Octopus
Security Analysis tool for WebAssembly module (wasm) and Blockchain Smart Contracts (BTC/ETH/NEO/EOS)
Install / Use
/learn @FuzzingLabs/OctopusREADME
Octopus
<p align="center"> <img src="/images/logo-medium.png" height="300px"/> </p>Huge thanks to QuoScient for having sponsored this project.
Octopus is a security analysis framework for WebAssembly module and Blockchain Smart Contract.
The purpose of Octopus is to provide an easy way to analyze closed-source WebAssembly module and smart contracts bytecode to understand deeper their internal behaviours.
Features
- Explorer: Octopus JSON-RPC client implementation to communicate with blockchain platforms
- Disassembler: Octopus can translate bytecode into assembly representation
- Control Flow Analysis: Octopus can generate a Control Flow Graph (CFG)
- Call Flow Analysis: Octopus can generate a Call Flow Graph (function level)
- IR conversion (SSA): Octopus can simplify assembly into Static Single Assignment (SSA) representation
- Symbolic Execution: Octopus use symbolic execution to find new paths into a program
Platforms / Architectures
Octopus support the following types of programs/smart contracts:
- WebAssembly module (WASM)
- Bitcoin script (BTC script)
- Ethereum smart contracts (EVM bytecode & Ewasm)
- EOS smart contracts (WASM)
- NEO smart contracts (AVM bytecode)
|| BTC | ETH (EVM) | ETH (WASM) | EOS | NEO || WASM | |:--------------------:|:---:|:---:|:---:|:---:|:---:|:---:|:---:| | Explorer | :heavy_check_mark: | :heavy_check_mark:| :heavy_check_mark: | :heavy_check_mark: | :heavy_check_mark: | | :o: | |Disassembler | :heavy_check_mark: | :heavy_check_mark: | :heavy_check_mark: | :heavy_check_mark: | :heavy_check_mark: | | :heavy_check_mark: | |Control Flow Analysis | :heavy_multiplication_x: | :heavy_check_mark: | :heavy_check_mark: | :heavy_check_mark: | :heavy_check_mark: | | :heavy_check_mark: | |Call Flow Analysis | :heavy_multiplication_x: | :heavy_plus_sign: | :heavy_check_mark: | :heavy_check_mark: | :heavy_plus_sign: | | :heavy_check_mark: | |IR conversion (SSA) | :heavy_multiplication_x: | :heavy_check_mark: | :heavy_plus_sign: | :heavy_plus_sign: | :heavy_multiplication_x: | | :heavy_check_mark: | |Symbolic Execution | :heavy_multiplication_x: | :heavy_plus_sign: | :heavy_plus_sign: | :heavy_plus_sign: | :heavy_multiplication_x: | | :heavy_plus_sign: |
- PyPI package :heavy_check_mark:
- Docker :heavy_check_mark:
:heavy_check_mark: DONE / :heavy_plus_sign: WIP / :heavy_multiplication_x: TODO / :o: N/A
Requirements
Octopus is supported on Linux (ideally Ubuntu 16.04) and requires Python >=3.5 (ideally 3.6).
Dependencies:
Quick Start
- Install system dependencies
# Install system dependencies
sudo apt-get update && sudo apt-get install python-pip graphviz xdg-utils -y
- Install Octopus:
# Download Octopus
git clone https://github.com/pventuzelo/octopus
cd octopus
# Install Octopus library/CLI and its dependencies
python3 setup.py install
or
# but prefer the first way to install if possible
pip3 install octopus
- Run tests
# Run tests for all platforms (disassembly, CFG, ...)
./run_tests.sh
# Run tests that require internet access (explorer tests)
./run_explorer_tests.sh
# Run tests for only one platforms
# {btc, eth, eos, neo, wasm}_run_tests.sh
cd octopus/tests/
./wasm_run_tests.sh
Docker container
A docker container providing the toolset is available at docker hub. In a terminal, run the following commands:
docker pull smartbugs/octopus
docker run -it smartbugs/octopus
cd octopus
python3 octopus_eth_evm.py -s -f examples/ETH/evm_bytecode/61EDCDf5bb737ADffE5043706e7C5bb1f1a56eEA.bytecode
Command-line tools
- WebAssembly: octopus_wasm.py
- Ethereum (EVM): octopus_eth_evm.py
In-depth Examples using APIs
<details><summary>WebAssembly</summary> <p>Disassembler
Disassembly of a Wasm module:
from octopus.arch.wasm.disassembler import WasmDisassembler
FILE = "examples/wasm/samples/helloworld.wasm"
with open(FILE, 'rb') as f:
module_bytecode = f.read()
disasm = WasmDisassembler()
# return list of functions instructions (list)
print(disasm.disassemble_module(module_bytecode))
#[[<octopus.arch.wasm.instruction.WasmInstruction at 0x7f85e4904278>,<octopus.arch.wasm.instruction.WasmInstruction at 0x7f85e4904f60>,<octopus.arch.wasm.instruction.WasmInstruction at 0x7f85e4904ef0>]]
print()
# return text of functions code
print(disasm.disassemble_module(module_bytecode, r_format='text'))
# func 0
# i32.const 0
# call 0
# end
Disassembly of wasm bytecode:
from octopus.arch.wasm.disassembler import WasmDisassembler
# bytecode in WebAssembly is the function code (i.e. function body)
bytecode = b'\x02\x7fA\x18\x10\x1cA\x00\x0f\x0b'
# create a WasmDisassembler object
disasm = WasmDisassembler(bytecode)
# disassemble bytecode into a list of WasmInstruction
# attributes r_format='list' by default
print(disasm.disassemble())
#[<octopus.arch.wasm.instruction.WasmInstruction object at 0x7f85e4904eb8>, <octopus.arch.wasm.instruction.WasmInstruction object at 0x7f85e4904278>, <octopus.arch.wasm.instruction.WasmInstruction object at 0x7f85e4904390>, <octopus.arch.wasm.instruction.WasmInstruction object at 0x7f85e4904ef0>, <octopus.arch.wasm.instruction.WasmInstruction object at 0x7f85e4904f60>, <octopus.arch.wasm.instruction.WasmInstruction object at 0x7f85e4901048>]
print()
print(disasm.disassemble(r_format='reverse'))
#{0: <octopus.arch.wasm.instruction.WasmInstruction object at 0x7f85e4901048>, 1: <octopus.arch.wasm.instruction.WasmInstruction object at 0x7f85e4904240>, 2: <octopus.arch.wasm.instruction.WasmInstruction object at 0x7f85e4904f60>, 3: <octopus.arch.wasm.instruction.WasmInstruction object at 0x7f85e4904ef0>, 4: <octopus.arch.wasm.instruction.WasmInstruction object at 0x7f85e4904278>, 5: <octopus.arch.wasm.instruction.WasmInstruction object at 0x7f85e4904390>}
print()
print(disasm.disassemble(r_format='text'))
# block -1
# i32.const 24
# call 28
# i32.const 0
# return
# end
ModuleAnalyzer
from octopus.arch.wasm.analyzer import WasmModuleAnalyzer
FILE = "examples/wasm/samples/hello_wasm_studio.wasm"
with open(FILE, 'rb') as f:
module_bytecode = f.read()
# return list of functions instructions (list)
# attributes analysis=True by default
analyzer = WasmModuleAnalyzer(module_bytecode)
# show analyzer attributes
print(analyzer.func_prototypes)
# [('putc_js', 'i32', ''),
# ('__syscall0', 'i32', 'i32'),
# ('__syscall3', 'i32 i32 i32 i32', 'i32'),
# ('__syscall1', 'i32 i32', 'i32'),
# ('__syscall5', 'i32 i32 i32 i32 i32 i32', 'i32'),
# ('__syscall4', 'i32 i32 i32 i32 i32', 'i32'),
# ('$func6', '', ''),
# ('main', '', 'i32'),
# ('writev_c', 'i32 i32 i32', 'i32'),
# ('$func9', '', 'i32'),
# ('$func10', 'i32', 'i32'),
# ('$func11', 'i32', 'i32'),
# ('$func12', 'i32', ''),
# ('$func13', 'i32', 'i32'),
# ('$func14', 'i32 i32 i32 i32', 'i32'),
# ('$func15', 'i32 i32', 'i32'),
# ('$func16', 'i32 i32', 'i32'),
# ('$func17', 'i32', 'i32'),
# ('$func18', 'i32', 'i32'),
# ('$func19', 'i32', 'i32'),
# ('$func20', 'i32 i32 i32', 'i32'),
# ('$func21', 'i32 i32 i32', 'i32'),
# ('$func22', 'i32 i64 i32', 'i64'),
# ('$func23', 'i32 i32 i32', 'i32'),
# ('$func24', 'i32', 'i32'),
# ('$func25', 'i32 i32 i32 i32', '')]
print()
print(analyzer.contains_emscripten_syscalls())
#[('__syscall0', 'restart_syscall'),
# ('__syscall3', 'read'),
# ('__syscall1', 'exit'),
# ('__syscall5', 'open'),
# ('__syscall4', 'write')]
Control Flow Analysis
from octopus.arch.wasm.cfg import WasmCFG
# complete wasm module
file_name = "examples/wasm/samples/fib.wasm"
# read file
with open(file_name, 'rb') as f:
raw = f.read()
# create the cfg
cfg = WasmCFG(raw)
# visualize CFGGraph
# generate graph.dot and graph.pdf file
cfg.visualize()
<p align="center">
<img src="/images/wasm-cfg-fib.png" height="400px"/>
</p>
Functions' instructions analytics
from octopus.arch.wasm.cfg import WasmCFG
# complete wasm module
file_name = "examples/wasm/samples/hello_wasm_studio.wasm"
# read file
with open(file_name, 'rb') as f:
raw = f.read()
# create the cfg
cfg = WasmCFG(raw)
# visualization
cfg.visualize_instrs_per_funcs()
<p align="center">
<img src="/images/wasm-instr-func-analytics.png" height="400px"/>
</p>
Call Flow Analysis
from octopus.arch.wasm.cfg import WasmCFG
# fibonacci wasm module
file_name = "examples/wasm/samples/hello_wasm_studio.wasm"
# read file
with open(file_name, 'rb') as f:
raw = f.read()
# create the cfg
cfg = WasmCFG(raw)
# visualize Call Flow Graph
# generate call_graph.dot and call_graph.pdf file
#
# color similar to https://webassembly.studio/
# imported func == turquoise / exported func == grey
# edge label = number of different calls to the function
cfg.visualize_call_flow()
<p align="center">
<img src="/images/wasm-callflow-hello-studio.png" height="400px"/>
</p>
Legend:
<p align="center"> <img src="/images/legend_callgraph.png" height="400px"/> </p>IR conversion (SSA)
from octopus.arch.wasm.emulator import WasmSSAEmulatorEngine
# complete wasm module
file_name = "examples/wasm/samples/fib.wasm"
# read file
with open(file_name, 'rb') as f:
raw = f.read()
# run the emulator for SSA
emul = WasmSSAEmulatorEngine(raw)
emul.emulate_one_function('fib')
# or emul.emulate_functions(['fib'])
# or e
