Ouroboros — A Symbolic-Execution Decompiler

A fast, research-friendly decompiler built in Rust that recovers high-level structure from binaries using symbolic execution, constraint tracking instead of single static assignment IR models (though it can ingest SSA IR). It emphasizes block pre/post conditions, register/memory deltas, dominator/post-dominator analysis, and structured reconstruction (if/else, loops, SESE regions).

---

Features

• Symbolic Execution Core Tracks path constraints and symbolic values to recover call arguments, globals, and side effects.

• Constraint-Tracking SSA IR VEX-inspired, def-use chains, register/memory deltas, and expression rewriting for clean lifting.

• CFG + Structural Recovery Dominators, post-dominators, and SESE discovery to form if/else, loops, and early-exit patterns.

• Calling Convention Inference Heuristics + dataflow to infer parameter/return passing across common ABIs.

• Rust-first Performance & Safety Zero-cost abstractions where possible; explicit, testable passes.

• Beautiful UI Thanks and a big shoutout to

  • egui for this beautiful work.
  • egui_dock for managing dockable windows like a boss.
  • egui_graphs for CFG drawing

---

Quick Start

Prerequisites

• Rust (stable). Install via https://rustup.rs/
• Recommended: llvm-objdump, gdb, or your favorite disassembler for comparison.

Build

git clone https://github.com/Hexorg/Ouroboros.git
cd Ouroboros
cargo build --release

Run

cargo run --release

Upon start, Use the menu bar and select File > Open... Then select a PE32 (or PE32+) binary. Currently, no automated analysis is available. Navigate to a know function start address, hover over the first byte of the instruction and press D to disassemble the instructions until NOP or an invalid instruction is encountered. Then Hover over the first address of the function and press F to decompile the function. This will symbolically execute the function and present decompiled result.

License

Ouroboros is free, open source and permissively licensed! You can choose from

• MIT License (LICENSE-MIT or http://opensource.org/licenses/MIT)
• Apache License, Version 2.0 (LICENSE-APACHE or http://www.apache.org/licenses/LICENSE-2.0)

at your option. This means you can select the license you prefer!