Skip to main content

👻 Spectral Software Solutions - Your codebase just got possessed by supernatural development powers

Project description

👻 Haunted - Spectral Software Solutions

Your codebase just got possessed by supernatural development powers.

Transform your development workflow with an autonomous AI spirit that thinks, codes, and ships features while you focus on the bigger picture. Haunted doesn't just assist - it possesses your repository and handles complete development cycles from planning to deployment.

🌙 No API Séance Required - Seamlessly channels your Claude Code authentication for effortless spectral integration.

繁體中文 README | English

🔮 Supernatural Features

  • 👻 Spectral Authentication: No API keys needed - channels your Claude Code powers directly
  • 🎭 Autonomous Possession: AI takes complete ownership of features from concept to deployment
  • 🌙 Concurrent Hauntings: Multiple issues developed simultaneously across your codebase
  • 🕯️ Self-Exorcising: When bugs appear, the spirit debugs and fixes itself automatically
  • 🦇 Git Manifestation: Automatic branch creation, testing, and merging with ghostly precision
  • 👺 Issue-Driven Séances: All development starts from clear issue descriptions and priorities
  • 🎃 Spectral Workflow: Complete development lifecycle - Plan → Code → Test → Debug → Ship

🕯️ Summoning Requirements

  • Python 3.10+ - The vessel for spectral powers
  • Node.js 18+ - Required for Claude Code channeling
  • Claude Code CLI - Your gateway to the supernatural realm:
# Install Claude Code CLI
npm install -g @anthropic-ai/claude-code

# Authenticate with the spirit realm (this opens the portal!)
claude login

🎭 Possession Ritual

# Instant spectral possession (recommended)
uvx haunted

# Or summon globally with pipx
pipx install haunted

# For development possession
git clone <repository-url>
cd haunted
uv sync

🌙 The possession is complete! No API keys needed - Haunted channels your Claude Code authentication automatically.

🔮 Confirm the Haunting

# Test the spectral connection
uvx haunted --help

# For pipx installation
haunted --help

🌙 Summoning Your Spectral Developer

1. Begin the Possession

# Invite the spirit into your project
uvx haunted init

# Or if globally installed with pipx
haunted init

This spectral ritual will:

  • Manifest .haunted/ sanctuary with database and config
  • Verify your Git repository is ready for haunting
  • Establish supernatural configuration

2. Whisper Your Desires

# Communicate your high-priority wish to the spirit
uvx haunted issue create "Implement user authentication" --priority high --description "Add login/logout functionality with JWT tokens"

# Organize supernatural work into phases
uvx haunted phase create "Phase 1 - Core Features" --description "Essential features for MVP"

# Channel additional requests into specific phases
uvx haunted issue create "Add password reset" --phase <phase-id> --priority medium

3. Release the Autonomous Spirit

# Unleash your spectral developer
uvx haunted start

Your ghostly assistant will:

  • Scan for open Issues by supernatural priority
  • Manifest Git branches for each spectral task
  • Possess your codebase through the complete development cycle
  • Automatically merge completed hauntings

4. Monitor Progress

# Check overall status
uvx haunted status

# List all issues
uvx haunted issue list

# View specific issue details
uvx haunted issue show <issue-id>

# View issues by status
uvx haunted issue list --status in_progress

Workflow

Haunted implements the development workflow from docs/DEVELOPMENT_WORKFLOW.md:

  1. Plan: AI analyzes requirements and creates implementation strategy
  2. Implement: AI writes code following the plan
  3. Unit Test: AI creates and runs unit tests
  4. Fix Issues: AI fixes any test failures
  5. Integration Test: AI runs integration tests
  6. Diagnose: If integration tests fail, AI diagnoses and replans
  7. Done: Issue completed and merged

Commands

Core Commands

  • uvx haunted init - Initialize Haunted in current project
  • uvx haunted start - Start the AI daemon
  • uvx haunted stop - Stop the daemon
  • uvx haunted status - Show current status

Issue Management

  • uvx haunted issue create <title> - Create new issue
  • uvx haunted issue list - List all issues
  • uvx haunted issue show <id> - Show issue details
  • uvx haunted issue comment <id> <message> - Add comment to issue

Phase Management

  • uvx haunted phase create <name> - Create new phase
  • uvx haunted phase list - List all phases

Configuration

Configuration is stored in .haunted/config.yaml:

api:
  # No API key needed! Uses Claude Code authentication
  model: claude-3-5-sonnet-20241022  # Claude Code model
  max_concurrent_issues: 3
  rate_limit_retry: true

daemon:
  scan_interval: 30  # seconds
  max_iterations: 3  # max workflow cycles per issue

git:
  auto_merge: true
  auto_commit: true

database:
  url: sqlite:///.haunted/haunted.db

Architecture

haunted/
├── cli/           # Command-line interface
├── core/          # Core business logic
│   ├── claude_wrapper.py # Claude Code SDK integration
│   ├── workflow.py       # Workflow engine
│   ├── database.py       # Database management
│   └── git_manager.py    # Git operations
├── models/        # SQLModel data models
├── daemon/        # Background service
├── mcp/           # MCP tools for Claude
└── utils/         # Utilities and config

Development Workflow Integration

Haunted is designed to work with your existing development workflow:

  1. Create Issues for features, bugs, or tasks
  2. Let AI work - Haunted processes Issues autonomously
  3. Review Results - Check AI's work in Git branches
  4. Provide Feedback - Add comments to blocked Issues
  5. Merge & Deploy - Completed Issues are auto-merged

Git Branch Strategy

  • main: Production branch
  • phase/: Phase branches for organizing work
  • issue/: Individual Issue branches
  • Auto-merge: Issues -> Phases -> Main (when ready)

MCP Tools

Claude has access to comprehensive tools:

  • File operations (read, write, list)
  • Command execution
  • Git operations
  • Issue management
  • Code search and analysis

Troubleshooting

Common Issues

  1. Claude Code not authenticated: Run claude login first
  2. Claude Code not installed: Install with npm install -g @anthropic-ai/claude-code
  3. Python version < 3.10: Upgrade to Python 3.10 or higher
  4. Not a Git repository: Run git init first
  5. Database errors: Delete .haunted/haunted.db and reinitialize

Logs

Enable verbose logging:

uvx haunted --verbose start

Or specify log file:

uvx haunted --log-file haunted.log start

Examples

Basic Workflow

# Initialize project
uvx haunted init

# Create issues
uvx haunted issue create "Add user model" --priority high
uvx haunted issue create "Implement API endpoints" --priority high
uvx haunted issue create "Add input validation" --priority medium

# Start AI processing
uvx haunted start

# Monitor progress
watch uvx haunted status

Issue Management

# View issue details
uvx haunted issue show abc123

# Add clarification comment
uvx haunted issue comment abc123 "Please use bcrypt for password hashing"

# Check all open issues
uvx haunted issue list --status open

Contributing

  1. Fork the repository
  2. Create feature branch: git checkout -b feature/name
  3. Make changes and test
  4. Submit pull request

License

MIT License - see LICENSE file for details.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

haunted-0.1.0.tar.gz (87.6 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

haunted-0.1.0-py3-none-any.whl (46.9 kB view details)

Uploaded Python 3

File details

Details for the file haunted-0.1.0.tar.gz.

File metadata

  • Download URL: haunted-0.1.0.tar.gz
  • Upload date:
  • Size: 87.6 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? Yes
  • Uploaded via: twine/6.1.0 CPython/3.13.7

File hashes

Hashes for haunted-0.1.0.tar.gz
Algorithm Hash digest
SHA256 27ac9588fbbe02e6209d412d6bf31891c6f3c27b1d76611837293af45339f908
MD5 2274e0fb2246e509aacd3170998f166d
BLAKE2b-256 2b57d09b5ac8d77852ef3ba27210433a696be0c217fd9b9149fad3c1a39d2bb8

See more details on using hashes here.

Provenance

The following attestation bundles were made for haunted-0.1.0.tar.gz:

Publisher: python-publish.yml on Pr0gCat/haunted

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

File details

Details for the file haunted-0.1.0-py3-none-any.whl.

File metadata

  • Download URL: haunted-0.1.0-py3-none-any.whl
  • Upload date:
  • Size: 46.9 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? Yes
  • Uploaded via: twine/6.1.0 CPython/3.13.7

File hashes

Hashes for haunted-0.1.0-py3-none-any.whl
Algorithm Hash digest
SHA256 053f8debe8d557252bac412fabf8aa4ff790a2c123359148b25330c6d39d29bc
MD5 f63e229f1563fb046577f37867073759
BLAKE2b-256 9382a37e67ec30ebcccf5d85cafb134c8f129085a806670d92e7c8fdaf6eb61f

See more details on using hashes here.

Provenance

The following attestation bundles were made for haunted-0.1.0-py3-none-any.whl:

Publisher: python-publish.yml on Pr0gCat/haunted

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page