Changelog

v0.3 - Current

Breaking Changes

• Ollama action removed - Replaced with preset: ollama
• Old syntax: ollama: {state: present, ...}
• New syntax: preset: {name: ollama, with: {state: present, ...}}
• Migration: Update your configs to use preset syntax
• Benefit: 81% code reduction, user-extensible workflows

Preset System

• Preset action - Reusable, parameterized workflows
• Package complex workflows as YAML files
• Type-safe parameters with validation (string, bool, array, object)
• Parameter namespacing prevents collisions
• Discovery paths: ./presets/, ~/.mooncake/presets/, system paths
• Support for both flat (name.yml) and directory (name/preset.yml) formats
• Full integration with planner (includes, loops, conditionals work in presets)
• Aggregate result tracking (changed = any step changed)
• Ollama preset - AI runtime management
• Replaces 1,400 lines of Go code with 250 lines of YAML
• Multi-platform installation (apt, dnf, yum, brew, script fallback)
• Service configuration with environment variables
• Model management (pull, force pull)
• Uninstallation with optional model cleanup
• Preset authoring - User extensibility
• Create custom presets without Go knowledge
• Share presets as files (git, package managers, direct distribution)
• Template rendering in preset steps
• Platform-aware logic using system facts

New Actions

• Print action - Simple output to console
• Display messages during execution
• Variable interpolation support
• Useful for debugging and user feedback
• Package action - Package management
• Cross-platform package installation
• Support for apt, dnf, yum, brew, and more
• State management (present/absent/latest)
• Idempotent operations
• Service management (systemd on Linux, launchd on macOS)
• Full lifecycle control (start, stop, restart, reload)
• Enable/disable services on boot
• Unit/plist file management with templates
• Drop-in configuration files (systemd)
• Idempotent operations with change detection
• Assert action for state verification
• Command assertions (exit code verification)
• File assertions (existence, content, permissions, ownership)
• HTTP assertions (status codes, response body validation)
• Always returns changed: false (verify without modifying)
• Fail-fast behavior with detailed error messages
• Download action - Download files from URLs
• Checksum verification (SHA256, MD5)
• Idempotent downloads (skip if exists, verify checksums)
• Timeout support
• Unarchive action - Extract archives
• Support for tar, tar.gz, zip formats
• Automatic format detection
• Idempotent extraction
• Permission preservation
• Command action - Structured command execution
• Multiple interpreter support (bash, sh, pwsh, cmd)
• Stdin input support
• Output capture control
• Copy action - Advanced file copying
• Recursive directory copying
• Content filtering with glob patterns
• Backup creation before overwrite
• Checksum-based change detection
• Permission and ownership preservation
• Enhanced file action - Extended file management
• Multiple states: file, directory, absent, touch, link, hardlink, perms
• Ownership management (owner/group)
• Permission management with recursive option
• Backup creation before modifications
• Symbolic and hard link creation
• Safety checks (prevent accidental root deletion)

Planning & Execution

• Deterministic plan compiler
• Expand all loops and includes into linear plan
• Origin tracking (file:line:col) for every step
• Export plans as JSON/YAML
• Execute from saved plans (--from-plan)
• Plan inspection with mooncake plan command
• Tag filtering at plan time
• Artifacts system - Persist execution data
• Run results and summaries
• Event stream capture
• Full stdout/stderr capture (--capture-full-output)
• Unique run IDs with timestamps
• Structured artifact directory (.mooncake/runs/<runid>/)
• Event system - Real-time observability
• Structured event emission (JSON)
• Full lifecycle events (run start, step execution, completion)
• Integration-friendly for monitoring and auditing
• JSON output format for CI/CD (--output-format json)

System Facts

• Comprehensive system facts collection
• CPU: model, cores, instruction set flags (AVX, AVX2, SSE4_2, FMA, AES)
• Memory: total/free, swap total/free
• GPU detection: NVIDIA (nvidia-smi), AMD (rocm-smi), Intel/Apple
• Disk information: mounts, filesystem type, size/used/available
• Network: interfaces with MAC/IP, default gateway, DNS servers
• Toolchain versions: Docker, Git, Go, Python
• Package manager detection (apt, dnf, yum, brew, pacman, zypper, apk, port)
• Distribution info: name, version, kernel
• mooncake facts command with text/JSON output

Execution Control & Flow

• Timeout support - Prevent long-running commands
• Per-step timeout configuration
• Standard exit code (124) on timeout
• Retry logic - Automatic retry on failure
• Configurable retry count and delay
• Works with shell, command, download actions
• Idempotency controls
• creates - Skip step if path exists
• unless - Skip step if command succeeds
• Custom result evaluation
• changed_when - Override change detection
• failed_when - Define custom failure conditions

Template & Expression Language

• Expression language (expr-lang) - Powerful expression evaluation
• Used in when, changed_when, failed_when, with_items
• Rich operators and functions
• Type-safe evaluation
• New template functions
• has() - Check if map contains key
• All facts available as template variables

Security & Privilege Escalation

• Improved sudo handling
• Interactive password prompt (-K, --ask-become-pass) - recommended
• File-based password (--sudo-pass-file) with 0600 permission check
• CLI password with explicit opt-in (--insecure-sudo-pass)
• Better error messages for privilege escalation failures

User Interface & Output

• Animated TUI with progress tracking
• Real-time step status updates
• Spinner animations
• Color-coded results
• Optional raw mode (--raw) for CI/CD
• Dry-run mode for safe preview
• Preview all changes without execution
• Shows what would be changed
• Template rendering validation
• Execution time tracking
• Per-step duration
• Total run time
• Performance metrics in summary

Code Quality & Reliability

• Critical bug fixes - Resolved 5 critical runtime issues
• Fixed nil pointer dereference in HandleVars (prevented production crashes)
• Fixed panic on type assertion in config reader (improved error handling)
• Fixed division by zero in progress bar (TUI stability)
• Fixed error shadowing in defer closures (proper error propagation)
• Fixed shallow copy in Clone() method (prevented shared state bugs)
• High priority improvements - Enhanced 5 incomplete features
• Implemented SetData method for custom result data
• Fixed/documented expandUser template filter
• Completed expression evaluation in planner
• Improved Windows facts collection
• Enhanced git fetching for preset registry
• Performance optimizations - Reduced overhead
• Added logging for dropped events (improved debugging)
• Implemented regex compilation caching (faster expression evaluation)
• Fixed string redaction overlap handling (security improvement)
• Code stability - Reduced panic risk
• Removed panics from handler registration (graceful errors)
• Removed panics from template renderer initialization (better error handling)

Documentation & Architecture

• Architecture Decision Records (ADRs) - Documented key design decisions
• ADR-001: Handler-Based Action Architecture
• ADR-002: Preset Expansion System
• ADR-003: Planner and Execution Model
• AI Specification - Guidelines for AI agents using mooncake
• Safety guidelines and best practices
• Action risk assessment matrix
• Idempotency patterns and examples
• Comprehensive guides
• Using Presets guide (600+ lines)
• Preset Authoring guide (800+ lines)
• Quick-start examples for all major features
• Documentation improvements
• Restructured navigation with feature-first approach
• Professional documentation style (removed emojis)
• Streamlined quick-start guide with copyable examples
• API documentation for programmatic usage
• Positioned as "The Standard Runtime for AI System Configuration"

Developer Experience

• Handler-based action architecture - Modular action system
• Each action self-contained in one file (100-1000 lines)
• Registry pattern for automatic action discovery
• No dispatcher updates needed for new actions
• Net reduction of ~16,000 lines of code
• Testing infrastructure
• Docker-based testing for multiple Linux distributions
• Comprehensive test suite covering all actions and utilities
• Integration tests for sudo, file operations, and system interactions
• Race detector enabled in CI
• Test coverage improvements across the codebase
• Code quality tooling
• Security scanning (gosec) with zero issues
• Vulnerability checking (govulncheck)
• Linter integration with automatic fixes
• Custom error types with structured error handling
• Build and release automation
• Multi-platform builds (Linux, macOS, Windows)
• Automated GitHub releases
• Version management and changelog generation
• Configuration validation
• JSON Schema-based validation with detailed error messages
• Template syntax validation at parse time
• Type checking for all action parameters
• Clear error messages with file:line:col references
• CLI improvements
• mooncake explain command for system fact inspection
• mooncake validate command for config validation
• mooncake plan command for plan inspection
• Better help text and usage examples
• Improved flag organization and defaults

v0.2

Features

• Loop iteration
• with_items - Iterate over lists/arrays
• with_filetree - Recursively iterate over directory contents
• Loop variables: item, item.index, item.first, item.last
• Tag filtering - Filter execution by tags with --tags
• Register - Capture step output for reuse
• Capture stdout, stderr, rc (exit code), changed, failed, skipped
• Use in subsequent steps via variables
• Works with all action types
• Expression language migration - Migrated from govaluate to expr-lang
• Nested field access (e.g., result.stdout)
• Built-in functions: len(), contains(), string operations
• Better error messages and type safety

v0.1 - Initial Release

Features

• Shell command execution - Run shell commands
• File and directory operations - Basic file/directory creation and management
• Template rendering - Pongo2 template engine with variable substitution
• Variables - Define and use variables in configurations
• Basic system facts - OS, architecture, hostname, user home
• Conditionals - Execute steps conditionally with when
• Include files - Split configurations across multiple files
• Include variables - Load variables from separate YAML files
• Sudo/privilege escalation - Run steps with elevated privileges using become