puaros/packages/guardian/CHANGELOG.md

Changelog

All notable changes to @samiyev/guardian will be documented in this file.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

[0.7.5] - 2025-11-25

Changed

• ♻️ Refactored AnalyzeProject use-case - improved maintainability and testability:
  • Split 615-line God Use-Case into focused pipeline components
  • Created FileCollectionStep.ts for file scanning and basic parsing (66 lines)
  • Created ParsingStep.ts for AST parsing and dependency graph construction (51 lines)
  • Created DetectionPipeline.ts for running all 7 detectors (371 lines)
  • Created ResultAggregator.ts for building response DTO (81 lines)
  • Reduced AnalyzeProject.ts from 615 to 245 lines (60% reduction)
  • All 345 tests pass, no breaking changes
  • Improved separation of concerns and single responsibility
  • Easier to test and modify individual pipeline steps

Added

• 🤖 AI Agent Instructions in CLI help - dedicated section for AI coding assistants:
  • Step-by-step workflow: scan → fix → verify → expand scope
  • Recommended commands for each step (--only-critical --limit 5)
  • Output format description for easy parsing
  • Priority order guidance (CRITICAL → HIGH → MEDIUM → LOW)
  • Helps Claude, Copilot, Cursor, and other AI agents immediately take action

Run guardian --help to see the new "AI AGENT INSTRUCTIONS" section.

[0.7.4] - 2025-11-25

Fixed

• 🐛 TypeScript-aware hardcode detection - dramatically reduces false positives by 35.7%:
  • Ignore strings in TypeScript union types (type Status = 'active' | 'inactive')
  • Ignore strings in interface property types (interface { mode: 'development' | 'production' })
  • Ignore strings in type assertions (as 'read' | 'write')
  • Ignore strings in typeof checks (typeof x === 'string')
  • Ignore strings in Symbol() calls for DI tokens (Symbol('LOGGER'))
  • Ignore strings in dynamic import() calls (import('../../module.js'))
  • Exclude tokens.ts/tokens.js files completely (DI container files)
  • Tested on real-world TypeScript project: 985 → 633 issues (352 false positives eliminated)
• ✅ Added 13 new tests for TypeScript type context filtering

[0.7.3] - 2025-11-25

Fixed

• 🐛 False positive: repository importing its own aggregate:
  • Added isInternalBoundedContextImport() method to detect internal imports
  • Imports like ../aggregates/Entity from repositories/Repo are now allowed
  • This correctly allows ICodeProjectRepository to import CodeProject from the same bounded context
  • Cross-aggregate imports (with multiple ../..) are still detected as violations

[0.7.2] - 2025-11-25

Fixed

• 🐛 False positive: errors folder detected as aggregate:
  • Added errors and exceptions to DDD_FOLDER_NAMES constants
  • Added to nonAggregateFolderNames — these folders are no longer detected as aggregates
  • Added to allowedFolderNames — imports from errors/exceptions folders are allowed across aggregates
  • Fixes issue where domain/code-analysis/errors/ was incorrectly identified as a separate aggregate named "errors"

[0.7.1] - 2025-11-25

Fixed

• 🐛 Aggregate Boundary Detection for relative paths:

  • Fixed regex pattern to support paths starting with domain/ (without leading /)
  • Now correctly detects violations in projects scanned from parent directories

• 🐛 Reduced false positives in Repository Pattern detection:

  • Removed findAll, exists, count from ORM technical methods blacklist
  • These are now correctly recognized as valid domain method names
  • Added exists, count, countBy[A-Z] to domain method patterns

• 🐛 Non-aggregate folder exclusions:

  • Added exclusions for standard DDD folders: constants, shared, factories, ports, interfaces
  • Prevents false positives when domain layer has shared utilities

Changed

• ♻️ Extracted magic strings to constants:
  • DDD folder names (entities, aggregates, value-objects, etc.) moved to DDD_FOLDER_NAMES
  • Repository method suggestions moved to REPOSITORY_METHOD_SUGGESTIONS
  • Fallback suggestions moved to REPOSITORY_FALLBACK_SUGGESTIONS

Added

• 📁 Aggregate boundary test examples:
  • Added examples/aggregate-boundary/domain/ with Order, User, Product aggregates
  • Demonstrates cross-aggregate entity reference violations

[0.7.0] - 2025-11-25

Added

🔒 Aggregate Boundary Validation

New DDD feature to enforce aggregate boundaries and prevent tight coupling between aggregates.

• ✅ Aggregate Boundary Detector:

  • Detects direct entity references across aggregate boundaries
  • Validates that aggregates reference each other only by ID or Value Objects
  • Supports multiple folder structure patterns:
    • domain/aggregates/order/Order.ts
    • domain/order/Order.ts
    • domain/entities/order/Order.ts

• ✅ Smart Import Analysis:

  • Parses ES6 imports and CommonJS require statements
  • Identifies entity imports from other aggregates
  • Allows imports from value-objects, events, services, specifications folders

• ✅ Actionable Suggestions:

  • Reference by ID instead of entity
  • Use Value Objects to store needed data from other aggregates
  • Maintain aggregate independence

• ✅ CLI Integration:

  • --architecture flag includes aggregate boundary checks
  • CRITICAL severity for violations
  • Detailed violation messages with file:line references

• ✅ Test Coverage:

  • 41 new tests for aggregate boundary detection
  • 333 total tests passing (100% pass rate)
  • Examples in examples/aggregate-boundary/

Technical

• New AggregateBoundaryDetector in infrastructure layer
• New AggregateBoundaryViolation value object in domain layer
• New IAggregateBoundaryDetector interface for dependency inversion
• Integrated into AnalyzeProject use case

[0.6.4] - 2025-11-24

Added

🎯 Smart Context-Aware Suggestions for Repository Method Names

Guardian now provides intelligent, context-specific suggestions when it detects non-domain method names in repositories.

• ✅ Intelligent method name analysis:

  • queryUsers() → Suggests: search, findBy[Property]
  • selectById() → Suggests: findBy[Property], get[Entity]
  • insertUser() → Suggests: create, add[Entity], store[Entity]
  • updateRecord() → Suggests: update, modify[Entity]
  • upsertUser() → Suggests: save, store[Entity]
  • removeUser() → Suggests: delete, removeBy[Property]
  • fetchUserData() → Suggests: findBy[Property], get[Entity]
  • And more technical patterns detected automatically!

• 🎯 Impact:

  • Developers get actionable, relevant suggestions instead of generic examples
  • Faster refactoring with specific naming alternatives
  • Better learning experience for developers new to DDD

Fixed

• ✅ Expanded domain method patterns support:

  • find*() methods - e.g., findNodes(), findNodeById(), findSimilar()
  • saveAll() - batch save operations
  • deleteBy*() methods - e.g., deleteByPath(), deleteById()
  • deleteAll() - clear all entities
  • add*() methods - e.g., addRelationship(), addItem()
  • initializeCollection() - collection initialization

• 🐛 Removed findAll from technical methods blacklist:

  • findAll() is now correctly recognized as a standard domain method
  • Reduced false positives for repositories using this common pattern

Technical

• Added suggestDomainMethodName() method in RepositoryPatternDetector.ts with keyword-based suggestion mapping
• Updated getNonDomainMethodSuggestion() in RepositoryViolation.ts to extract and use smart suggestions
• Refactored suggestion logic to reduce cyclomatic complexity (22 → 9)
• Enhanced domainMethodPatterns with 9 additional patterns
• All 333 tests passing

[0.6.3] - 2025-11-24

Fixed

🐛 Repository Pattern Detection - Reduced False Positives

Fixed overly strict repository method name validation that was flagging valid DDD patterns as violations.

• ✅ Added support for common DDD repository patterns:

  • has*() methods - e.g., hasProject(), hasPermission()
  • is*() methods - e.g., isCached(), isActive()
  • exists*() methods - e.g., existsById(), existsByEmail()
  • clear*() methods - e.g., clearCache(), clearAll()
  • store*() methods - e.g., storeMetadata(), storeFile()
  • Lifecycle methods: initialize(), close(), connect(), disconnect()

• 🎯 Impact:

  • Reduced false positives in real-world DDD projects
  • Better alignment with Domain-Driven Design best practices
  • More practical for cache repositories, connection management, and business queries

• 📚 Why these patterns are valid:

  • Martin Fowler's Repository Pattern allows domain-specific query methods
  • DDD recommends using ubiquitous language in method names
  • Lifecycle methods are standard for resource management in repositories

Technical

• Updated domainMethodPatterns in RepositoryPatternDetector.ts with 11 additional valid patterns
• All existing functionality remains unchanged

[0.6.2] - 2025-11-24

Added

📚 Research-Backed Documentation

Guardian's detection rules are now backed by scientific research and industry standards!

• ✅ New Documentation

  • docs/WHY.md - User-friendly explanations for each rule with authoritative sources
  • docs/RESEARCH_CITATIONS.md - Complete academic and industry references (551 lines)
  • Organized by detection type with quick navigation

• ✅ Micro-Citations in README

  • Each feature now includes one-line citation with "Why?" link
  • Examples: "Based on MIT 6.031, SonarQube RSPEC-109"
  • Non-intrusive, opt-in for users who want to learn more

• ✅ CLI Help Enhancement

  • Added "BACKED BY RESEARCH" section to --help output
  • Mentions MIT, Martin Fowler, Robert C. Martin, industry standards
  • Link to full documentation

Changed

• Documentation Structure: Moved RESEARCH_CITATIONS.md to docs/ directory for better organization
• All internal links updated to reflect new documentation structure

Backed By

Our rules are supported by:

• 🎓 Academia: MIT Course 6.031, ScienceDirect peer-reviewed studies
• 📚 Books: Clean Architecture (Martin 2017), DDD (Evans 2003), Enterprise Patterns (Fowler 2002)
• 🏢 Industry: Google, Microsoft, Airbnb style guides, SonarQube standards
• 👨‍🏫 Experts: Martin Fowler, Robert C. Martin, Eric Evans, Alistair Cockburn

[0.6.1] - 2025-11-24

Improved

📖 Enhanced CLI Help System

Guardian's --help command is now comprehensive and AI-agent-friendly!

• ✅ Detailed Main Help

  • Complete detector descriptions with quick fix instructions
  • Severity level explanations (CRITICAL → LOW)
  • Step-by-step workflow guide for fixing violations
  • 7 practical usage examples
  • "HOW TO FIX COMMON ISSUES" reference section

• ✅ Better Organization

  • Clear DETECTS section with all 8 violation types
  • Each detector includes → what to do to fix it
  • Severity system with priority guidance
  • Examples cover all major use cases

• ✅ AI Agent Ready

  • Help output provides complete context for autonomous agents
  • Actionable instructions for each violation type
  • Clear workflow: run → review → fix → verify

Fixed

• Code Quality: Extracted all hardcoded strings from help text to constants
  • Moved 17 magic strings to CLI_HELP_TEXT constant
  • Improved maintainability and i18n readiness
  • Follows Clean Code principles (Single Source of Truth)

Technical

• All CLI help strings now use CLI_HELP_TEXT from constants
• Zero hardcode violations in Guardian's own codebase
• Passes all quality checks (format, lint, build, self-check)

[0.6.0] - 2025-11-24

Added

🎯 Output Limit Control

Guardian now supports limiting detailed violation output for large codebases!

• ✅ --limit Option
  • Limit detailed violation output per category: guardian check src --limit 10
  • Short form: -l <number>
  • Works with severity filters: guardian check src --only-critical --limit 5
  • Shows warning when violations exceed limit
  • Full statistics always displayed

📋 Severity Display Constants

• Extracted severity labels and headers to reusable constants
• Improved CLI maintainability and consistency
• SEVERITY_DISPLAY_LABELS and SEVERITY_SECTION_HEADERS

📚 Complete Development Workflow

• Added comprehensive workflow documentation to CLAUDE.md
• 6-phase development process (Planning → Quality → Documentation → Verification → Commit → Publication)
• Quick checklists for new features
• Common workflows and debugging tips

Changed

• ESLint Configuration: Optimized with CLI-specific overrides, reduced warnings from 129 to 0
• Documentation: Updated README with all 8 detector types and latest statistics
• TODO: Added technical debt tracking for low-coverage files

Fixed

• Removed unused SEVERITY_LEVELS import from AnalyzeProject.ts
• Fixed unused fileName variable in HardcodeDetector.ts
• Replaced || with ?? for nullish coalescing

Removed

• Deleted unused IBaseRepository interface (dead code)
• Fixed repository pattern violations detected by Guardian on itself

Technical Details

• All 292 tests passing (100% pass rate)
• Coverage: 90.63% statements, 82.19% branches, 83.51% functions
• ESLint: 0 errors, 0 warnings
• Guardian self-check: ✅ No issues found
• No breaking changes - fully backwards compatible

[0.5.2] - 2025-11-24

Added

🎯 Severity-Based Prioritization

Guardian now intelligently prioritizes violations by severity, helping teams focus on critical issues first!

• ✅ Severity Levels

  • 🔴 CRITICAL: Circular dependencies, Repository pattern violations
  • 🟠 HIGH: Dependency direction violations, Framework leaks, Entity exposures
  • 🟡 MEDIUM: Naming violations, Architecture violations
  • 🟢 LOW: Hardcoded values

• ✅ Automatic Sorting

  • All violations automatically sorted by severity (most critical first)
  • Applied in AnalyzeProject use case before returning results
  • Consistent ordering across all detection types

• ✅ CLI Filtering Options

  • --min-severity <level> - Show only violations at specified level and above
  • --only-critical - Quick filter for critical issues only
  • Examples:
    • guardian check src --only-critical
    • guardian check src --min-severity high

• ✅ Enhanced CLI Output

  • Color-coded severity labels (🔴🟠🟡🟢)
  • Visual severity group headers with separators
  • Severity displayed for each violation
  • Clear filtering messages when filters active

Changed

• Updated all violation interfaces to include severity: SeverityLevel field
• Improved CLI presentation with grouped severity display
• Enhanced developer experience with visual prioritization

Technical Details

• All 292 tests passing (100% pass rate)
• Coverage: 90.63% statements, 82.19% branches, 83.51% functions
• No breaking changes - fully backwards compatible
• Clean Architecture principles maintained

---

[0.5.1] - 2025-11-24

Changed

🧹 Code Quality Refactoring

Major internal refactoring to eliminate hardcoded values and improve maintainability - Guardian now fully passes its own quality checks!

• ✅ Extracted Constants

  • All RepositoryViolation messages moved to domain constants (Messages.ts)
  • All framework leak template strings centralized
  • All layer paths moved to infrastructure constants (paths.ts)
  • All regex patterns extracted to IMPORT_PATTERNS constant
  • 30+ new constants added for better maintainability

• ✅ New Constants Files

  • src/infrastructure/constants/paths.ts - Layer paths, CLI paths, import patterns
  • Extended src/domain/constants/Messages.ts - 25+ repository pattern messages
  • Extended src/shared/constants/rules.ts - Package placeholder constant

• ✅ Self-Validation Achievement

  • Reduced hardcoded values from 37 to 1 (97% improvement)
  • Guardian now passes its own src/ directory checks with 0 violations
  • Only acceptable hardcode remaining: bin/guardian.js entry point path
  • All 292 tests still passing (100% pass rate)

• ✅ Improved Code Organization

  • Better separation of concerns
  • More maintainable codebase
  • Easier to extend with new features
  • Follows DRY principle throughout

Technical Details

• No breaking changes - fully backwards compatible
• All functionality preserved
• Test suite: 292 tests passing
• Coverage: 96.77% statements, 83.82% branches

---

[0.5.0] - 2025-11-24

Added

📚 Repository Pattern Validation

Validate proper implementation of the Repository Pattern to ensure domain remains decoupled from infrastructure.

• ✅ ORM Type Detection in Interfaces

  • Detects ORM-specific types (Prisma, TypeORM, Mongoose) in domain repository interfaces
  • Ensures repository interfaces remain persistence-agnostic
  • Supports detection of 25+ ORM type patterns
  • Provides fix suggestions with clean domain examples

• ✅ Concrete Repository Usage Detection

  • Identifies use cases depending on concrete repository implementations
  • Enforces Dependency Inversion Principle
  • Validates constructor and field dependency types
  • Suggests using repository interfaces instead

• ✅ Repository Instantiation Detection

  • Detects new Repository() in use cases
  • Enforces Dependency Injection pattern
  • Identifies hidden dependencies
  • Provides DI container setup guidance

• ✅ Domain Language Validation

  • Checks repository methods use domain terminology
  • Rejects technical database terms (findOne, insert, query, execute)
  • Promotes ubiquitous language across codebase
  • Suggests business-oriented method names

• ✅ Smart Violation Reporting

  • RepositoryViolation value object with detailed context
  • Four violation types: ORM types, concrete repos, new instances, technical names
  • Provides actionable fix suggestions
  • Shows before/after code examples

• ✅ Comprehensive Test Coverage

  • 31 new tests for repository pattern detection
  • 292 total tests passing (100% pass rate)
  • Integration tests for multiple violation types
  • 96.77% statement coverage, 83.82% branch coverage

• ✅ Documentation & Examples

  • 6 example files (3 bad patterns, 3 good patterns)
  • Comprehensive README with patterns and principles
  • Examples for ORM types, concrete repos, DI, and domain language
  • Demonstrates Clean Architecture and SOLID principles

Changed

• Updated test count: 261 → 292 tests
• Added REPOSITORY_PATTERN rule to constants
• Extended AnalyzeProject use case with repository pattern detection
• Added REPOSITORY_VIOLATION_TYPES constant with 4 violation types
• ROADMAP updated with completed repository pattern validation (v0.5.0)

---

[0.4.0] - 2025-11-24

Added

🔀 Dependency Direction Enforcement

Enforce Clean Architecture dependency rules to prevent architectural violations across layers.

• ✅ Dependency Direction Detector

  • Validates that dependencies flow in the correct direction
  • Domain layer can only import from Domain and Shared
  • Application layer can only import from Domain, Application, and Shared
  • Infrastructure layer can import from all layers
  • Shared layer can be imported by all layers

• ✅ Smart Violation Reporting

  • DependencyViolation value object with detailed context
  • Provides fix suggestions with concrete examples
  • Shows both violating import and suggested layer placement
  • CLI output with severity indicators

• ✅ Comprehensive Test Coverage

  • 43 new tests for dependency direction detection
  • 100% test pass rate (261 total tests)
  • Examples for both good and bad architecture patterns

• ✅ Documentation & Examples

  • Good architecture examples for all layers
  • Bad architecture examples showing common violations
  • Demonstrates proper layer separation

Changed

• Updated test count: 218 → 261 tests
• Optimized extractLayerFromImport method to reduce complexity
• Updated getExampleFix to avoid false positives
• ROADMAP updated with completed dependency direction feature

---

[0.3.0] - 2025-11-24

Added

🚫 Entity Exposure Detection

Prevent domain entities from leaking to API responses, enforcing proper DTO usage at boundaries.

• ✅ Entity Exposure Detector

  • Detects when controllers/routes return domain entities instead of DTOs
  • Scans infrastructure layer (controllers, routes, handlers, resolvers, gateways)
  • Identifies PascalCase entities without Dto/Request/Response suffixes
  • Parses async methods with Promise return types

• ✅ Smart Remediation Suggestions

  • EntityExposure value object with step-by-step fix guidance
  • Suggests creating DTOs with proper naming
  • Provides mapper implementation examples
  • Shows how to separate domain from presentation concerns

• ✅ Comprehensive Test Coverage

  • 24 new tests for entity exposure detection (98% coverage)
  • EntityExposureDetector: 98.07% coverage
  • Overall project: 90.6% statements, 83.97% branches

• ✅ Documentation & Examples

  • BadUserController and BadOrderController examples
  • GoodUserController showing proper DTO usage
  • Integration with CLI for helpful output

Changed

• Updated test count: 194 → 218 tests
• Added entity exposure to violation pipeline
• ROADMAP updated with completed entity exposure feature

---

[0.2.0] - 2025-11-24

Added

🔌 Framework Leak Detection

Major new feature to detect framework-specific imports in domain layer, ensuring Clean Architecture compliance.

• ✅ Framework Leak Detector

  • Detects 250+ framework patterns across 12 categories
  • HTTP frameworks: Express, Fastify, Koa, Hapi, NestJS, etc.
  • ORMs/Databases: Prisma, TypeORM, Sequelize, Mongoose, Drizzle, etc.
  • Loggers: Winston, Pino, Bunyan, Log4js, etc.
  • Caches: Redis, IORedis, Memcached, etc.
  • Message Queues: Bull, BullMQ, KafkaJS, RabbitMQ, etc.
  • And 7 more categories (HTTP clients, validation, DI, email, storage, testing, templates)

• ✅ Smart Violation Reporting

  • Framework type categorization
  • Detailed suggestions for proper abstraction
  • CLI output with severity indicators
  • Integration with existing violation pipeline

• ✅ Comprehensive Test Coverage

  • 35 new tests for framework leak detection
  • 100% coverage of detection logic
  • Examples for all major framework types

• ✅ Documentation & Examples

  • New bad architecture examples showing framework leaks
  • Updated README with framework leak detection section
  • Integration guides for preventing framework coupling

Changed

• Updated test count: 159 → 194 tests across 7 files
• Updated examples: 36 → 38 files (29 good + 9 bad)
• CLI now reports 5 types of violations (added framework leaks)

---

[0.1.0] - 2025-11-24

Added

🎉 Initial Release of @samiyev/guardian

Code quality guardian for vibe coders and enterprise teams - your AI coding companion that keeps code clean while you move fast.

Core Features

• ✨ Hardcode Detection

  • Detects magic numbers (timeouts, ports, limits, retries, delays)
  • Detects magic strings (URLs, connection strings, API endpoints, error messages)
  • Smart context analysis to reduce false positives
  • Automatic constant name suggestions based on context
  • Location suggestions for extracted constants (domain/shared/infrastructure)
  • Ignores allowed numbers: -1, 0, 1, 2, 10, 100, 1000
  • Ignores console.log, imports, tests, and exported constants

• 🔄 Circular Dependency Detection

  • Detects import cycles in codebase (A → B → A, A → B → C → A, etc.)
  • Shows complete dependency chain for each cycle
  • CLI output with detailed cycle path and severity
  • Supports detection of multiple independent cycles
  • Handles complex graphs with both cyclic and acyclic parts

• 📝 Naming Convention Enforcement

  • Layer-based naming rules for Clean Architecture
  • Domain Layer:
    • Entities: PascalCase nouns (User.ts, Order.ts)
    • Services: *Service suffix (UserService.ts)
    • Repository interfaces: I*Repository prefix (IUserRepository.ts)
    • Forbidden patterns: Dto, Controller, Request, Response
  • Application Layer:
    • Use cases: Verb + Noun in PascalCase (CreateUser.ts, UpdateProfile.ts)
    • DTOs: *Dto, *Request, *Response suffixes
    • Mappers: *Mapper suffix
  • Infrastructure Layer:
    • Controllers: *Controller suffix
    • Repository implementations: *Repository suffix
    • Services: *Service or *Adapter suffixes
  • Smart exclusion system for base classes
  • Support for 26 standard use case verbs

• 🏗️ Architecture Violations

  • Clean Architecture layer validation
  • Dependency rules enforcement:
    • Domain → can only import Shared
    • Application → can import Domain, Shared
    • Infrastructure → can import Domain, Application, Shared
    • Shared → cannot import anything
  • Layer detection from file paths
  • Import statement analysis

• 🔌 Framework Leak Detection

  • Detects framework-specific imports in domain layer
  • Identifies coupling between domain and infrastructure concerns
  • Checks for HTTP framework leaks (Express, Fastify, Koa, Hapi, NestJS)
  • Checks for ORM/Database leaks (Prisma, TypeORM, Sequelize, Mongoose, Drizzle)
  • Checks for external service leaks (AWS SDK, Firebase, Stripe, Twilio)
  • Reports violations with framework type and suggested fixes
  • Helps maintain clean domain boundaries

CLI Interface

• 🛠️ Command-line tool (guardian command)
  • guardian check <path> - analyze project
  • --exclude <dirs> - exclude directories
  • --verbose - detailed output
  • --no-hardcode - skip hardcode detection
  • --no-architecture - skip architecture checks
  • --version - show version
  • --help - show help

Reporting & Metrics

• 📊 Comprehensive metrics
  • Total files analyzed
  • Total functions count
  • Total imports count
  • Layer distribution statistics (domain/application/infrastructure/shared)
  • Detailed violation reports with file:line:column
  • Context snippets for each violation
  • Smart suggestions for fixing issues

Developer Experience

• 🤖 Built for AI-Assisted Development

  • Perfect companion for GitHub Copilot, Cursor, Windsurf, Claude, ChatGPT, Cline
  • Catches common AI code smells (hardcoded values, architecture violations)
  • Educational error messages with fix suggestions
  • Designed for vibe coding workflow: AI writes → Guardian reviews → AI fixes → Ship

• 🏢 Enterprise-Ready

  • Enforce architectural standards at scale
  • CI/CD integration ready
  • JSON/Markdown output for automation
  • Security: catch hardcoded secrets before production
  • Metrics export for dashboards

Examples & Documentation

• 📚 Comprehensive examples (38 files)
  • Good Architecture (29 files): Complete DDD/Clean Architecture patterns
    • Domain: Aggregates, Entities, Value Objects, Events, Services, Factories, Specifications
    • Application: Use Cases, DTOs, Mappers
    • Infrastructure: Repositories, Controllers
  • Bad Architecture (9 files): Anti-patterns to avoid
    • Hardcoded values, Circular dependencies, Framework leaks, Entity exposure, Naming violations
  • All examples fully documented with explanations
  • Can be used as templates for new projects

Testing & Quality

• ✅ Comprehensive test suite

  • 194 tests across 7 test files
  • All tests passing
  • 80%+ code coverage on all metrics
  • Test fixtures for various scenarios
  • Integration and unit tests

• 🧹 Self-analyzing

  • Guardian passes its own checks with 0 violations
  • All constants extracted (no hardcoded values)
  • Follows Clean Architecture
  • No circular dependencies
  • Proper naming conventions

Technical Details

Architecture:

• Built with Clean Architecture principles
• Domain-Driven Design (DDD) patterns
• Layered architecture (Domain, Application, Infrastructure, Shared)
• TypeScript with strict type checking
• Tree-sitter based AST parsing

Dependencies:

• commander ^12.1.0 - CLI framework
• simple-git ^3.30.0 - Git operations
• tree-sitter ^0.21.1 - Abstract syntax tree parsing
• tree-sitter-javascript ^0.23.0 - JavaScript parser
• tree-sitter-typescript ^0.23.0 - TypeScript parser
• uuid ^13.0.0 - UUID generation

Development:

• TypeScript 5.7.3
• Vitest 4.0.10 for testing
• Node.js >= 18.0.0 required
• CommonJS output with full TypeScript declarations
• Source maps included

Package:

• Size: ~58 KB compressed
• Unpacked: ~239 KB
• 172 files included
• Public npm package (@samiyev/guardian)
• CLI binary: guardian

Documentation

• 📖 Comprehensive README (25KB+)

  • Quick start for vibe coders (30-second setup)
  • Enterprise integration guides (CI/CD, pre-commit, metrics)
  • Real-world examples and workflows
  • API documentation
  • FAQ for both vibe coders and enterprise teams
  • Success stories and use cases

• 🗺️ Roadmap - Future features and improvements

• 📋 Contributing guidelines

• 📝 TODO list - Technical debt tracking

• 📄 MIT License

Notes

• First public release on npm
• Production-ready for both individual developers and enterprise teams
• Perfect for AI-assisted development workflows
• Enforces Clean Architecture at scale
• Zero violations in own codebase (self-tested)

---

Future Releases

Planned features for upcoming versions:

• Configuration file support (.guardianrc)
• Custom rule definitions
• Plugin system
• Multi-language support
• Watch mode
• Auto-fix capabilities
• Git integration (check only changed files)
• Performance optimizations

See ROADMAP.md for detailed feature roadmap.

---

Version Guidelines

Semantic Versioning

MAJOR.MINOR.PATCH (e.g., 1.2.3)

• MAJOR - Incompatible API changes
• MINOR - New features, backwards compatible
• PATCH - Bug fixes, backwards compatible

Release Checklist

Before releasing a new version:

• Update CHANGELOG.md with all changes
• Update version in package.json
• Run pnpm test - all tests pass
• Run pnpm build - clean build
• Run pnpm test:coverage - coverage >= 80%
• Update ROADMAP.md if needed
• Update README.md if API changed
• Create git tag: git tag v0.1.0
• Push to GitHub: git push origin main --tags
• Publish to npm: npm publish

---

Links:

• Official Website
• GitHub Repository
• npm Package
• Documentation
• Roadmap
• Issues