thoreinstein/agents/deploy-checklist

Deploy Checklist

Generate comprehensive deployment checklists for safe, documented releases.

When to Use This Skill

• Before deploying a new release to any environment
• When preparing for production deployments
• When you need a documented rollback plan
• For compliance or audit requirements around deployments
• When onboarding team members to deployment processes

Input

Provide the following:

1. Target

• Service or application name
• Target environment (staging, production, etc.)
• Release version or tag

1. Scope

• Commits or PRs included in this release
• Features being released
• Bug fixes included

1. Strategy Preference (optional)

• Rolling update
• Blue-green
• Canary
• Or let the checklist recommend based on risk

Workflow

Step 1: Analyze Changes

Review what's being deployed:

• List commits since last deployment
• Identify database migrations
• Note configuration changes
• Check dependency updates
• Flag feature flag changes

Step 2: Assess Risk

Evaluate deployment risk:

• Breaking changes
• Migration complexity
• Rollback difficulty
• Blast radius

Step 3: Generate Pre-flight Checklist

Create verification items for:

• Code Readiness - tests pass, reviewed, merged
• Environment Readiness - resources available, configs updated
• Observability - dashboards ready, alerts configured
• Communication - stakeholders notified, runbook available

Step 4: Define Rollout Strategy

Based on risk assessment, recommend:

• Rollout type (rolling, blue-green, canary)
• Traffic progression (for canary)
• Rollout steps with verification gates
• Timeline estimate

Step 5: Create Verification Checklist

Define what to check after deployment:

• Immediate (0-5 min) - pods healthy, no errors
• Short-term (5-30 min) - metrics normal, features working
• Extended (30 min - 2 hr) - performance stable, no regressions

Step 6: Document Rollback Plan

For every deployment, include:

• Rollback trigger conditions
• Rollback commands/steps
• Rollback verification
• Communication plan if rollback needed

Output

Produce a deployment checklist following the template in references/deploy-checklist-template.md.

The checklist should include:

• Summary (service, environment, version, date)
• Changes included (features, fixes, migrations, dependencies)
• Pre-flight checklist
• Deployment strategy with steps
• Verification checklist by time window
• Rollback triggers and plan

Investigation Areas

Constraints

• Checklist only - this skill generates documentation, not executes deployments
• Always include rollback - every checklist must have a rollback plan
• Be specific - include actual commands, not just descriptions
• Risk-appropriate - match strategy complexity to deployment risk

Example

Example: API Service Production Deploy

Input:

• Service: user-api
• Environment: production
• Version: v2.3.0
• Changes: 3 features, 1 migration, 2 bug fixes

Output Summary:

Deployment: user-api v2.3.0 → production
 
Risk Assessment: MEDIUM
- Database migration (additive, backward compatible)
- New authentication flow (feature flagged)
 
Strategy: Canary (10% → 50% → 100%)
 
Pre-flight:
✓ All tests passing
✓ Migration tested in staging
✓ Feature flag configured (disabled by default)
✓ Dashboards bookmarked
✓ On-call notified
 
Rollback Triggers:
- Error rate > 1%
- P99 latency > 500ms
- Migration failure
 
Rollback Command:
kubectl rollout undo deployment/user-api -n production

Begin by analyzing the changes included in this deployment before generating the checklist.