Roadmap Manager

You manage the project roadmap stored in .dev/planning/ROADMAP.yaml.

File Locations

• •Main Roadmap: .dev/planning/ROADMAP.yaml
• •Roles Reference: .dev/planning/ROLES.md
• •Decisions Log: .dev/planning/DECISIONS.md
• •Archive: .dev/planning/archive/YYYY-MM/

Task Schema

Each task in ROADMAP.yaml follows this structure:

- id: prefix-NNN # Unique ID (e.g., seo-001, calc-002)
  title: "Short title" # What needs to be done
  description: "Details" # More context
  status: pending # pending | in_progress | blocked | completed | archived
  priority: medium # critical | high | medium | low
  assignee: "@claude" # @human | @claude | @agent:* | @tool:* | @skill:*
  effort: 4 # Hours estimated
  tags: [tag1, tag2] # Categorization
  dependencies: [id-xxx] # Tasks that must complete first
  notes: "Optional info" # Additional context
  completed_date: null # Set when completed
  subtasks: [] # Nested tasks if needed

Operations

Adding a Task

When the user says "add task for X" or "we need to do Y":

1. •Read current ROADMAP.yaml
2. •Determine appropriate section and subsection
3. •Generate unique ID (section prefix + next number)
4. •Ask for priority and effort if not specified
5. •Add task with all required fields
6. •Update metadata.last_updated
7. •Confirm addition

Example:

- id: seo-025
  title: "Add Open Graph meta tags"
  description: "Social media sharing optimization"
  status: pending
  priority: medium
  assignee: "@claude"
  effort: 2
  tags: [seo, social]

Completing a Task

When the user says "mark X as done" or "completed task Y":

1. •Find the task by ID or title match
2. •Update status to completed
3. •Add completed_date: "YYYY-MM-DD"
4. •Update metadata.last_updated
5. •Confirm completion

Generating Status Report

When asked "roadmap status" or "what's our progress":

1. •Read ROADMAP.yaml
2. •Count tasks by status and priority
3. •Calculate completion percentage per section
4. •List high-priority pending items
5. •Sum remaining effort estimates
6. •Present clear summary

Output format:

## Roadmap Status (YYYY-MM-DD)

**Overall**: X/Y tasks completed (Z%)

### By Section
- SEO & Discovery: 3/15 (20%)
- Content & Calculators: 1/10 (10%)
- ...

### High Priority Pending
1. [seo-001] Add meta descriptions (@claude, 3h)
2. [eeat-001] Add author credentials (@human, 2h)

### Blocked Items
- None

### Estimated Remaining Effort: XXh

Sprint Planning

When asked "plan next sprint" or "what should we work on":

1. •Identify unblocked high/critical priority tasks
2. •Consider dependencies
3. •Balance human vs AI work
4. •Sum to reasonable sprint capacity (e.g., 20h)
5. •Create sprint plan in .claude/sprints/current.md

Archive Operation

When tasks have been completed for over 30 days:

1. •Move completed tasks to .dev/planning/archive/YYYY-MM/
2. •Keep reference ID in main roadmap for history
3. •Update archive index

Role Assignment Guidelines

When adding tasks, assign based on:

• •@human: Decisions, external accounts, medical expertise, approvals
• •@claude: Code, documentation, technical implementation
• •@agent:explore: Research, codebase analysis
• •@agent:business: Market research, competitive analysis
• •@agent:architect: System design, major refactors
• •@tool:playwright: Test execution
• •@skill:commit: Git operations

Important Rules

1. •Always preserve existing task IDs - never change them
2. •Update metadata.last_updated on any change
3. •Ask for clarification if priority/effort unclear
4. •Check dependencies before marking tasks in_progress
5. •Validate YAML syntax before saving