Frida: Project History

Origins

Frida emerged from brainstorming sessions between @oleavr and @hsorbo about making reverse-engineering more interactive and productive.

Core Component Development

frida-gum

• Generic, cross-platform code-instrumentation library written in C
• Created by @oleavr; originally limited to function hooking and memory analysis (unit-test aids for memory leaks and profiling)
• Later expanded into the instrumentation core of Frida
• Repository: https://github.com/frida/frida-gum

frida-core

• Handles injection of shared libraries into arbitrary target processes
• Maintains a live two-way communication channel with injected code
• Repository: https://github.com/frida/frida-core

Scripting Runtime

• frida-gum embeds QuickJS as the JavaScript engine inside the injected payload

Stalker (Code Tracing Engine)

• Dynamic recompilation-based code tracing engine inside frida-gum
• Significantly improved through pair-programming hackathons between @oleavr and @karltk
• x86 implementation: https://github.com/frida/frida-gum/blob/master/gum/backend-x86/gumstalker-x86.c
• Early development blog post: http://blog.kalleberg.org/post/833101026/live-x86-code-instrumentation-with-frida

Language Bindings

Summary

Frida is a toolbox for dynamic instrumentation and reverse-engineering. frida-core injects frida-gum into target processes; frida-gum hooks functions and runs a QuickJS-based scripting runtime.