majiayu000 vibe-production-mindset
The gap between "works on my machine" and "works in production" is where incidents live.
When to Use This Skill
- Starting any implementation that will run in production
- When implementing API endpoints, data processing, or user-facing features
- Before declaring a feature "done"
- When the user cares about production quality
When NOT to Use This Skill
- Prototypes or spikes explicitly marked as throwaway
- Internal scripts that run once
- Test code (different quality criteria)
- When the user says "just get it working"
The Checklist
Imagine this code serves 1 million users. Check:
1. Observability
- Structured logging at key decision points (not just errors)
- Tracing/correlation ID propagated through the request
- Metrics for key operations (latency, success/failure counts)
- Alerts defined for anomalies
2. Error Handling at System Boundaries
- External API calls have timeouts
- External API calls have retries with backoff
- File I/O handles "file not found" and "permission denied"
- Network errors are handled (connection refused, timeout, DNS failure)
- Database errors are handled (connection lost, constraint violation)
3. Input Validation
- All external input validated before processing
- Size limits on string/array inputs
- SQL injection prevention (parameterized queries)
- XSS prevention (output encoding)
- Authentication checked before authorization
4. Graceful Degradation
- Circuit breakers for external dependencies
- Fallback behavior when dependencies are unavailable
- Graceful shutdown (drain connections, finish in-flight)
- Health check endpoint
5. Resource Management
- Connections closed/returned to pool
- Goroutines/threads have lifecycle management
- Memory bounded (no unbounded growth)
- File handles closed
Steps
- Read the implementation
- Run through the checklist — not every item applies to every feature
- Flag gaps — only for items that DO apply
- Recommend fixes — specific, not generic
Output Format
Production Readiness: [Feature Name]
Overall: READY / NEEDS_WORK / NOT_READY
| Category | Status | Gaps |
|---|---|---|
| Observability | ✓/◐/✗ | [count] |
| Error Handling | ✓/◐/✗ | [count] |
| Input Validation | ✓/◐/✗ | [count] |
| Graceful Degradation | ✓/◐/✗ | [count] |
| Resource Management | ✓/◐/✗ | [count] |
Critical Gaps
- [Gap with specific file:line and fix suggestion]
Note
This is NOT about over-engineering. It's about not under-engineering the critical paths. A simple feature should have simple production hardening — logging, error handling, and input validation. Not every feature needs circuit breakers.