/retro — Eng Manager

What shipped. What broke. What to improve.

When to Use

• End of sprint or week
• After a major ship
• After an incident
• When you feel stuck

Process

Step 1: Gather Data

# What shipped this week/sprint
git log --oneline --since="1 week ago" --no-merges

# PRs merged
gh pr list --state merged --search "merged:>=$(date -v-7d +%Y-%m-%d)"

# Test health
{test_command} 2>&1 | tail -5

Step 2: Analyze

What went well?

• Features shipped on time
• Bugs caught before production
• Clean PRs (few review rounds)

What went wrong?

• Bugs that escaped to production
• Scope creep
• Blocked work
• Flaky tests

What to improve?

• Patterns that keep recurring
• Missing automation
• Knowledge gaps

Step 3: Report

## Retro: Week of {date}

### Shipped
- feat: notification triage (PR #42)
- fix: duplicate delivery bug (PR #43)
- docs: API reference update (PR #44)

### Metrics
- PRs merged: 3
- Tests added: 9
- Coverage: 83% → 87%
- Bugs escaped to prod: 0

### What went well
- /qa caught the duplicate delivery bug before ship
- Plan doc kept scope tight

### What to improve
- Queue worker has no monitoring — add /perf next sprint
- Test coverage on src/queue.ts still below target

### Action Items
- [ ] Add monitoring to queue worker
- [ ] Write integration tests for queue

Chaining

• Before: /ship (retro after shipping)
• After: /plan (retro findings inform next sprint)