# Production Ready

> Never deploy, publish or change production config unless the user asked for it in this task.

- Skill: `amsborse/production-ready` (Agent Skill)
- Install (CLI): `npx get-skill add amsborse/production-ready`
- Raw SKILL.md: https://api.skillmd.com/api/skills/amsborse/production-ready/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: DevOps & Infra
- Author: amsborse (https://skillmd.com/u/amsborse)
- Updated: 2026-11-17
- Page: https://skillmd.com/skills/amsborse/production-ready

---


# Production ready

Never deploy, publish or change production config unless the user asked for it in this task.
Leave the work on a branch and say where.

## 1. Audit (read-only)

Score each area **ready / gap / unknown** with file:line evidence. Unknown is not ready.

- **Security:** run the `security-audit` skill (repository mode).
- **Tests:** suite runs green from a clean checkout; the main user flow has an end-to-end test
  (`test-everything`).
- **Errors and observability:** no swallowed errors; structured logs without secrets; health check;
  error reporting; alerts on the failures that matter.
- **Performance:** measured load time or latency for the main flow, bundle or image sizes, slow
  queries, timeouts and retries with backoff on network calls.
- **Config and secrets:** all config from environment, an `.env.example`, nothing secret in the
  repo, separate dev and prod settings.
- **Build and deploy:** reproducible build, pinned dependencies, CI runs tests and lint,
  a documented deploy and a tested rollback.
- **Data:** migrations are reversible or backed up first, backups exist and a restore was tried.
- **Docs:** a README that gets a new developer running, and a runbook for incidents.

## 2. Fix

Ask once with the numbered gap list if fixing needs decisions (new services, paid tools, schema
changes). Fix the safe gaps highest-risk first, one small change at a time, each with a test.
Delegate isolated chunks with `coordinate-agents` when it helps.

## 3. Prove

Re-run the audit. Report the before and after scores, evidence for every "ready", and what is
still a gap or unknown. Verify the main flow in the running app (`verify`), and get an independent
review of the diff (`review-code`) for large changes.

