Under construction: we're in the process of extracting this from our toolshed repository; please pardon the flaws.

Ruckus is a:

A tool used in security testing to generate pathological inputs for target code. Two common use cases:

• Generating malicious protocol messages to attack network software

• Creating malicious files in specific file formats to feed to target programs

I'm stealing Mike Eddington's term; Smart Fuzzers distinguish themselves from "just plain fuzzers" by being aware of the data format they're being used to test. In both Eddington's Peach Fuzzer and Ruckus, you accomplish that by defining data models (structures) to describe protocols and file formats.

Peach Fuzzer is written in Python. So is Sully, Pedram Amini's fuzzer. SPIKE is written in C. Ruckus is Ruby's answer, and it tries to play to Ruby's strengths:

• It's much more DSL-y than Peach Fuzzer or Sully

• Unlike XML-bound Peach Fuzzer, it's "configuration files" are code

• You don't really need to know Ruby to write those files

Long term I'm hoping Ruckus bears the same relationship to Ruby as Expect did to Tcl.

Ruckus separates metadata and actual content, and structures packets and file formats as trees of nodes, each with classes and (when desired) DOM-style id's (we call them "tags"). Ruckus data can be manipulated with tree traversal and "mutated" with cascading style sheet selector-type logic.