nu-lint

A static analysis tool for Nushell scripts based on the official style guide.

Can detect stylistic issues in Nushell scripts and suggest improvements. Functions complementary to the basic checks done by nu-check.

(Inspired by Rust's Clippy)

Installation

From crates.io:

cargo install nu-lint

Or build from source:

cargo install --path .

Nix

Add to your configuration.nix:

let
  nu-lint = pkgs.callPackage (pkgs.fetchFromGitHub {
    owner = "wvhulle";
    repo = "nu-lint";
    rev = "main"; # or specific commit hash
    sha256 = ""; # nix will tell you the correct hash
  }) {};
in
{
  environment.systemPackages = [
    nu-lint
  ];
}

Or in a shell.nix:

{ pkgs ? import <nixpkgs> {} }:

let
  nu-lint = pkgs.callPackage (pkgs.fetchFromGitHub {
    owner = "wvhulle";
    repo = "nu-lint";
    rev = "main";
    sha256 = ""; # nix will tell you the correct hash
  }) {};
in
pkgs.mkShell {
  buildInputs = [
    nu-lint
  ];
}

Usage

nu-lint script.nu                          # Lint a file
nu-lint directory/                         # Lint all .nu files in a directory
nu-lint --config custom.toml script.nu     # Use custom config
nu-lint list-rules                         # Show all rules
nu-lint explain snake_case_variables       # Explain a rule

Configuration

Create .nu-lint.toml in your project root (or any parent directory):

[general]
max_severity = "warning"

[rules]
snake_case_variables = "warning"
prefer_error_make = "info"
kebab_case_commands = "warning"

[style]
line_length = 100
indent_spaces = 4

[exclude]
patterns = ["*.test.nu", "vendor/**"]

The linter will automatically find and use this config file when you run it.

Rules

Categories:

• naming
• formatting
• idioms
• error handling
• code quality
• documentation
• type safety

Example Output

info[prefer_parse_over_each_split]

  ℹ Manual splitting with 'each' and 'split row' - consider using 'parse'
   ╭─[example.nu:5:1]
 5 │ $data | each { |line| $line | split row " " | get 0 }
   ·         ───────────────────┬──────────────────────────
   ·                            ╰── AST-based detection of structured text processing pattern
   ╰────
  help: Use 'parse "{field1} {field2}"' for structured text extraction instead of 'each' with 'split row'

This example demonstrates AST traversal detecting command patterns that the regex-based rules cannot catch.

Contributing

Contributions are welcome. Please run tests and formatting before submitting:

cargo test
cargo +nightly fmt
cargo clippy

License

MIT