Simple Web Proxy

[Mark Charsley]

I'd like a simple web-proxy that 
- intercepts each HTML page,  
- passes it through a list of python functions which can, for instance, 
search for and destroy for links to images on doubleclick.net, 
particularly offensive Javascript functions,  etc.
- pass the modified page back to the client.

A couple of questions:
1) does this make sense, or am I suffering from a particularly simplistic 
view of how HTTP works (I'd be shocked if it worked with HTTPS for 
instance)
2) is there an existing skeleton out there I could build upon (capable of 
running on 9x or linux with apache: doesn't matter which)

TIA

Mark