During my employers annual PCI audit, our auditor requested that we perform a search of all of our servers for credit card data. He recommended a tool called “ccsrch” – which like many open source projects had a couple of issues, and hadn’t been updated in years. So, I fixed it.

CCSRCH is a cross-platform, command-line application that reads every file from the starting point passed in, and scans them for what looks like credit card numbers (and using the Luhn algorithm to check each possible result). It’s fairly brute-force, but it gets the scans required for PCI – though I would be careful about using it during production hours, it can have a pretty significant impact on a server’s I/O performance.

I’ve forked the application and setup a new ccsrch project over at github (the original is on SourceForge), and made a few modifications to better suit my needs (from the change log):

• Added option to output the file name, and how many hits were found to the console when using -o (see -c in usage).
• Added option to limit the number of results from a single file before going on to the next file (see -l in usage).
• Added option to exclude certain file types from the scan (see -n in usage).
• Fix for ignoring NULL, CR & LF.
• Ignore dash when scanning.
• Exclude results with the last 8 digits repeating (very unlikely to be a real PAN).

I’ve uploaded a Windows build of the new 1.0.5 release to github, and for *nix systems, you can just download the latest tag.