Truffle uses lerna to manage multi-package repositories. Each Truffle module is defined in its own npm package in the packages/ directory.

The entry point of these modules is @truffle/core. This is where the command line parser is setup.

Install lerna:

$ npm install -g lerna
$ npm install -g yarn

The command flow

The heart of Truffle lies in the @truffle/core package. Whenever a command is run, packages/core/cli.js gets run with everything following truffle (on the command line) being passed in as arguments. In other words, if you run truffle migrate --network myNetwork, then packages/core/cli.js gets run with "migrate" and "--network myNetwork" as arguments.

Throughout the course of running packages/core/cli.js, Truffle parses out what commands and options the user has provided, instantiates an instance of the Command class, and calls the run method on that instance. The run method is the interface that cli.js uses for ALL commands. You can find all of the specific command files (one file for each command) at packages/core/lib/commands. From the run method of each command you should be able to trace the command lifecycle to libraries and modules in @truffle/core as well as other packages in the monorepo.

Add a new command in truffle

Create a new lerna package

$ lerna create mycmd

Add the package to @truffle/core

$ lerna add mycmd --scope=@truffle/core

Create a new command in @truffle/core

1. Create a new directory in packages/core/lib/commands/, let's call it mycmd.
2. Create 3 javascript files inside the mycmd directory with the filenames run.js, meta.js and index.js:
   • run.js contains the entry function after a user calls your command.
   • meta.js contains information such as command name and command description.
   • index.js exports both the run module and the meta module.

run.js

$ cat << EOF > core/lib/commands/mycmd/run.js

module.exports = async function (options){
  const mycmd = require("mycmd");
  // TODO: write the run command here, something like:
  // mycmd(options)
};
EOF

meta.js

$ cat << EOF > core/lib/commands/mycmd/meta.js
module.exports = {
  command: "mycmd",
  description: "Run mycmd",
  builder: {},
  help: {
    usage: "truffle mycmd",
    options: []
  }
EOF

index.js

$ cat << EOF > core/lib/commands/mycmd/index.js
module.exports = {
  run: require("./run"),
  meta: require("./meta")
};
EOF

Link it from the commands/index.js file

--- packages/core/lib/commands/index.js
+++ packages/core/lib/commands/index.js
@@ -1,4 +1,5 @@
 module.exports = {
+  mycmd: require("./mycmd"),

Link it from the commands/commands.js file

--- packages/core/lib/commands/commands.js
+++ packages/core/lib/commands/commands.js
@@ -1,4 +1,5 @@
 module.exports = [
+  "mycmd",

From there, you should see it in the help screen:

$ cd packages/core
$ node cli.js
Truffle v5.0.0-beta.1 - a development framework for Ethereum

Usage: truffle <command> [options]

Commands:
  mycmd     Run mycmd
  build     Execute build pipeline (if configuration present)
[...]

Write your module/command

The setup is done, you can now write your command and organize your module as you want in: packages/mycmd/. You can have a look at packages/box/ which is a good starting example to follow.

Execute the command

$ cd packages/core
$ node cli.js mycmd