Agile Backlog MCP

# Example Slash Commands for Agile MCP Workflows ## Overview These are example slash commands you can add to your Claude Code configuration (`.claude/commands/`) to streamline common agile workflows. --- ## Developer Slash Commands ### `/start-feature` - Begin work on a new feature **File:** `.claude/commands/start-feature.md` ```markdown You are starting work on a new feature. Follow this workflow: 1. Ask the user for: - Feature name/description - Project identifier (or list available projects) - Priority (high/medium/low) - Estimated story points (if known) 2. Create the story using the Agile MCP: - Tool: create_story - Include: title, description, priority, acceptance criteria 3. Ask the user to break down implementation steps 4. Create tasks for each step: - Tool: create_task (for each implementation task) 5. Mark the first task as "in_progress" 6. Provide a summary showing: - Story created (with ID) - Tasks created (with IDs) - What you'll work on first Then proceed to implement the first task. **Output format:** ``` ✓ Created Story #123: [Feature Name] - Priority: [X] - Estimated: [Y] points ✓ Created Tasks: 1. Task #1: [name] (in_progress) 2. Task #2: [name] 3. Task #3: [name] Starting implementation of Task #1... ``` ``` --- ### `/complete-task` - Mark current task done and move to next **File:** `.claude/commands/complete-task.md` ```markdown You are completing the current task and moving to the next one. 1. Ask which task was just completed (or infer from context) 2. Update the task to "done": - Tool: update_task - status: "done" 3. Ask if there are any technical notes to add: - If yes: create_note with markdown content - If no: continue 4. Check for remaining tasks: - Tool: list_tasks(story_id: X, status: "todo") 5. If tasks remain: - Mark next task as "in_progress" - Summarize what's next 6. If no tasks remain: - Ask if story is complete - If yes: update_story(status: "done") - Celebrate completion! **Output format:** ``` ✓ Task #45 marked as done ✓ Added technical note Next up: Task #46 - [task name] [Brief description of what this task involves] Ready to proceed? ``` ``` --- ### `/add-tech-note` - Document technical decision **File:** `.claude/commands/add-tech-note.md` ```markdown You are documenting a technical decision or discovery. 1. Ask the user: - What entity type (story/task) - Entity ID (or infer from current work) - What to document 2. Create a well-formatted markdown note: - Tool: create_note - parent_type: "story" or "task" - parent_id: <ID> - content: Use markdown formatting 3. Structure the note with: - ## Heading for the topic - **Decision:** What was decided - **Rationale:** Why this choice - **Alternatives considered:** What else was evaluated - **Impact:** How this affects the code **Example note format:** ```markdown ## Authentication Library Choice **Decision:** Using `jsonwebtoken` v9.0.0 for JWT handling **Rationale:** - Industry standard with 25k+ stars - Active maintenance and security updates - Simple API matching our needs **Alternatives considered:** - jose: More modern but smaller community - passport-jwt: Too heavyweight for our use case **Impact:** - Added dependency (23kb gzipped) - Standardized auth across all services ``` ``` --- ### `/check-blockers` - Identify and report blockers **File:** `.claude/commands/check-blockers.md` ```markdown You are checking for blocked work in the current project. 1. Get project identifier from context or ask user 2. Check for blocked stories: - Tool: list_stories(status: "blocked") 3. Check for blocked tasks: - Tool: list_tasks(status: "blocked") 4. For each blocked item: - Get details: get_story() or get_task() - Check notes for blocker description - Check dependencies: list_dependencies() 5. Generate a blocker report: **Output format:** ``` ## Blocker Report ### Blocked Stories (2) **Story #123: User Authentication** - Blocked by: Story #120 (Database Migration) - Status: Database schema not finalized - Impact: Cannot implement user table **Story #145: Payment Integration** - Blocked by: External API access - Status: Waiting for vendor API keys - Impact: Cannot test payment flow ### Blocked Tasks (1) **Task #67: Deploy to staging** - Blocked by: CI/CD pipeline setup - Parent Story: #123 ### Recommendations 1. Prioritize Story #120 to unblock authentication 2. Follow up with vendor for API keys 3. Assign DevOps resource to CI/CD setup ``` ``` --- ## Project Manager Slash Commands ### `/plan-sprint` - Interactive sprint planning **File:** `.claude/commands/plan-sprint.md` ```markdown You are a Project Manager planning the next sprint. 1. Get project identifier from user 2. Analyze team velocity: - Tool: get_velocity_report() - Calculate average from last 3 sprints - Present velocity data to user 3. Review backlog: - Tool: list_stories(status: "todo") - Filter by priority: high first - Show stories with estimates 4. Ask user for sprint details: - Sprint name (e.g., "Sprint 24") - Start and end dates - Sprint goal (1 sentence) - Desired capacity (suggest based on velocity) 5. Create the sprint: - Tool: create_sprint() 6. Recommend stories based on: - Priority - Estimates fitting in capacity - No blocking dependencies - Alignment with sprint goal 7. For each recommended story, ask for confirmation 8. Add confirmed stories to sprint: - Tool: add_story_to_sprint() 9. Show sprint summary: - Total points committed - Capacity utilization % - List of stories 10. Ask if ready to start sprint: - If yes: start_sprint() - If no: allow adjustments **Output format:** ``` ## Sprint 24 Planning ### Velocity Analysis - Last 3 sprints: 42, 38, 45 points - Average: 41.7 points - Recommended capacity: 40 points (96%) ### Proposed Stories (38 points) 1. ✓ Story #123: User Dashboard (13 pts) - High priority 2. ✓ Story #124: Dark Mode (8 pts) - High priority 3. ✓ Story #130: Notifications (13 pts) - Medium priority 4. ? Story #135: Mobile Layout (21 pts) - Too large, recommend splitting ### Sprint Summary - **Goal:** Improve user experience with dashboard and dark mode - **Capacity:** 40 points - **Committed:** 34 points (85% utilization) - **Buffer:** 6 points for unknowns Ready to start Sprint 24? (yes/no) ``` ``` --- ### `/daily-standup` - Generate standup report **File:** `.claude/commands/daily-standup.md` ```markdown You are generating a daily standup report for the active sprint. 1. Get project identifier 2. Find active sprint: - Tool: list_sprints(status: "active") - If multiple, ask which one 3. Get burndown data: - Tool: get_sprint_burndown() - Calculate: completed points, remaining points - Compare actual vs. ideal burndown 4. Check for blockers: - Tool: list_stories(status: "blocked") - Tool: list_tasks(status: "blocked") 5. Get sprint details: - Tool: get_sprint() - Show story status breakdown 6. Generate report: **Output format:** ``` ## Daily Standup - [Date] ### Sprint 24 Progress 📊 **Burndown Status** - Total Points: 40 - Completed: 18 (45%) - Remaining: 22 (55%) - Days Remaining: 6 of 10 **Trend:** ⚠️ Slightly behind (should be at 20 pts by now) ### Story Status - ✅ Done: 2 stories (18 points) - 🔄 In Progress: 3 stories (15 points) - ⏸️ Todo: 2 stories (7 points) ### Blockers 🚨 - Story #145: Payment Integration - Waiting for API keys from vendor - Impact: 1 developer idle ### Recommendations - Consider pulling Story #150 (5 pts) from backlog to keep team busy - Follow up with vendor on API keys today - Team pace needs to increase slightly to hit sprint goal ``` ``` --- ### `/sprint-retro` - Complete sprint and retrospective **File:** `.claude/commands/sprint-retro.md` ```markdown You are facilitating a sprint retrospective. 1. Get project identifier and sprint ID (or find active sprint) 2. Complete the sprint: - Tool: complete_sprint() - This generates final metrics 3. Get sprint details: - Tool: get_sprint() - Show final status of all stories 4. Get velocity data: - Tool: get_velocity_report() - Compare this sprint to historical average 5. Ask the team for retrospective feedback: - What went well? - What could be improved? - Action items for next sprint? 6. Document retrospective: - Tool: create_note - parent_type: "sprint" - parent_id: <sprint_id> - content: Structured retrospective notes 7. Handle incomplete work: - List stories still in_progress - Ask if they should move to next sprint or backlog 8. Generate sprint summary: **Output format:** ``` ## Sprint 24 Retrospective ### Final Results - **Committed:** 40 points - **Completed:** 35 points - **Completion Rate:** 87.5% - **Stories Done:** 5 of 7 ### Velocity Trend - Sprint 22: 45 pts - Sprint 23: 38 pts - Sprint 24: 35 pts ⬇️ - Average: 39.3 pts ### Retrospective Notes #### What Went Well ✅ - [Team feedback] - [Team feedback] #### What to Improve 📈 - [Team feedback] - [Team feedback] #### Action Items 🎯 - [ ] [Action item 1] - [ ] [Action item 2] ### Incomplete Work - Story #145: Payment Integration (8 pts) - 80% complete - Recommendation: Move to Sprint 25 **Sprint 24 marked as complete. Ready to plan Sprint 25?** ``` ``` --- ### `/backlog-health` - Assess backlog health **File:** `.claude/commands/backlog-health.md` ```markdown You are assessing the health of the product backlog. 1. Get project identifier 2. Gather backlog metrics: - Tool: list_stories() - Count by status - Count by priority - Identify stories missing estimates - Identify stories missing acceptance criteria 3. Check epics: - Tool: list_epics() - Find stories without epic (orphaned) 4. Check dependencies: - Tool: list_dependencies() - Identify complex dependency chains 5. Calculate health score: - Stories with estimates: +1 point each - Stories with acceptance criteria: +1 point each - Stories in epics: +1 point each - Score = (total points / possible points) * 100 6. Generate health report: **Output format:** ``` ## Backlog Health Report ### Overall Health: 🟡 Yellow (68%) ### Metrics - **Total Stories:** 45 - Todo: 30 - In Progress: 8 - Done: 7 - **By Priority:** - Critical: 2 - High: 12 - Medium: 20 - Low: 11 ### Issues Found ❌ **Missing Estimates** (12 stories) - Story #150: User Settings - Story #151: Email Templates - [... 10 more] ⚠️ **Missing Acceptance Criteria** (8 stories) - Story #145: Payment Integration - Story #148: Search Feature - [... 6 more] ℹ️ **Orphaned Stories** (5 stories) - Not assigned to any epic - Recommend creating thematic epics ### Recommendations 1. **Immediate Actions:** - Add estimates to top 12 stories (enables sprint planning) - Define acceptance criteria for high-priority work 2. **Backlog Grooming:** - Create epics for orphaned stories - Archive low-priority stories older than 6 months 3. **Target Health:** 🟢 Green (>85%) - Requires: Estimating 12 stories + criteria for 8 stories **Schedule backlog grooming session?** ``` ``` --- ## Utility Slash Commands ### `/export-docs` - Export backlog to markdown **File:** `.claude/commands/export-docs.md` ```markdown You are exporting the project backlog to markdown documentation. 1. Get project identifier 2. Export backlog: - Tool: export_backlog() - This creates `.agile-backlog/` directory with markdown files 3. Explain the export structure: **Output format:** ``` ✓ Backlog exported to .agile-backlog/ ### Structure Created: ``` .agile-backlog/ ├── backlog.json # Raw JSON export ├── epics/ │ ├── user-authentication/ │ │ ├── epic.md # Epic overview │ │ └── features/ │ │ ├── login.md # Story details │ │ └── logout.md │ └── payment-system/ │ ├── epic.md │ └── features/ └── orphan-stories/ └── standalone-feature.md ``` ### What's Included: - ✓ All epics with descriptions - ✓ All stories with tasks - ✓ Acceptance criteria - ✓ Story points and priorities - ✓ Dependencies - ✓ All notes (technical decisions, etc.) ### Use Cases: - Share with stakeholders - Version control your backlog - Generate external documentation - Review historical work Files are now in `.agile-backlog/` - would you like me to open any specific files? ``` ``` --- ### `/show-dependencies` - Visualize dependency graph **File:** `.claude/commands/show-dependencies.md` ```markdown You are visualizing the dependency graph for the project. 1. Get project identifier 2. Get all dependencies: - Tool: list_dependencies() 3. Get all stories: - Tool: list_stories() 4. Build dependency graph: - Map which stories block which - Identify critical path (longest chain) - Find independent work 5. Present in text format: **Output format:** ``` ## Dependency Graph ### Critical Path (20 points, ~2 sprints) ``` Story #120: Database Schema (8 pts) ↓ blocks Story #123: User Authentication (5 pts) ↓ blocks Story #125: User Profile (5 pts) ↓ blocks Story #130: Settings Page (2 pts) ``` ### Parallel Work (can start immediately) - Story #140: Email Service (8 pts) - No dependencies - Story #141: UI Components (13 pts) - No dependencies - Story #145: Documentation (3 pts) - No dependencies ### Blocked Work (cannot start yet) - Story #150: Admin Panel - Blocked by: #123 (User Authentication) - Can start after: Sprint 24 ### Recommendations 1. **Prioritize critical path** - Start #120 immediately 2. **Parallelize independent work** - Assign to different team members 3. **Plan ahead** - Story #150 ready for Sprint 25 **Would you like me to create a sprint plan based on this dependency analysis?** ``` ``` --- ## Installation Instructions To use these slash commands: 1. Create `.claude/commands/` directory in your project 2. Create individual `.md` files for each command 3. Copy the command content (without the outer markdown code fence) 4. Use in Claude Code with `/command-name` Example: ```bash mkdir -p .claude/commands cat > .claude/commands/start-feature.md << 'EOF' You are starting work on a new feature. Follow this workflow: ... EOF ``` --- ## Customization You can customize these commands by: - Changing output format - Adding project-specific defaults - Combining multiple commands - Adding team-specific workflows --- ## Command Chaining You can chain commands for common workflows: ```markdown # Example: Full feature implementation workflow User: "Let's build user authentication" Claude: 1. /start-feature (creates story + tasks) 2. [Implements code] 3. /add-tech-note (documents decisions) 4. /complete-task (marks tasks done) 5. [Repeat for each task] 6. Story automatically marked done when all tasks complete ``` --- ## Summary These slash commands make agile workflows **effortless**: - **Developers:** `/start-feature`, `/complete-task`, `/add-tech-note` - **PM:** `/plan-sprint`, `/daily-standup`, `/sprint-retro`, `/backlog-health` - **Utilities:** `/export-docs`, `/show-dependencies` Copy, customize, and use them to streamline your agile process!