fuzzing-dictionary

The fuzzing-dictionary skill helps improve fuzzing effectiveness by supplying domain-specific tokens that guide fuzzers toward meaningful and structurally valid inputs. Instead of relying entirely on random mutations, fuzzing dictionaries provide known keywords, protocol commands, magic bytes, and format-specific strings that help bypass early validation checks and trigger deeper execution paths. This technique is especially useful when fuzzing parsers, protocol handlers, and file format processors where valid structure or recognizable tokens are required before more complex logic is reached.

The skill supports creating and using dictionary files compatible with common fuzzing tools such as libFuzzer, AFL++, and cargo-fuzz. It explains how to define dictionary entries using quoted strings, key-value pairs, and hexadecimal escape sequences for binary data. The documentation also demonstrates multiple approaches for generating dictionary content, including extracting tokens from header files, binaries, command-line manuals, and large language models. Additional examples cover common patterns such as HTTP protocol keywords and file format magic bytes, making it easier to adapt dictionaries for different fuzzing targets.

This skill is designed for security researchers, fuzzing engineers, software testers, and developers working on parser-heavy or protocol-driven applications. It is particularly valuable for testing structured inputs such as JSON, XML, HTTP, DNS, media codecs, and archive formats. Teams performing coverage-guided fuzzing can use dictionaries to overcome stagnating coverage and improve code path exploration without requiring complex custom mutators.