Multi-project Claude Code orchestration dashboard for tracking progress across a portfolio of projects.
- Scan projects - Automatically detect projects by marker files (package.json, pyproject.toml, Cargo.toml, etc.)
- Parse progress - Extract completion %, phases, and next actions from TODO.md, PROGRESS.md files
- Health scoring - Calculate health scores based on completion, activity, documentation, and git state
- Decision tracking - Surface pending decisions (Option A/B patterns) that need resolution
- Claude Code integration - Generate context-aware continue prompts and launch sessions
- Batch launching - Launch multiple Claude Code sessions in parallel with iTerm2/tmux
# Clone the repo
git clone https://github.com/dbbuilder-org/claude-code-project-manager.git
cd claude-code-project-manager
# Run setup script
./setup.sh
# Activate virtual environment
source venv/bin/activate# Scan your projects directory
pm scan ~/dev2
# View portfolio summary
pm summary
# View project status table
pm status
# View health scores (best first)
pm health
# View projects needing attention
pm health --asc
# Generate continue prompt for a project
pm continue myproject --dry-run
# Launch Claude Code for a project
pm launch myproject
# Open dashboard
pm dashboardScan a directory for projects and update the database.
pm scan ~/dev2 # Scan with default settings
pm scan ~/dev2 --verbose # Show detailed outputProject detection:
- Node.js:
package.json - Python:
pyproject.toml,setup.py - Rust:
Cargo.toml - Go:
go.mod - Ruby:
Gemfile
Container folders: Directories named clients/ are scanned recursively for client projects.
Show project status table with progress and phase information.
pm status # Show all projects
pm status --filter type:client # Filter by category
pm status --filter status:active # Only incomplete projects
pm status --sort completion # Sort by completion %
pm status --limit 10 # Limit results
pm status --json # JSON outputQuick summary of portfolio health and distribution.
Show projects sorted by health score.
pm health # Healthiest first
pm health --asc # Needs attention (lowest first)
pm health --filter type:client # Filter by category
pm health --limit 10 # Limit resultsHealth score factors:
- Completion % (40 points max)
- Recent activity (20 points)
- CLAUDE.md present (15 points)
- Clean git state (15 points)
- No pending decisions (10 points)
Generate a context-aware continue command for Claude Code.
pm continue myproject # Generate and copy to clipboard
pm continue myproject --dry-run # Preview without copying
pm continue myproject --mode decision # Focus on pending decision
pm continue --filter type:client # Pick from filtered projectsLaunch Claude Code sessions for one or more projects.
pm launch myproject # Launch single project
pm launch proj1 proj2 proj3 # Launch multiple projects
pm launch --filter type:client # Launch all client projects
pm launch --dry-run myproject # Preview without launching
pm launch --parallel 3 proj1 proj2 proj3 # Limit parallel sessionsOpen the Streamlit dashboard for visual project management.
The parser recognizes these patterns in TODO.md and PROGRESS.md files:
- [x] Completed task
- [ ] Pending task
- [ ] IN PROGRESS task
- [ ] BLOCKED task## Completion: 65%
## Progress: 8 of 10 complete**Status:** Phase 2 in progress
**Current Phase:** Implementation
**Current Focus:** Building the API layer
**Next Step:** Add unit tests## Architecture Decision
### Option A: PostgreSQL
More scalable for production.
### Option B: SQLite
Simpler for development.
Recommended: Option A# Run tests
pytest tests/
# Run with coverage
pytest tests/ --cov=pm --cov-report=html
# Run specific test file
pytest tests/test_parser.py -vproject-manager/
├── pm/ # Main package
│ ├── cli.py # Click CLI commands
│ ├── database/ # SQLAlchemy models
│ ├── scanner/ # Project detection and parsing
│ └── generator/ # Prompt generation
├── dashboard/ # Streamlit dashboard
├── tests/ # Test suite
│ ├── conftest.py # Fixtures
│ ├── test_cli.py # CLI tests
│ ├── test_detector.py # Scanner tests
│ ├── test_parser.py # Parser tests
│ ├── test_models.py # Database tests
│ ├── test_prompts.py # Generator tests
│ └── test_e2e.py # End-to-end tests
└── pyproject.toml # Package configuration
MIT