sudo

Robert Coggeshall and Cliff Spencer wrote the original subsystem around 1980 at the Department of Computer Science at SUNY/Buffalo.[13] Robert Coggeshall brought sudo with him to the University of Colorado Boulder. Between 1986 and 1993, the code and features were substantially modified by the IT staff of the University of Colorado Boulder Computer Science Department and the College of Engineering and Applied Science, including Todd C. Miller.[13] The current version has been publicly maintained by OpenBSD developer Todd C. Miller since 1994,[13] and has been distributed under an ISC-style license since 1999.[13]

In November 2009 Thomas Claburn, in response to concerns that Microsoft had patented sudo,[14] characterized such suspicions as overblown.[15] The claims were narrowly framed to a particular GUI, rather than to the sudo concept.[16]

The logo is a reference to an xkcd strip.[17][18]

Unlike the command su, users supply their personal password to sudo (if necessary)[19] rather than that of the superuser or other account. This allows authorized users to exercise altered privileges without compromising the secrecy of the other account's password. After authentication, and if the configuration file permits the user access, the system invokes the requested command. sudo retains the user's invocation rights through a grace period (typically 5 minutes) per pseudo terminal, allowing the user to execute several successive commands as the requested user without having to provide a password again.

As a security and auditing feature, sudo may be configured to log each command run. When a user attempts to invoke sudo without being listed in the configuration file, an exception indication is presented to the user indicating that the attempt has been recorded. If configured, the root user will be alerted via mail. By default, an entry is recorded in the system.[20]

The /etc/sudoers file contains a list of users or user groups with permission to execute a subset of commands while having the privileges of the root user or another specified user. The program may be configured to require a password.[21]

In some system distributions, sudo has largely supplanted the default use of a distinct superuser login for administrative tasks, most notably in some Linux distributions as well as Apple's macOS.[22][23] This allows for more secure logging of admin commands and prevents some exploits.

In association with SELinux, sudo can be used to transition between roles in role-based access control (RBAC).[24]

visudo is a command-line utility that allows editing the sudo configuration file in a fail-safe manner. It prevents multiple simultaneous edits with locks and performs sanity and syntax checks.

Sudoedit is a program that symlinks to the sudo binary.[25] When sudo is run via its sudoedit alias, sudo behaves as if the -e flag has been passed and allows users to edit files that require additional privileges to write to.[26]

The program runas provides similar functionality in Microsoft Windows, but it cannot pass current directories, environment variables or long command lines to the child. And while it supports running the child as another user, it does not support simple elevation. A true su and sudo for Windows that can pass all of that state information and start the child either elevated or as another user (or both) is included with Hamilton C shell.[27][28]

Graphical user interfaces exist for sudo – notably gksudo – but are deprecated in Debian and no longer included in Ubuntu.[29][30] Other user interfaces are not directly built on sudo, but provide similar temporary privilege elevation for administrative purposes, such as pkexec in Unix-like operating systems, User Account Control in Microsoft Windows and Mac OS X Authorization Services.[31]

doas, available since OpenBSD 5.8 (October 2015), has been written in order to replace sudo in the OpenBSD base system, with the latter still being made available as a port.[32]

• chroot
• doas
• runas
• Comparison of privilege authorization features

• Official website