// This skill should be used when the user asks to "create a deployment plan", "plan the release", "design the rollout strategy", "define rollback procedures", "plan the go-live", "create a release checklist", or needs to plan how a feature or service will be deployed to production.
| name | Deployment Plan |
| description | This skill should be used when the user asks to "create a deployment plan", "plan the release", "design the rollout strategy", "define rollback procedures", "plan the go-live", "create a release checklist", or needs to plan how a feature or service will be deployed to production. |
| version | 1.0.0 |
Produce a detailed deployment plan with rollout strategy, go/no-go criteria, rollback procedures, and a release checklist. Treat every production deployment as a risk management exercise.
A deployment plan is not an afterthought — it is the bridge between working code and a production service. This skill produces a plan that is executable by any engineer on the team, not only the person who wrote the code.
Select the appropriate strategy based on risk and system characteristics:
| Strategy | When to Use | Rollback Speed | Risk |
|---|---|---|---|
| Rolling update | Stateless services, low risk | Fast (redeploy) | Low |
| Blue-Green | Zero-downtime required, fast rollback needed | Instant (switch) | Low |
| Canary | High-traffic, gradual risk validation needed | Fast (redirect traffic) | Very Low |
| Feature flags | Business-controlled release, A/B testing | Instant (toggle) | Very Low |
| Big bang | Schema migrations, monolith, unavoidable downtime | Slow (manual rollback) | High |
For database schema changes: always use backwards-compatible migrations (expand/contract pattern).
# Deployment Plan: [Service/Feature Name]
Release Date: [YYYY-MM-DD] | Release Manager: [Role]
## Deployment Strategy
[Selected strategy and rationale]
## Pre-Deployment Checklist
- [ ] All acceptance criteria passing in staging
- [ ] Performance baseline validated in staging
- [ ] Security gate PASSED
- [ ] Database migrations tested and reversible
- [ ] Runbook reviewed by on-call engineer
- [ ] Rollback procedure tested in staging
- [ ] Monitoring dashboards prepared
- [ ] On-call engineer briefed and available
- [ ] Change record raised (if required)
## Deployment Steps
| Step | Action | Owner | Duration | Validation |
|---|---|---|---|---|
| 1 | [Run DB migration] | DevOps | 5 min | [verify row count] |
| 2 | [Deploy new container version] | DevOps | 3 min | [health check passes] |
| 3 | [Smoke test critical paths] | QA | 10 min | [checklist passes] |
| 4 | [Enable feature flag for 10% of traffic] | Product | 1 min | [error rate stable] |
| 5 | [Monitor for 30 minutes] | On-call | 30 min | [SLOs met] |
| 6 | [Enable for 100% of traffic] | Product | 1 min | [error rate stable] |
## Go/No-Go Criteria
**Go**: All pre-deployment checklist items checked AND error rate < [threshold] in first 10 minutes
**No-Go**: Any pre-deployment checklist item not met OR error rate > [threshold] at any point
## Rollback Procedure
| Trigger | Action | Owner | Duration |
|---|---|---|---|
| Error rate > [X]% | Disable feature flag / revert traffic | On-call | < 2 min |
| Service unavailable | Redeploy previous image | DevOps | < 5 min |
| Data corruption | Roll back DB migration | DBA | [estimate] |
## Post-Deployment Checklist
- [ ] SLOs met for 1 hour post-deployment
- [ ] No P0 or P1 incidents raised
- [ ] Deployment notes updated in runbook
- [ ] Feature flag removal task created (with date)
- [ ] Stakeholders notified of successful release
references/migration-patterns.md — Backwards-compatible database migration patternsassets/deployment-plan-template.md — Blank deployment plan templateThis skill should be used when the user asks to "use the VAMFI templates", "get the proposal template", "find the PRD template", "use the HLD template", "load the runbook template", "access artifact templates", or needs a blank template for any VAMFI consultancy document.
This skill should be used when the user asks to "shape stories", "create a backlog", "break this epic into stories", "write user stories with acceptance criteria", "define the product backlog", "estimate stories", or needs to convert a PRD or delivery plan into an executable sprint backlog.
This skill should be used when the user asks to "implement this feature", "write the code for this story", "execute this implementation plan", "make this code change", "add this functionality", or needs hands-on code implementation following a VAMFI implementation plan.
This skill should be used when the user asks to "review this code", "review the PR", "check code quality", "refactor this module", "improve this code", "run a code review", or needs systematic review of recently written or changed code before it merges.
This skill should be used when the user asks to "optimise costs", "reduce cloud spend", "improve performance", "fix slow queries", "analyse FinOps", "reduce latency", "optimise infrastructure", "right-size resources", or needs to improve the efficiency of a running system.
This skill should be used when the user asks to "assess the current architecture", "document the existing system", "analyse the codebase", "identify technical debt", "review the legacy system", "produce a current-state report", or needs to understand what exists before designing a target architecture.