justfile-expert（正义文件专家）

Justfile Expert

Expert knowledge for Just command runner, recipe development, and task automation with focus on cross-platform compatibility and project standardization.

When to Use This Skill

Core Expertise

Command Runner Mastery

• Justfile syntax and recipe structure
• Cross-platform task automation (Linux, macOS, Windows)
• Parameter handling and argument forwarding
• Module organization for large projects

Recipe Development Excellence

• Recipe patterns for common operations
• Dependency management between recipes
• Shebang recipes for complex logic
• Environment variable integration

Project Standardization

• Golden template with standard naming and section structure
• Self-documenting project operations
• Portable patterns across projects
• Integration with CI/CD pipelines

Recipe Naming Conventions

Semantic Workflow Recipes

Standard composite recipes with defined meanings:

# Composite: code quality only (no tests)
check: format-check lint typecheck

# Pre-commit checks (fast, non-mutating)
pre-commit: format-check lint typecheck test-unit
    @echo "Pre-commit checks passed"

# Full CI simulation
ci: check test-coverage build
    @echo "CI simulation passed"

# Clean build artifacts
clean:
    rm -rf dist build .next

# Clean everything including deps
clean-all: clean
    rm -rf node_modules .venv __pycache__

Key Capabilities

Recipe Parameters

• Required parameters: recipe param: - must be provided
• Default values: recipe param="default": - optional with fallback
• Variadic +: recipe +FILES: - one or more arguments
• **Variadic ***: recipe *FLAGS: - zero or more arguments
• Environment export: recipe $VAR: - parameter as env var

Settings Configuration

• set dotenv-load: Load .env file automatically
• set positional-arguments: Enable $1, $2 syntax
• set export: Export all variables as env vars
• set shell: Custom shell interpreter
• set quiet: Suppress command echoing

Recipe Attributes

• [private]: Hide from --list output
• [no-cd]: Don't change directory
• [no-exit-message]: Suppress exit messages
• [unix] / [windows] / [linux] / [macos]: Platform-specific recipes
• [positional-arguments]: Per-recipe positional args
• [confirm] / [confirm("message")]: Require confirmation before running
• [group: "name"]: Group recipes in --list output
• [working-directory: "path"]: Run in specific directory

Module System

• mod name: Declare submodule
• mod name 'path': Custom module path
• Invocation: just module::recipe or just module recipe

Essential Syntax

Basic Recipe Structure

# Comment describes the recipe
recipe-name:
    command1
    command2

Recipe with Parameters

build target:
    @echo "Building {{target}}..."
    cd {{target}} && make

test *args:
    uv run pytest {{args}}

Recipe Dependencies

default: build test

build: _setup
    cargo build --release

_setup:
    @echo "Setting up..."

Variables and Interpolation

version := "1.0.0"
project := env('PROJECT_NAME', 'default')

info:
    @echo "Project: {{project}} v{{version}}"

Conditional Recipes

[unix]
open:
    xdg-open http://localhost:8080

[windows]
open:
    start http://localhost:8080

Standard Recipes

Every project should provide these standard recipes, organized by section:

# Justfile - Project task runner
# Run `just` or `just help` to see available recipes

set dotenv-load
set positional-arguments

# Default recipe - show help
default:
    @just --list

# Show available recipes with descriptions
help:
    @just --list --unsorted

####################
# Development
####################

# Start development environment
dev:
    # bun run dev / uv run uvicorn app:app --reload / skaffold dev

# Build for production
build:
    # bun run build / cargo build --release / docker build

# Clean build artifacts
clean:
    # rm -rf dist build .next

####################
# Code Quality
####################

# Run linter (read-only)
lint *args:
    # bun run lint / uv run ruff check {{args}}

# Auto-fix lint issues
lint-fix:
    # bun run lint:fix / uv run ruff check --fix .

# Format code (mutating)
format *args:
    # bun run format / uv run ruff format {{args}}

# Check formatting without modifying (non-mutating)
format-check *args:
    # bun run format:check / uv run ruff format --check {{args}}

# Type checking
typecheck:
    # bunx tsc --noEmit / uv run basedpyright

####################
# Testing
####################

# Run all tests
test *args:
    # bun test {{args}} / uv run pytest {{args}}

# Run unit tests only
test-unit *args:
    # bun test --grep unit {{args}} / uv run pytest -m unit {{args}}

####################
# Workflows
####################

# Composite: code quality (no tests)
check: format-check lint typecheck

# Pre-commit checks (fast, non-mutating)
pre-commit: format-check lint typecheck test-unit
    @echo "Pre-commit checks passed"

# Full CI simulation
ci: check test-coverage build
    @echo "CI simulation passed"

Section Structure

Organize recipes into these standard sections:

Use #################### comment blocks as section dividers for readability.

Common Patterns

Setup/Bootstrap Recipe

# Initial project setup
setup:
    #!/usr/bin/env bash
    set -euo pipefail
    echo "Installing dependencies..."
    uv sync
    echo "Setting up pre-commit..."
    pre-commit install
    echo "Done!"

Docker Integration

# Build container image
docker-build tag="latest":
    docker build -t {{project}}:{{tag}} .

# Run container
docker-run tag="latest" *args:
    docker run --rm -it {{project}}:{{tag}} {{args}}

# Push to registry
docker-push tag="latest":
    docker push {{registry}}/{{project}}:{{tag}}

Database Operations

# Run database migrations
db-migrate:
    uv run alembic upgrade head

# Create new migration
db-revision message:
    uv run alembic revision --autogenerate -m "{{message}}"

# Reset database
db-reset:
    uv run alembic downgrade base
    uv run alembic upgrade head

CI/CD Recipes

# Full CI check (lint + test + build)
ci: lint test build
    @echo "CI passed!"

# Release workflow
release version:
    git tag -a "v{{version}}" -m "Release {{version}}"
    git push origin "v{{version}}"

MCP Integration (just-mcp)

The just-mcp MCP server enables AI assistants to discover and execute justfile recipes through the Model Context Protocol, reducing context waste since the AI doesn't need to read the full justfile.

Installation:

# Via npm
npx just-mcp --stdio

# Via pip/uvx
uvx just-mcp --stdio

# Via cargo
cargo install just-mcp

Claude Desktop configuration (.claude/mcp.json):

{
  "mcpServers": {
    "just-mcp": {
      "command": "npx",
      "args": ["-y", "just-mcp", "--stdio"]
    }
  }
}

Available MCP Tools:

• list_recipes - Discover all recipes and parameters
• run_recipe - Execute a recipe with arguments
• get_recipe_info - Get detailed recipe documentation
• validate_justfile - Check for syntax errors

Agentic Optimizations

Best Practices

Recipe Development Workflow

1. Name clearly: Use descriptive, verb-based names (build, test, deploy)
2. Document always: Add comment before each recipe
3. Use defaults: Provide sensible default parameter values
4. Group logically: Organize with section comments
5. Hide internals: Mark helper recipes as [private]
6. Test portability: Verify on all target platforms

Critical Guidelines

• Always provide default recipe pointing to help
• Use @ prefix to suppress command echo when appropriate
• Use shebang recipes for multi-line logic
• Prefer set dotenv-load for configuration
• Use modules for large projects (>20 recipes)
• Include variadic *args for passthrough flexibility
• Quote all variables in shell commands

Comparison with Alternatives

When to use Just:

• Cross-project standard recipes
• Simple, readable task automation
• No tool version management needed

When to use mise tasks:

• Project-specific with tool version pinning
• Already using mise for tool management

When to use Make:

• Legacy projects with existing Makefiles
• Build systems requiring incremental compilation

For the golden justfile template, detailed syntax reference, advanced patterns, and troubleshooting, see REFERENCE.md.