This is the official Sia wallet app for the Ledger Nano S.
When installed on a Nano S, the app allows you to generate Sia addresses, calculate transaction hashes, sign those hashes, and use those signatures to construct valid transactions. The Sia app is the most secure method currently available for performing these actions.
This code also serves as a walkthrough for writing your own Ledger Nano S app.
The code is heavily commented and describes both the high-level architecture
and low-level implementation details of Nano S app development. Begin by
reading src/main.c, and the comments will tell you which files to read next.
No binaries are provided at this time. To build and install the Sia app on your Ledger Nano S, follow Ledger's setup instructions.
After the app is installed, build the sialedger.go binary to interact with
the device. ./sialedger --help will print a list of commands.
Please refer to our standalone guide for a walkthrough that demonstrates how to generate addresses and sign transactions using the app.
The attack surface for using the Sia wallet app on a Ledger Nano S comprises the Sia app itself, the system firmware running on the Nano S, the computer that the Nano S is connected to, and posession/control of the device. For our purposes, the app only needs to ensure its own correctness and protect the user from the computer that the Nano S is connected to. Other attack surfaces are beyond our control; we assume that the user physically controls the device, is not running malicious/buggy software on the device, and follows proper security protocols. The goal of the Sia app is to achieve perfect security given these assumptions.
The main attack vector that we are concerned with, then, is a computer running
malicious sofware. This software may imitate programs like sialedger in such
a way that the user cannot tell the difference, but secretly act maliciously.
Specifically, the computer can do the following:
- Lie to the user about which actions it is performing. Example: the user
runs
./sialedger addr 1, but the computer secretly runs./sialedger addr 2instead. - Lie to the user about what it received from the Nano S. Example: the Nano S generates address X, but the computer displays address Y.
- Exfiltrate data supplied by the user or the Nano S. Example: the user generates addresses A and B; the computer "phones home" to report that A and B are owned by the same person.
- Refuse to comply with the user's requests. Example: the user runs
./sialedger addr 1, but the computer does nothing.
Clearly the app cannot prevent the computer from performing type-3 or type-4 actions. Fortunately, these actions do not allow the attacker to steal coins; they can only gather metadata and temporarily prevent the user from performing certain actions. Type-1 and type-2 actions, on the other hand, are much more dangerous, and can be trivially exploited to steal the user's coins.
To combat these attacks, apps must make use of the embedded display on the Nano S. If data sent to/from the Nano S is displayed on the screen, the user can verify that the computer is not lying about what it sent or received. In the interest of user-friendliness, we would like to display as little information as much as possible, but each omission brings with it the risk of introducing a vulnerability. Therefore, an app should display all data by default, and omit data only after subjecting the omission to extreme scrutiny. The Sia app adheres to this principle more closely than most Ledger apps, and as a result is not affected by certain vulnerabilities affecting those apps.