\[/b\](?!.*==)
<strong>$1</strong>
i`(?<!==.*)\[i\](.*)\[/i\](?!.*==)
<em>$1</em>
i`(?<!==.*)\[(.)\](.*)\[/\1\](?!.*==)
<$1>$2</$1>
i`(?<!==.*)\[color=(.*)\](.*)\[/color\](?!.*==)
<span style="color:$1">$2</span>
i`(?<!==.*)\[size=(.*)\](.*)\[/size\](?!.*==)
<span style="font-size:$1">$2</span>
i`(?<!==.*)\[url\](.*)\[/url\](?!.*==)
[url=$1]$1[/url]
i`(?<!==.*)\[url=(.*)\](.*)\[/url\](?!.*==)
<a href="$1">$2</a>
i`(?<!==.*)\[img\](.*)\[/img](?!.*==)
<img src="$1">
i`(?<!==.*)\[quote=(.*)\](.*)\[/quote\](?!.*==)
[quote]<cite>$1</cite>$2[/quote]
i`(?<!==.*)\[quote\](.*)\[/quote\](?!.*==)
<blockquote>$1<blockquote>
==(.*)==
<code>$1</code>

Try it online!

a big mess but by the first 5 minutes i already wanted to be done writing regex.

Explanation

• all exressions are case-insensitive
• First, replace all code tags with == which is guaranteed not to appear elsewhere
• I use negative look(ahead|behind) to assert the things i'm matching aren't enclosed in == in all subsequent expressions, properly escaping whatever is in code tags (this is shorter than checking for the full [code] tag everytime)
• i replace b and i tags with their respective html
• i replace the remaining 1-letter tags with the same tag as html
• i replace color and size tags (an optimization here could allow first replacing size with font-size then matching and replacing both but i couldn't find a way where that's shorter)
• I replace urls without an = with one that has
• I replace all urls properly
• i replace all images
• i replace quotes with an = to one without, that just contains the literal <cite> tag
• I replace quote tags properly
• finally, i replace my custom == tags with their contents in a <code> tag.

Retina, 675 bytes

i`\[code\](.*)\[/code]
==$1==
i`(?<!==.*)\[b](.*)\[/b](?!.*==)
<strong>$1</strong>
i`(?<!==.*)\[i](.*)\[/i](?!.*==)
<em>$1</em>
i`(?<!==.*)\[(.)](.*)\[/\1](?!.*==)
<$1>$2</$1>
i`(?<!==.*)\[color=(.*)](.*)\[/color](?!.*==)
<span style="color:$1">$2</span>
i`(?<!==.*)\[size=(.*)](.*)\[/size](?!.*==)
<span style="font-size:$1">$2</span>
i`(?<!==.*)\[url](.*)\[/url](?!.*==)
[url=$1]$1[/url]
i`(?<!==.*)\[url=(.*)](.*)\[/url](?!.*==)
<a href="$1">$2</a>
i`(?<!==.*)\[img](.*)\[/img](?!.*==)
<img src="$1">
i`(?<!==.*)\[quote=(.*)](.*)\[/quote](?!.*==)
[quote]<cite>$1</cite>$2[/quote]
i`(?<!==.*)\[quote](.*)\[/quote](?!.*==)
<blockquote>$1<blockquote>
==(.*)==
<code>$1</code>

Try it online!

-20 bytes: I discovered by accident that ] matches that character and it doesn't need to be escaped when it doesn't close a character class

a big mess but by the first 5 minutes i already wanted to be done writing regex.

Explanation

• all exressions are case-insensitive
• First, replace all code tags with == which is guaranteed not to appear elsewhere
• I use negative look(ahead|behind) to assert the things i'm matching aren't enclosed in == in all subsequent expressions, properly escaping whatever is in code tags (this is shorter than checking for the full [code] tag everytime)
• i replace b and i tags with their respective html
• i replace the remaining 1-letter tags with the same tag as html
• i replace color and size tags (an optimization here could allow first replacing size with font-size then matching and replacing both but i couldn't find a way where that's shorter)
• I replace urls without an = with one that has
• I replace all urls properly
• i replace all images
• i replace quotes with an = to one without, that just contains the literal <cite> tag
• I replace quote tags properly
• finally, i replace my custom == tags with their contents in a <code> tag.