Mythril

Mythril is a security analysis tool for Ethereum smart contracts. It uses the LASER-ethereum symbolic virtual machine to detect various types of issues. Use it to analyze source code or as a nmap-style black-box blockchain scanner (an "ethermap" if you will).

• Installation and setup
• Security analysis
  • Analyzing Solidity code
    • Specifying Solc versions
    • Output formats
    • Analyzing a Truffle project
  • Analyzing on-chain contracts
  • Speed vs. Coverage
• Control flow graph
• Blockchain exploration
  • Searching from the command line
  • Reading contract storage
• Utilities
  • Disassembler
  • Calculating function hashes
  • Function signatures
• Credit

Installation and setup

Build the Docker image:

$ git clone https://github.com/ConsenSys/mythril/
$ docker build mythril

Install from Pypi:

$ pip3 install mythril

If you plan to analyze Solidity code you'll also need the native version of solc. Solcjs is not supported.

Running tests

python version

First, make sure your python version is 3.6.x. Some tests will fail with 3.5.x since some generated easm code is different from 3.6.x.

truffle

In the tests, we tested the command --truffle, which required the truffle command is installed.

npm install -g truffle

geth

In order to run tests and coverage reports, you need to run geth locally, since some tests depend on it.

Install geth from here: https://github.com/ethereum/go-ethereum/wiki/Building-Ethereum

Then you can run geth version and if you see Version: 1.8.2-stable or above, it's OK for testing.

Don't forget to run geth account new to generate an account for you if this is the first time you use it.

Then start it like this:

geth --syncmode full --rpc --shh --debug

We use --syncmode full here because the eth.blockNumber will get increased soon in this mode, which is useful in tests.

If there is no error thrown, you can wait 1 or 2 minutes before running tests.

Run the tests

pip3 install -r requirements.txt
./all_tests.sh

It may cost you about 3 minutes to run all the tests.

Generating test coverage report

./coverage_report.sh

It will generate a coverage testing report coverage_html_report/index.html, which will be automatically opened in browser. You can find coverage rate and tested/missing code from the report.

Notice there are some tests are running by shell commands(tests/cmd_line_test.py), not calling by python, so they are not included in the coverage analysis.

It may cost you about 5 minutes to generate the report.

Security analysis

Run myth -x with one of the input options described below to run the analysis. This will run the Python modules in the /analysis/modules directory.

Mythril detects a range of security issues, including integer underflows, owner-overwrite-to-Ether-withdrawal, and others. However, the analysis will not detect business logic issues and is not equivalent to formal verification.

Analyzing Solidity code

In order to work with Solidity source code files, the solc command line compiler needs to be installed and in path. You can then provide the source file(s) as positional arguments, e.g.:

$ myth -x solidity_examples/ether_send.sol 
==== Ether send ====
Type: Warning
Contract: Crowdfunding
Function name: withdrawfunds()
PC address: 816
In the function 'withdrawfunds()' a non-zero amount of Ether is sent to msg.sender.

There is a check on storage index 7. This storage slot can be written to by calling the function 'crowdfunding()'.
--------------------
In file: solidity_examples/ether_send.sol:18

msg.sender.transfer(this.balance)

If an input file contains multiple contract definitions, Mythril analyzes the last bytecode output produced by solc. You can override this by specifying the contract name explicitly:

$ myth -x OmiseGo.sol:OMGToken

Specifying Solc versions

You can specify a version of the solidity compiler to be used with --solv <version number>. Please be aware that this uses py-solc and will only work on Linux and OS X versions of Mavericks, Yosemite and El Capitan as of the time of this writing. It will check you locally installed compiler, if this is not what is specified, it will download binaries on Linux or try to compile from source on OS X.

Output formats

By default, analysis results are printed to the terminal in text format. You can change the output format with the -o argument:

$ myth -xo json underflow.sol

The json format is useful for integration into other tools, while -o markdown creates a human-readable report.

Analyzing a Truffle project

Truffle Suite is a popular development framework for Ethereum. To analyze the smart contracts in a Truffle project, change in the project root directory and make run truffle compile followed by myth --truffle.

Analyzing on-chain contracts

When analyzing contracts on the blockchain, Mythril will by default query a local node via RPC. You can also use the built-in INFURA support. Alternatively, you can override the RPC settings with the --rpc argument.

Argument	Description
None	Connect to local Ethereum node
-i	Connect to INFURA Mainnet via HTTPS
--rpc ganache	Connect to local Ganache
--rpc infura-[netname]	Connect to infura-mainnet, rinkeby, kovan or ropsten
--rpc HOST:PORT	Custom RPC connection Custom
--rpctls <True/False>	RPC connection over TLS (default: False)
--ipc	Connect to local Ethereum node via IPC

To analyze a mainnet contract via local RPC:

$ myth -xa 0x5c436ff914c458983414019195e0f4ecbef9e6dd

Or, using INFURA instead:

$ myth -xia 0x5c436ff914c458983414019195e0f4ecbef9e6dd

Adding the -l flag will cause Mythril to automatically retrieve dependencies, such as dynamically linked library contracts:

$ myth -xia 0xEbFD99838cb0c132016B9E117563CB41f2B02264 -l -v1

Speed vs. Coverage

The maximum recursion depth for the symbolic execution engine can be controlled with the --max-depth argument. The default value is 12. Lowering this value reduces the analysis time as well as the coverage / number of explored states.

$ myth -xia 0x5c436ff914c458983414019195e0f4ecbef9e6dd --max-depth 8

Control flow graph

The -g FILENAME option generates an interactive jsViz graph:

$ myth -ig ./graph.html -a 0x5c436ff914c458983414019195e0f4ecbef9e6dd --max-depth 8

The "bounce" effect, while awesome (and thus enabled by default), sometimes messes up the graph layout. Try adding the --enable-physics flag for a very entertaining "bounce" effect that unfortunately completely destroys usability.

Statespace JSON for Traceview Explorer

The -j FILENAME option dumps the statespace to json in the format that is required by the Symbolic Trace Explorer GUI.

$ ./myth -ij ./statespace.json -a 0x5c436ff914c458983414019195e0f4ecbef9e6dd --max-depth 8

Blockchain exploration

If you are planning to do batch operations or use the contract search features, running a go-ethereum node is recommended. Start your local node as follows:

$ geth --syncmode fast --rpc

Mythril builds its own contract database to enable fast search operations. This enables operations like those described in the legendary "Mitch Brenner" blog post in seconds minutes instead of days. Unfortunately, the initial sync process is slow. You don't need to sync the whole blockchain right away though: If you abort the syncing process with ctrl+c, it will be auto-resumed the next time you run the --init-db command.

$ myth --init-db
Starting synchronization from latest block: 4323706
Processing block 4323000, 3 individual contracts in database
(...)

Note that only contracts with non-zero balance are added to the database.

If you experience syncing errors on Mac OS High Sierra, run the following command before starting the sync:

export OBJC_DISABLE_INITIALIZE_FORK_SAFETY=YES

Searching from the command line

The search feature allows you to find contract instances that contain specific function calls and opcode sequences. It supports simple boolean expressions, such as:

$ myth --search "func#changeMultisig(address)#"
$ myth --search "code#PUSH1 0x50,POP#"
$ myth --search "func#changeMultisig(address)# and code#PUSH1 0x50#"

Reading contract storage

You can read the contents of storage slots from a deployed contract as follows.

$ myth --storage 0,1 -a "0x76799f77587738bfeef09452df215b63d2cfb08a"
0x0000000000000000000000000000000000000000000000000000000000000003

Utilities

Disassembler

Use the -d flag to disassemble code. The disassembler accepts a bytecode string or a contract address as its input.

$ myth -d -c "0x6060"
0 PUSH1 0x60

Specifying an address via -a ADDRESS will download the contract code from your node.

$ myth -d -a "0x2a0c0dbecc7e4d658f48e01e3fa353f44050c208"
0 PUSH1 0x60
2 PUSH1 0x40
4 MSTORE
(...)
1135 - FUNCTION safeAdd(uint256,uint256) -
1136 CALLVALUE
1137 ISZERO

Calculating function hashes

To print the Keccak hash for a given function signature:

$ myth --hash "setOwner(address)"
0x13af4035

Function signatures

Whenever you disassemble or analyze binary code, Mythril will try to resolve function names using its local signature database. The database must be provided at ~/.mythril/signatures.json. You can start out with the default file as follows:

$ mkdir ~/.mythril
$ cd ~/.mythril
$ wget https://raw.githubusercontent.com/b-mueller/mythril/master/signatures.json

When you analyze Solidity code, new function signatures are added to the database automatically.

Credit

• JSON RPC library is adapted from ethjsonrpc (it doesn't seem to be maintained anymore, and I needed to make some changes to it).

• The signature data in signatures.json was initially obtained from the Ethereum Function Signature Database.

• Many features, bugfixes and analysis modules have been added by contributors.