production-readiness-review

Consolidates SLOs, rollback tests, runbooks, and capacity evidence into a go/no-go launch checklist.

Updated May 21, 2026
One-click install
npx skills add https://github.com/CagesThrottleUs/private-ai-harness --skill production-readiness-review-cagesthrottleus
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: production-readiness-review
Source: https://github.com/CagesThrottleUs/private-ai-harness/tree/main/skills/production-readiness-review
Command: npx skills add https://github.com/CagesThrottleUs/private-ai-harness --skill production-readiness-review-cagesthrottleus

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Launch readiness is often scattered across deployment, observability, and testing workflows, so a service can pass every individual gate and still be unready for production traffic. This Skill forces a single consolidated Production Readiness Review (PRR) with human sign-off before go-live. ## Core Features & Use Cases - Seven-dimension checklist artifact: Produces a PRR document covering Service Levels, Architecture & Dependencies, Performance & Capacity, Observability, Deployment & Rollback, Operability, and Testing & Security, with every line marked PASS / FAIL / N-A plus evidence. - Evidence-based validation: Rejects plausible-but-unverified claims — rollback must be tested, SLO targets must trace to real measurements, and dependencies must be verified to exist. - Human go/no-go gate: Acts as a hard gate in the epic lane before first production traffic; a FAIL blocks launch and a named human signs off. - Use Case: Before launching a new externally-reachable API service, run the PRR to confirm SLOs are measured, the rollback procedure was actually executed, runbooks were exercised, and on-call is defined — then present the artifact for human sign-off. ## Quick Start Run a production readiness review for the payments service before its first production traffic and produce the go/no-go checklist artifact.

Frequently Asked Questions about production-readiness-review

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I run a production readiness review before launch?

Produce a PRR checklist artifact covering seven dimensions — Service Levels, Architecture, Performance, Observability, Deployment & Rollback, Operability, and Testing & Security. Each line must be PASS, FAIL, or N-A with linked evidence, then a human makes the go/no-go call.

What should a production readiness checklist include?

It should include defined SLIs and SLOs, enumerated dependencies with failure behavior, load test results, golden-signals dashboards and alerts, a tested rollback procedure, exercised runbooks with on-call defined, and security review of auth and PII surfaces.

When should I skip a production readiness review?

Skip the PRR for internal-only libraries, documentation changes, or work that adds no new production surface. It is required only when a new externally-reachable service or materially new production surface is about to go live.

Why does a PRR reject written claims without evidence?

AI-authored readiness artifacts read plausibly while being untested, so the PRR only accepts evidenced claims. A rollback line needs a link to the rollback actually being run, and SLO targets must trace to real measurements rather than guessed numbers.

Can a production readiness review be auto-approved?

No. The reviewer agent validates completeness and evidence, but a human always makes the go/no-go decision with a named sign-off and date. A FAIL blocks launch and the gate never auto-advances.