AFL++ is an advanced, high-performance version of American Fuzzy Lop (AFL), a coverage-guided fuzzer used to find bugs and security vulnerabilities in software.

Does this skill work with source code or binaries?

It supports both, allowing for instrumentation-based fuzzing of source code and black-box fuzzing of compiled binaries using specialized modes like QEMU or FRIDA.

How does this integrate with Claude Code?

The skill provides Claude with the domain-specific knowledge required to configure targets, initiate fuzzing sessions, and analyze resulting crashes within your development environment.

Which programming languages benefit most from this skill?

AFL++ is most effective for compiled languages like C, C++, Rust, and Go, though it can be applied to any binary target.

Who created the AFL++ skill?

This skill was developed by Trail of Bits, a premier cybersecurity research and consulting firm, as part of their security testing handbook.

AFL++ Fuzzing

Name: AFL++ Fuzzing
Author: plurigrid

byplurigrid

•

セキュリティとテスト

Performs comprehensive binary and source code fuzzing using the AFL++ framework to identify software vulnerabilities.

The AFL++ skill integrates the industry-leading American Fuzzy Lop plus plus framework into Claude Code, enabling developers to conduct advanced coverage-guided security testing. It facilitates the discovery of memory corruption, logic errors, and complex crashes by automatically generating and mutating test inputs for both source code and compiled binaries. This skill is particularly valuable for security researchers and developers looking to harden critical infrastructure, network protocols, and file parsers through automated vulnerability discovery.