Production Bug Pre-Mortem
Find code-backed production failure paths, rank their impact, and get targeted fixes, regression tests, and release tripwires.
Before & after
Same prompt, same agent — the only difference is the skill.
That's smart to think ahead! There could definitely be edge cases you haven't considered. You'll want to test your error handling, think about what happens under load, check your database queries for N+1 problems, and make sure logging is good. Authentication and permissions are common gotchas. You should probably load test it and review the code one more time. Also think about what happens if your dependencies go down or if data is malformed. Have you tested with realistic data volumes? I'd suggest going through the code and thinking about what could break. What parts of the system are you most concerned about?
PREDICTED PRODUCTION BUGS (pre-mortem analysis): CRITICAL (fix before ship) • Database connection pool exhaustion under spike traffic → query timeout cascades to 500s FIX: Set max_connections=50, add circuit breaker, test with k6 at 200 req/s • Payment webhook missing idempotency key → duplicate charges on retry FIX: Hash request body, cache responses for 24h, return 200 if already processed • Unindexed user_id on orders table → reports endpoint times out on enterprise accounts FIX: Add composite index (user_id, created_at), verify explain plan shows index_scan HIGH (handle gracefully before ship) • Missing CORS headers → browser requests from app.example.com fail silently FIX: Add Access-Control-Allow-Origin, test cross-origin via localhost:3000 • Null check missing on stripe.customer_email → crashes on legacy accounts FIX: Default to support@example.com, log event for manual review MEDIUM (monitor first week) • Timezone mismatch (server UTC, client local) → date-based reports show off-by-one FIX: Explicit serialize to ISO-8601, document client-side normalization Testing checklist: Load test, chaos inject dependency failures, replay production payloads.
About this skill
“The tests pass, but what breaks when this ships?” This pre-mortem traces how real production conditions can turn apparently correct code into duplicate work, lost data, stalled requests, or cascading failures. Each finding connects source evidence to a concrete failure sequence, a targeted fix, and a test that could expose it before release.
What it does
- Traces requests and jobs across database commits, queues, external calls, and acknowledgments.
- Reconstructs race conditions and crash windows with explicit execution sequences.
- Checks whether retries, timeouts, and duplicate delivery preserve business invariants.
- Examines rolling deployments, migrations, and rollback paths for incompatible intermediate states.
- Ranks supported risks by impact and exposure, keeping confidence separate.
- Proposes focused regression tests, detection signals, and containment measures.
What you get
A scoped production-risk report with source locations, failure triggers, incident sequences, fixes, and proposed verification. Unresolved assumptions and inspection gaps stay visible so hypothetical risks do not become false claims about your system.
Best for
Engineers preparing releases that touch background jobs, payments, shared state, external integrations, or database migrations, and teams investigating where their next recurring incident might originate.
SKILL.md preview
---
name: production-bug-pre-mortem
description: >-
Finds latent production bugs by tracing business invariants through concurrency, retries, partial failures, and deployment transitions. Produces evidence-backed failure scenarios with ranked impact, targeted fixes, and reproduction tests. Use when the user says "pre-mortem this release", "what could break in production?", "the tests pass but I'm worried", "review this webhook before we ship", "my jobs keep running twice", "check this migration for production risks", or "find the next incident before it happens".
---
# Production Bug Pre-Mortem
… (sign up to view the full skill)More development skills
View all Development skills →React and Next.js Regression Test Coverage
Generate React and Next.js tests with Jest, Testing Library, MSW, and Playwright, prioritized by regression risk.
Playwright E2E Tests and Flake Repair
Build reliable Playwright tests, repair flakes from evidence, migrate suites, and configure faster, diagnosable CI.
Pre-Ship Verification for Code Changes
Before calling a code change done, Claude grades every claim by evidence, tries to break it, and reports what's verified vs assumed.
TestRail Case and CI Result Sync
Sync automated tests to TestRail cases and scoped runs with stable mappings, retry handling, and verified result counts.