REFUSR

REFUSR is a tool that uses genetic algorithms to reverse engineer the internal processes of a programmable logic controller given access to its inputs and outputs. It was made under the ReMath DARPA Initiative.

This tool can be used to detect equations and formulaic repetitions of inputs and outputs from PLC controllers. This could catch hidden mistakes in industrial processes before they cause massive and costly recalls, or even lead to injuries and/or deaths.

This project, beyond being immensely useful in thiese industrial error-checking applications, is also an excellent showcase of where SC excels —– existing at an intersection of programming, language theory, and math —– all of which our team is very good at; a more conventional company could not have created this.

AIMEE

AIMEE is a system that Special Circumstances makes for DARPA's AIMEE initiative. It uses reinforcement and machine learning to discover weird machines and uses gadgets to explore the illegal state space revealed through the discovered exploit.

Certain kinds of attacks are the result of mixing instructions and program data in the same memory. AIMEE is an advanced example of a class of tools that automate the finding of these attack methods.

SafeDocs

SafeDocs is a project that addresses security issues within documents, messages, and immages. Specifically, security issues coming from those files potentially carrying untrustworthy commands to underlying systems. It both catches those commands in the act, and can convert the file as a whole to a safer document format.

SafeDocs is both open source and easily usable by personnel who lack an understanding of the project's technical background. These are, obviously, necessities for any tool to which you might want to entrust you or your organization's security.

It was developed in collaboration with DARPA, the PDF Association, BAE Systems, Galois Inc, SRI International, Northrop Grumman Systems Corporation, Lockheed Martin Corporation, Kudu Dynamics, and NASA JPL.

Our part in the project has fallen under the aegis of BAE Systems, with whom we have subcontracted.

Extensive work was done on the project by our Sam Atman, who contributed to it both in his associated work with Bridge/Valorin, and through PARSEC – a project for the improvement of Hammer, a parsing library. Like many modern parsing libraries, it provides a parser combinator interface for writing grammars as inline domain-specific languages, but Hammer also provides a variety of parsing backends. It's also bit-oriented rather than character-oriented, making it ideal for parsing binary data such as images, network packets, audio, and executables. Hammer is written in C, but will provide bindings for other languages. Hammer currently builds under Linux, Windows, and OS X.

Bridge

Bridge is the runtime environment and binary that runs Deck, which is itself used to run Orb, Espalier, and Helm. Orb is a literate programming language driven by Espalier, which is itself a PEG parsing engine. Both are important for the functioning of Helm, which is the REPL and command center which functions as the live interaction environment for Orb. Innovatively, one can save Helm sessions as unit tests. Bridge, Deck, Orb, and Helm were all developed by our Sam Atman.