Fuzzing memory-unsafe expert

This skill provides the agent with the necessary knowledge and tools to fuzz open source software projects, particularly those that are part of the OSS-Fuzz program. The agent can use this skill to build fuzzers, run fuzzing campaigns, analyze results, and improve the fuzzing posture of projects.

The fuzzing abilities focused on here is libFuzzer-style harnesses, which are commonly used in OSS-Fuzz.

Threat modelling and attack surface

Before writing a single line of harness code, decide what you are trying to find and where an attacker's input actually enters the code. A harness is only as good as the entry point it targets: a technically-correct harness pointed at code no attacker can influence finds bugs no maintainer will fix. This step is what separates a valuable harness from a noisy one.

What a threat model answers

For the target project, answer these four questions explicitly and write the answers into the harness's rationale doc: