- New to the project? → CLAUDE.md for setup and overview
- Looking for strategic vision? → ROADMAP.md
- Looking for implementation details? → CLAUDE.md - Task Navigation
- Looking for completed features? → CHANGELOG.md
For AI Agents: Understand current priorities and context before implementing features. You'll create your own task-level breakdown using TodoWrite.
For Humans: High-level status overview bridging ROADMAP strategy to execution tools (GitHub Issues, Jira, etc.).
PROJECT_PLAN.md focuses on what and when:
- What are we building/doing right now?
- When is this happening? (now / next / later)
- Who is responsible? (optional, useful for delegation)
This is NOT:
- Task-level tracking (GitHub Issues/Jira handle this)
- Strategic vision (see ROADMAP.md)
- Implementation details (see CLAUDE.md)
Timeframe: Current focus + next 1-2 months, reviewed weekly or biweekly
Theme: [What we're focused on - e.g., "Baseline Model Development" or "Multi-tenant Foundation"]
Strategic Alignment: Supports [ROADMAP milestone - e.g., "Q1 2025: Establish ML pipeline foundation"]
1. [Initiative/Feature Name]
- What: [Brief description of what this is - e.g., "Data preprocessing pipeline with feature engineering"]
- Why: [Business/technical reason - e.g., "Required for baseline model training"]
- Who: [Team/Person/Agent - optional, e.g., "Data team" or "TBD"]
- Status: [Phase or % - e.g., "70% complete - validation in progress"]
- Tracking: GitHub Issues #45-52 or Jira Epic
- Next milestone: [What's coming next - e.g., "Complete feature validation by end of month"]
2. [Initiative/Feature Name]
- What: [Description]
- Why: [Reason]
- Who: [Assignee - optional]
- Status: [Status]
- Tracking: [Link to detailed work]
- Blocker: [If applicable - e.g., "Waiting on dataset access approval"]
1. [Initiative/Feature Name]
- What: [Description]
- Why: [Reason]
- Who: [Assignee - optional]
- Depends on: [Prerequisites - e.g., "Completing data pipeline work above"]
- Target start: [Timeframe - e.g., "Early February" or "After current sprint"]
2. [Initiative/Feature Name]
- What: [Description]
- Why: [Reason]
- Decision needed: [If applicable - e.g., "Model architecture selection - awaiting experiments"]
- [High-level item 1 - e.g., "Model deployment pipeline"]
- [High-level item 2 - e.g., "A/B testing framework"]
🚧 [Blocker Name - e.g., "Production Dataset Access"]
- Blocks: [What's affected - e.g., "Model training with real data"]
- Owner: [Who's unblocking - e.g., "Compliance team"]
- ETA: [When resolved - e.g., "February 1"]
- Workaround: [What we're doing meanwhile - e.g., "Training on synthetic data"]
🚧 [Blocker Name]
- Blocks: [Impact]
- Owner: [Responsible party]
- ETA: [Timeline]
✅ [Accomplishment - e.g., "Exploratory Data Analysis"]
- Completed: [When - e.g., "Mid-January"]
- Outcome: [What was delivered - e.g., "Identified 15 key features, documented in notebook"]
- Documentation: [Where documented - e.g., "notebooks/eda/initial_analysis.ipynb" or CHANGELOG.md]
✅ [Accomplishment]
- Completed: [When]
- Outcome: [What was delivered]
- Documentation: [Link]
Detailed Task Tracking:
Progress & Experiments (if applicable):
Related Documentation:
- ROADMAP.md - Strategic vision
- CLAUDE.md - Implementation guidance
- CHANGELOG.md - Feature history
When you're asked to implement a feature:
- Read this document to understand current focus and alignment
- Check blockers to avoid working on blocked items
- Use TodoWrite to create your own task-level breakdown
- Work incrementally through your tasks, marking complete as you go
- Update documentation when your work is done (CHANGELOG, CLAUDE.md if architecture changed)
Your task lists are ephemeral (conversation-level). This document stays high-level.
Last Updated: [YYYY-MM-DD] Review Cadence: [e.g., "Weekly on Mondays" or "Biweekly"] Current Period: [e.g., "January 2025" or "Sprint 12"]