/retro

Automated retrospective generator that analyzes project data to surface insights, patterns, and actionable improvements using the Start/Stop/Continue format.

Quick Start

Copy/agileflow:retro

Generates a retrospective for the last 2 weeks using project data.

Parameters

Examples

Standard 2-Week Retrospective

Copy/agileflow:retro

Analyzes the last 2 weeks with ASCII-formatted output including:

• Sprint summary (stories, velocity, cycle time)
• What went well (Continue section)
• What needs improvement (Start/Stop sections)
• Actionable items by priority
• Team contributions
• Predictions for next sprint

Last 30 Days Retrospective

Copy/agileflow:retro TIMEFRAME=30d FORMAT=markdown

Analyzes the last 30 days and saves as markdown for easy sharing.

Epic-Specific Retrospective

Copy/agileflow:retro EPIC=EP-0010 FORMAT=markdown

Focuses retrospective analysis on a specific epic's completion with:

• Epic success metrics
• What went well during epic execution
• Challenges encountered
• Learnings for future epics

Export Without Saving

Copy/agileflow:retro SAVE=false FORMAT=markdown

Generates and displays retrospective without saving to disk (useful for quick analysis).

Output

ASCII Format (Default)

Creates a formatted report with Unicode box drawing:

Copy╔════════════════════════════════════════════════════════════════╗
║                  AGILEFLOW RETROSPECTIVE                        ║
║                  Sprint: Oct 17 - Oct 31 (2 weeks)              ║
╠════════════════════════════════════════════════════════════════╣
║                                                                 ║
║  📊 SPRINT SUMMARY                                              ║
║  Stories Completed:   17 (85% of planned 20)                   ║
║  Velocity:            8.5 stories/week (↗ +12%)                ║
║  Cycle Time:          3.2 days avg (↓ -8%)                     ║
║                                                                 ║
║  ✅ WHAT WENT WELL (Continue)                                   ║
║  1. Velocity increased +12% vs previous sprint                 ║
║  2. Completed EP-0010 (Authentication) on schedule             ║
║  3. 12 stories completed in \<2 days                            ║
║                                                                 ║
║  ⚠️ WHAT NEEDS IMPROVEMENT                                      ║
║  1. Average review time: 2.5 days (up from 1.8)                ║
║     🛑 STOP: Letting PRs sit unreviewed                         ║
║     ▶️ START: Daily 30min review slot                           ║
║                                                                 ║
║  🎯 ACTION ITEMS                                                ║
║  HIGH Priority:                                                ║
║  [ ] Unblock US-0045 (escalate for API keys)                   ║
║  [ ] Set up daily review time block (30min)                    ║
║                                                                 ║
║  🔮 PREDICTIONS FOR NEXT SPRINT                                 ║
║  Based on current velocity:                                    ║
║  EP-0011 (Payment): 40% → 75% (6 stories)                      ║
╚════════════════════════════════════════════════════════════════╝

Saved Files

Automatically saves to:

Copydocs/08-project/retrospectives/retro-<YYYYMMDD>.md
docs/08-project/retrospectives/README.md (updated with index)

Data Sources

Analyzes data from 4 sources:

1. docs/09-agents/bus/log.jsonl - Event patterns, status transitions, blockers
2. docs/09-agents/status.json - Current state, WIP levels, ownership
3. docs/06-stories//US-*.md** - Completion data, estimates vs actuals
4. Velocity data - Points completed, throughput trends

Retrospective Structure

Continue (What Went Well)

Patterns identified:

• High velocity increases
• Fast cycle times (under 2 days)
• Zero blocked stories
• Epic completions
• Good estimation accuracy (under 20% variance)
• Balanced agent workload

Start (New Practices)

Recommendations for:

• Daily standups to identify blockers earlier
• Breaking down stories into smaller chunks
• Finishing stories before starting new ones
• Pre-sprint dependency checks
• Dedicated review time blocks
• Planning poker for estimation
• Cross-training on bottleneck areas

Stop (Things to Eliminate)

Anti-patterns detected:

• Letting PRs sit unreviewed
• Starting stories with external dependencies
• Estimating without understanding complexity
• Over-committing in sprint planning
• Context switching between stories
• Large PRs (>400 lines)
• Assigning all work to one agent

Analysis Insights

Pattern Detection

The command identifies:

• Recurring blockers - Stories blocked multiple times
• Day-of-week patterns - When most work completes
• Handoff patterns - Knowledge silos or unclear ownership
• Story size correlations - What sizes are most efficient
• Velocity trends - Improving or declining productivity

Celebration Moments

Highlights wins like:

• Epic completions
• Velocity milestones (double-digit stories/week)
• Zero bugs reported
• Fast deliveries (under 4 hours)
• 100% test coverage achievements

Workflow

1. Load data sources (bus logs, status, story data)
2. Analyze what went well (velocity, cycle time, completion)
3. Analyze what needs improvement (drops, long cycles, WIP)
4. Detect advanced patterns
5. Generate prioritized action items
6. Create celebration section
7. Generate comprehensive report
8. Save to retrospectives directory
9. Update index with latest entry

Use When

• Sprint boundaries - After each sprint or iteration
• Epic completions - Retrospective on the full epic
• Team reviews - Monthly or quarterly check-ins
• Process improvement - Identifying patterns over time
• Before planning - Apply learnings to next sprint
• Metrics context - Understand what numbers mean

Integration

Works well with:

• /metrics - Contextualize metrics data
• /velocity - Analyze velocity trends
• /babysit - Suggests retro at sprint boundaries
• /status - References current story status

Related Commands

• /metrics - Detailed project metrics
• /velocity - Velocity tracking and trends
• /update - Stakeholder progress reports
• /feedback - Collect team feedback on stories