This document organizes the Claude Code Watchdog project into parallel workstreams to maximize development velocity and enable concurrent work across multiple tracks.
Project Duration: 4 weeks Team Size: 1-2 developers Parallel Workstreams: 8
┌─────────────────────────────────────────────────────────────────┐
│ WORKSTREAM DEPENDENCIES │
└─────────────────────────────────────────────────────────────────┘
Phase 1 (Week 1):
┌──────────────────┐
│ WS1: Core Infra │ (CRITICAL PATH)
└────────┬─────────┘
│
┌────┴────┬──────────┬──────────┐
│ │ │ │
┌───▼──┐ ┌──▼──┐ ┌───▼───┐ ┌──▼──────┐
│ WS2 │ │ WS3 │ │ WS4 │ │ WS5 │ (PARALLEL)
│State │ │Dec. │ │Action │ │Project │
└───┬──┘ └──┬──┘ └───┬───┘ └──┬──────┘
│ │ │ │
└────┬───┴──────────┴─────────┘
│
┌────▼────┐
│ WS6 │
│Logging │
└─────────┘
Phase 2 (Week 2):
WS3 (continues) - Add API intelligence
WS4 (continues) - Add skills
Phase 3 (Week 3):
WS5 (continues) - Multi-project
WS4 (continues) - Git operations
Phase 4 (Week 4):
┌─────────┐ ┌──────────┐
│ WS7 │ │ WS8 │ (PARALLEL)
│ Testing │ │ Docs │
└─────────┘ └──────────┘
Owner: Developer 1 Priority: P0 (Critical) Duration: Week 1 Total Effort: 7 hours
Establish the foundational project structure, module system, and Windows MCP integration layer.
- WI-1.1: Project Structure Setup (2h)
- WI-1.2: Windows MCP Integration Wrapper (3h)
- WI-1.9: Installation Script (2h)
- None (First workstream to start)
- Complete directory structure
- All module files with function signatures
- Windows MCP wrapper functions (State, Click, Type, Key)
- Basic installation script
- Module import system working
- All modules can be imported without errors
- Windows MCP tools callable from PowerShell
- Directory structure matches architecture
- Installation script runs on clean system
- Can work independently from other workstreams initially
- Must complete before WS2, WS3, WS4, WS5 can start core implementation
Owner: Developer 1 or 2 Priority: P0 (Critical) Duration: Week 1 (4h) + Week 3 (3h) Total Effort: 7 hours
Build robust UI state detection and classification system for Claude Code sessions.
Week 1:
- WI-1.3: State Detection Engine (4h)
Week 3:
- WI-3.1: Multi-Project Session Detection (4h)
- WI-2.5: Enhanced State Detection (3h) - moved from week 2
- Requires: WS1 (Windows MCP wrapper)
- Blocks: WS3 (Decision Engine needs state classification)
- State detection for 6 primary states
- TODO parsing with accuracy >95%
- Error detection and classification
- Processing indicator detection
- Multi-project session identification
- Enhanced error severity classification
- 98%+ accuracy on state classification
- Detects all active Claude Code sessions
- Correctly maps sessions to projects
- Handles edge cases gracefully
- Week 1: Can develop in parallel with WS4 (Action)
- Week 3: Can develop in parallel with WS4 (Git operations)
- Testing can happen in parallel with WS7
Owner: Developer 1 Priority: P0 (Critical) Duration: Week 1 (3h) + Week 2 (11h) Total Effort: 14 hours
Create intelligent decision-making system with both rule-based and AI-powered capabilities.
Week 1:
- WI-1.4: Rule-Based Decision Engine (3h)
Week 2:
- WI-2.1: Claude API Integration (4h)
- WI-2.2: Advanced Decision Engine (5h)
- WI-2.7: API Configuration Management (2h)
- Requires: WS2 (State Detection for input)
- Requires: WS1 (Core infrastructure)
- Blocks: WS4 (Action Executor needs decisions)
- Rule-based decision logic for all states
- Claude API integration with secure key storage
- Advanced decision engine with full context
- Decision history tracking
- Fallback from API to rules
- API configuration system
- Decisions make logical sense for each state
- API decisions more accurate than rule-based
- Confidence scores reflect decision quality
- Fallback works when API unavailable
- API costs within budget
- Week 1: Rule-based can develop in parallel with WS4 (Command Execution)
- Week 2: API integration can develop in parallel with WS4 (Skill resolution)
- Decision logging integrates with WS6
Owner: Developer 1 or 2 Priority: P0 (Critical) Duration: Week 1 (4h) + Week 2 (4h) + Week 3 (13h) Total Effort: 21 hours
Execute decisions through UI commands, skill invocations, and Git operations.
Week 1:
- WI-1.5: Command Execution Module (4h)
Week 2:
- WI-2.3: Skill-Based Error Resolution (4h)
Week 3:
- WI-3.3: Git Integration Module (5h)
- WI-3.4: Phase Transition Logic (4h)
- WI-3.5: GitHub Pull Request Creation (4h)
- Requires: WS1 (Windows MCP wrapper)
- Requires: WS3 (Decisions to execute)
- Command sending with retry logic
- Command verification
- Skill invocation system
- Git wrapper functions (branch, commit, push)
- Phase transition management
- Automated PR creation
- Commands sent successfully >95% of time
- Skills invoked correctly
- Git operations complete without errors
- PRs created with proper formatting
- Phase transitions work seamlessly
- Week 1: Can develop in parallel with WS2 (State Detection)
- Week 2: Can develop in parallel with WS3 (API integration)
- Week 3: Git operations can develop in parallel with WS2 (Multi-project detection)
Owner: Developer 1 or 2 Priority: P0 (Critical) Duration: Week 1 (7h) + Week 3 (7h) Total Effort: 14 hours
Build project registration, configuration, and multi-project orchestration systems.
Week 1:
- WI-1.6: Project Registration System (3h)
- WI-1.7: Main Watchdog Loop (4h)
Week 3:
- WI-3.2: Concurrent Project Processing (3h)
- WI-3.6: Session Recovery System (4h)
- Requires: WS1 (Core infrastructure)
- Integrates with: WS2, WS3, WS4 (orchestrates all)
- Project registration and validation
- Central registry management
- Main watchdog polling loop
- Multi-project processing
- Session recovery system
- State persistence
- Can register and manage 5+ projects
- Main loop runs continuously without crashes
- Projects isolated from each other
- Recovery works after interruptions
- Loop stops gracefully on Ctrl+C
- Week 1: Registration can develop while other workstreams build components
- Week 1: Main loop integrates components as they become available
- Week 3: Can develop in parallel with WS4 (Git operations)
Owner: Developer 2 Priority: P1 (High) Duration: Week 1 (2h) + Week 2 (2h) + Week 3 (3h) Total Effort: 7 hours
Provide comprehensive logging, notifications, and progress reporting.
Week 1:
- WI-1.8: Logging Infrastructure (2h)
Week 2:
- WI-2.6: Decision Log Enhancements (2h)
Week 3:
- WI-3.7: Progress Reporting (3h)
- Requires: WS1 (Core infrastructure for file I/O)
- Integrates with: All workstreams (provides logging)
- Markdown decision logs
- Console output with colors
- Windows toast notifications
- API metadata in logs
- Progress reports
- Daily summaries
- All decisions logged with context
- Notifications arrive when expected
- Logs readable and useful
- Reports provide actionable insights
- Can develop largely in parallel with all other workstreams
- Integrates incrementally as other features added
- Week 1-3: Can develop continuously alongside core features
Owner: Developer 1 & 2 (shared) Priority: P0 (Critical) Duration: Week 1 (3h) + Week 2 (3h) + Week 3 (4h) + Week 4 (14h) Total Effort: 24 hours
Ensure code quality through comprehensive unit and integration testing.
Week 1:
- WI-1.10: Integration Testing - Phase 1 (3h)
Week 2:
- WI-2.8: Integration Testing - Phase 2 (3h)
Week 3:
- WI-3.8: Integration Testing - Phase 3 (4h)
Week 4:
- WI-4.1: Comprehensive Error Handling (4h)
- WI-4.2: Unit Test Suite (6h)
- WI-4.3: Integration Test Suite (4h)
- Requires: Features from all workstreams to test
- Unit tests for all modules (80%+ coverage)
- Integration test suite
- Error handling audit
- Test automation
- Continuous testing framework
- 80%+ code coverage
- All critical paths tested
- No unhandled exceptions
- Tests run automatically
- All tests passing
- Week 1-3: Integration testing happens at end of each sprint
- Week 4: Unit tests can be developed in parallel with documentation
- Can split: Developer 1 on unit tests, Developer 2 on integration tests
Owner: Developer 1 or 2 Priority: P0 (Critical) Duration: Week 2 (2h) + Week 4 (16h) Total Effort: 18 hours
Create comprehensive documentation and prepare production release.
Week 2:
- WI-2.4: Cost Tracking System (includes docs) (3h)
Week 4:
- WI-4.4: Performance Optimization (3h)
- WI-4.5: User Documentation (4h)
- WI-4.6: Developer Documentation (3h)
- WI-4.7: Troubleshooting Guide (2h)
- WI-4.8: Installation Wizard Enhancement (3h)
- WI-4.9: Production Deployment Testing (4h)
- WI-4.10: Release Preparation (2h)
- Requires: All features complete before docs finalized
- README.md complete
- QUICKSTART.md
- ARCHITECTURE.md updated
- API-REFERENCE.md
- TROUBLESHOOTING.md
- Installation wizard
- v1.0 release package
- Documentation accurate and complete
- New user can install and use within 30 minutes
- Developers can understand architecture
- All issues have documented solutions
- Release package works on clean system
- Week 4: Documentation can be split between 2 developers
- User docs and developer docs can be written in parallel
- Testing and documentation can overlap
- One developer on docs, one on optimization
If working alone, follow this critical path:
Week 1 (Focus: Foundation):
- WS1: Core Infrastructure (Day 1-2)
- WS2: State Detection (Day 2-3)
- WS4: Command Execution (Day 3)
- WS3: Rule-Based Decisions (Day 4)
- WS5: Registration + Main Loop (Day 4-5)
- WS6: Logging (Day 5)
- WS7: Integration Test (Day 5)
Week 2 (Focus: Intelligence):
- WS3: API Integration (Day 1)
- WS3: Advanced Decisions (Day 2)
- WS4: Skill Resolution (Day 3)
- WS8: Cost Tracking (Day 3)
- WS2: Enhanced State Detection (Day 4)
- WS6: Decision Log Enhancements (Day 4)
- WS7: Integration Test (Day 5)
Week 3 (Focus: Scale):
- WS2: Multi-Project Detection (Day 1)
- WS5: Concurrent Processing (Day 1)
- WS4: Git Integration (Day 2)
- WS4: Phase Transitions (Day 3)
- WS4: PR Creation (Day 3-4)
- WS5: Session Recovery (Day 4)
- WS6: Progress Reporting (Day 5)
- WS7: Integration Test (Day 5)
Week 4 (Focus: Polish):
- WS7: Error Handling (Day 1)
- WS7: Unit Tests (Day 1-2)
- WS7: Integration Tests (Day 3)
- WS8: User Docs (Day 3)
- WS8: Developer Docs (Day 4)
- WS8: Troubleshooting Guide (Day 4)
- WS8: Production Testing (Day 5)
- WS8: Release Prep (Day 5)
Optimal parallel work distribution:
Developer 1 (Lead - Architecture & Core):
- Week 1: WS1, WS3, WS5 (Infrastructure, Decisions, Orchestration)
- Week 2: WS3 (API), WS8 (Cost Tracking)
- Week 3: WS4 (Git), WS5 (Multi-project)
- Week 4: WS7 (Unit Tests), WS8 (Developer Docs)
Developer 2 (Integration & Quality):
- Week 1: WS2, WS4, WS6 (State, Actions, Logging)
- Week 2: WS4 (Skills), WS2 (Enhanced State), WS6 (Logging)
- Week 3: WS2 (Multi-session), WS6 (Reporting), WS7 (Integration Tests)
- Week 4: WS7 (Integration Tests), WS8 (User Docs, Optimization)
| Workstream | Depends On | Blocks | Can Parallel With |
|---|---|---|---|
| WS1: Core Infra | None | All | None (first) |
| WS2: State Detection | WS1 | WS3 | WS4, WS5, WS6 |
| WS3: Decision Engine | WS1, WS2 | WS4 | WS6 |
| WS4: Action & Execution | WS1, WS3 | WS5 | WS2 (Week 1), WS6 |
| WS5: Project Mgmt | WS1 | None | WS2, WS3, WS4, WS6 |
| WS6: Logging | WS1 | None | All others |
| WS7: Testing | All features | None | WS8 (Week 4) |
| WS8: Documentation | All features | None | WS7 (Week 4) |
End of Week 1:
- Integration Point: Basic watchdog works end-to-end
- Components: WS1 + WS2 + WS3 + WS4 + WS5 + WS6
- Validation: Can monitor 1 project and auto-continue on TODOs
End of Week 2:
- Integration Point: AI-powered decisions operational
- Components: WS3 (enhanced) + WS4 (skills) + WS8 (costs)
- Validation: Makes smart decisions using API and invokes skills
End of Week 3:
- Integration Point: Multi-project with Git operations
- Components: WS2 (multi-session) + WS4 (git) + WS5 (concurrent)
- Validation: Monitors 3+ projects, creates commits and PRs
End of Week 4:
- Integration Point: Production-ready release
- Components: WS7 (all tests) + WS8 (all docs)
- Validation: Can be deployed by any user, runs 24+ hours
Risk: Windows MCP unreliable Mitigation: Comprehensive error handling, retry logic, fallback mechanisms
Risk: UI state detection inaccurate Mitigation: Extensive testing with various scenarios, screenshot debugging
Risk: API costs too high Mitigation: Strict budget limits, fallback to rules, cost tracking
Risk: Commands fail to send Mitigation: Multiple retries, verification, detailed logging
Risk: Projects interfere with each other Mitigation: Strong isolation, separate state files, error quarantine
Risk: Logs grow too large Mitigation: Log rotation, size limits, archival strategy
Risk: Insufficient test coverage Mitigation: 80% coverage target, code review, integration tests
Risk: Documentation becomes outdated Mitigation: Write docs incrementally, review at end of each sprint
- What did you complete yesterday?
- What will you work on today?
- Any blockers or dependencies?
- Any integration points needed?
- Merge all workstream branches
- Run full integration tests
- Review progress against sprint goals
- Plan next week's parallelization
WS1 → All: Must complete before any major work starts WS2 → WS3: State classification needed for decisions WS3 → WS4: Decisions needed before execution WS4 → WS5: Actions needed for orchestration All → WS7: Features needed for testing All → WS8: Complete system needed for documentation
| Week | WS1 | WS2 | WS3 | WS4 | WS5 | WS6 | WS7 | WS8 | Total |
|---|---|---|---|---|---|---|---|---|---|
| 1 | 7h | 4h | 3h | 4h | 7h | 2h | 3h | 0h | 30h |
| 2 | 0h | 3h | 11h | 4h | 0h | 2h | 3h | 3h | 26h |
| 3 | 0h | 4h | 0h | 13h | 7h | 3h | 4h | 0h | 31h |
| 4 | 0h | 0h | 0h | 0h | 0h | 0h | 14h | 16h | 30h |
| Total | 7h | 11h | 14h | 21h | 14h | 7h | 24h | 19h | 117h |
Track completion daily:
- Green: On track
- Yellow: Slight delay (<1 day)
- Red: Blocked or >1 day delayed
- All modules importable
- Windows MCP calls successful
- Installation works on clean system
- 98%+ state classification accuracy
- Detects all active sessions
- Handles edge cases
- API decisions work correctly
- Fallback to rules works
- Costs within budget
- 95%+ command send success
- Skills invoked correctly
- Git operations complete
- Manages 5+ projects simultaneously
- Recovery works after crashes
- Loop runs 24+ hours
- All decisions logged
- Notifications arrive
- Logs readable and useful
- 80%+ code coverage
- All integration tests pass
- No critical bugs
- User can install in <30 min
- Developers understand architecture
- Release package complete
At end of each sprint, review:
- What went well in each workstream?
- What blockers occurred?
- How well did parallel work work?
- What would we change next time?
- Any technical debt created?
Created: 2024-11-22 Last Updated: 2024-11-22 Version: 1.0