Skip to main content
fyrer’s configuration is organized around two core concepts: packages — the subdirectories that make up your monorepo — and tasks — the named commands to run inside each package. Everything in fyrer.yml flows from this two-level structure.

​
Packages

A package is a directory in your monorepo with a unique name and a root path. It acts as the scope for a group of related tasks, environment variables, and an optional .env file. Key rules:
  • Each package must have a name that is unique across the entire config.
  • root must be a relative path and must exist at the time fyrer is invoked.
  • Environment variables declared at the package level are inherited by every task in that package.
  • An optional env_file (relative to root) is parsed and merged into the environment before task-level variables are applied.
Here is a minimal package declaration:

​
Tasks

A task is a named command inside a package — think build, dev, lint, or test. Tasks are defined as a map under the package’s tasks key. Key rules:
  • cmd is required. It is run via sh -c on Unix or cmd /C on Windows, so any shell syntax is valid.
  • Task names must be unique within a package (the same name can appear in different packages).
  • Tasks can declare inputs, outputs, depends_on, cache, persistent, watch, timeout, cwd, and per-task env / env_file. Note: watch is reserved for a future release — the flag is accepted in config but automatic file-watching is not yet active.
  • Per-task environment variables take the highest precedence in the environment resolution order.

​
Task specifiers

When running fyrer run or fyrer plan, you tell fyrer which tasks to execute using a task specifier. Three forms are supported: For example:
Package names and task names cannot contain spaces. Use simple identifiers such as api, web, shared-ui, or build — anything you’d be comfortable using as a map key in YAML.