You can use the Internet Archive's Wayback Machine to retrieve old Craigslist pages.
For example, capture dates:
http://web.archive.org/web/*/http://sfbay.craigslist.org/search/ats
And one archived page:
http://web.archive.org/web/20150113174952/http://sfbay.craigslist.org/search/ats
All of these links go to archived listings. For example:
http://web.archive.org/web/20150113185626/http://sfbay.craigslist.org/sby/ats/4841850048.html
Instead of screen-scraping, there is an API, which makes the process of finding archive dates for individual pages.
This simple API for Wayback is a test to see if a given url is archived and currenlty accessible in the Wayback Machine.
Here is a small python script to access the API.
The good thing about Craigslist is that one you give the host-name (i.e. sfbay.craigslist.org), the rest of the URL structure should be similar. You can also scrape the main page to get all the host-names (sfbay, newyork, etc).
You'll find a complete list of categories sites on the CL bulk loading page.
Other ideas:
For other ideas, take a look at these questions from StackOverflow