This is a environment for measuring and observing the behaviors of fuzzers from inside the fuzzer itself.
This allows for comparing fuzzers without having to worry about them having different definitions of performance and coverage.
This tool provides real-time visualizations of coverage as it happens, so you can not only see what was hit, but what is being hit frequently!
It's important to note, while this tool currently mainly does visualizations, we'll likely start working on analyizing results statistically rather than eyeballing images!
To use this, simply run cargo run in prog_gen to create a random program.
This random program expects a filename as the argument once compiled which is
then used as the fuzz input for the randomly generated program.
This program then reports status via shared memory to a file which can be
observed and monitored in a web browser by using coverage_server. Run cargo run in coverage server and it will monitor the shared memory for changes on an
interval, and when new changes are observed the information will be pushed over
a websocket to a website which will visualize the coverage in realtime via
graphviz SVG.
Run cargo run in prog_gen to generate a random program.
Run cargo run in coverage_server to monitor coverage events.
Run python3 -m http.server in coverage_server to host the coverage viewer
website, and navigate to the locally hosted website.