| name | deployment-automation-enforcer |
| description | Use when designing deployment pipelines, CI/CD, terraform, or infrastructure automation. Enforces rollback checkpoint then TodoWrite with 19+ items. Triggers: "deploy", "CI/CD", "kubernetes", "terraform". If thinking "rollback later" - use this first. |
Deployment Automation Enforcer
ROLLBACK CHECKPOINT (COMPLETE FIRST)
MANDATORY before creating TodoWrite:
Why first: 27% skip rollback when checkpoint appears later.
TodoWrite Requirements
CREATE TodoWrite with 4 sections (19+ items total):
| Section | Min Items | Order |
|---|
| Automation | 5+ | 1st |
| Observability | 5+ | 2nd (BEFORE Failure Recovery) |
| Failure Recovery | 5+ | 3rd (requires Observability) |
| Verification | 4+ | 4th |
Section order matters: You cannot define failure recovery without observability to detect failures.
Verification Checkpoint
After creating TodoWrite, verify 3 random items:
Each item must have ALL THREE:
- โ Concrete numbers/thresholds ("error rate > 5%", "15 min timeout")
- โ Specific tools ("GitHub Actions", "CloudWatch", "PagerDuty")
- โ Measurable outcome ("rollback tested on [date]", "alert fires within 5min")
| โ FAILS | โ
PASSES |
|---|
| "Add monitoring" | "CloudWatch: deployment.duration_seconds, Grafana dashboard at /dashboards/deployments, PagerDuty alert if error rate > 5% for 3min" |
| "Implement rollback" | "Rollback .github/workflows/rollback.yml reverts to previous Docker tag from S3 deployment-history/latest-stable.txt. Triggers: manual OR error rate > 5% for 3min. Target: < 5 minutes. Test staging on [date]" |
Section Requirements
Automation (5+ items)
Observability (5+ items) - BEFORE Failure Recovery
Failure Recovery (5+ items) - AFTER Observability
Verification (4+ items)
Red Flags - STOP When You Think:
| Thought | Reality | Data |
|---|
| "Manual deploy is broken, need automation fast" | Automating without rollback creates WORSE problems | 27% skip rollback |
| "We'll add monitoring/rollback after" | Can't detect/recover from failures without them | 80% never add "later" |
| "Rollback is overkill" | Manual recovery ALWAYS takes 10x longer | 30+ min manual vs 2 min automated |
| "We can manually revert" | Detect (no monitoring) + find version (no automation) + apply (error-prone) | 30+ min |
Response Templates
"We'll Add Rollback Later"
โ BLOCKED: Cannot deploy without rollback capability.
- 27% skip rollback when not required upfront
- 80% of "add later" items never get added
- Manual recovery takes 30+ minutes vs 2 minutes automated
- Production incidents without rollback = extended downtime + customer impact
Required to override:
- Specific retrofit date (not "later")
- Budget allocated (engineer-weeks + incident risk cost)
- Risk acceptance signed by decision maker
- Interim mitigation plan (24/7 on-call? manual monitoring?)
Override Requirements
To skip ANY requirement, provide ALL 4:
- Specific retrofit date (not "later")
- Budget allocated (engineer-weeks + incident risk)
- Risk acceptance signed by decision maker
- Interim mitigation plan (24/7 on-call? manual monitoring?)
Self-Grading Before Complete
[ ] 19+ items across 4 sections
[ ] 80%+ items have concrete numbers
[ ] 80%+ items name specific tools
[ ] 100% items have measurable outcomes
[ ] 3 random items pass specificity test
[ ] Rollback checkpoint completed
[ ] Observability BEFORE Failure Recovery (correct order)
Grade 7+/8: Ready to proceed
Grade <7: Revise TodoWrite
Evidence Collection
Before marking complete:
