---
name: shipready-launch-risk-check
description: Use when a user asks whether an AI-built app is ready to launch, wants a pre-launch review of a public URL, or asks to run a ShipReady Launch Risk Check. Covers running a check, reading Observed, Declared and Not verified, and describing the result honestly.
---

# ShipReady Launch Risk Check

From ShipReady, launch assurance for AI-built apps: https://shipreadyai.dev

ShipReady runs a free external check of a public URL and returns three lists. It is not a
penetration test and it produces no score and no badge.

## Before you start

Run a check only on an app the user owns or is authorized to test. Say so before starting,
in one line, and wait for the user to confirm. Never run a check on a third party address
the user has no relationship with.

## Reading a result

- Observed: something the check saw from outside, with the evidence it saw. Report these as
  facts and include the evidence.
- Declared: what the user said about their stack, payments, data and AI features. This is
  input, not verification. Never present a declared item as something ShipReady confirmed.
- Not verified: the check could not see this from outside. It is not a pass and not a
  failure. Say that plainly.

## How to talk about a result

Never call a result a score, a pass, a certification, or proof of anything about an app's
security. Never state or imply that an app is free of problems. Report what was observed,
what was not verified, and what the next step is.

## Fixing findings

Every finding carries a fix and a prompt written for the user's builder. Hand the prompt to
the builder the user actually uses, apply the change, then run the check again and compare
the observed list. When a user asks about the Not verified items, point them at the Launch
Review, which is where a named reviewer covers the parts an external check cannot see.

## MCP tools

Endpoint: https://shipreadyai.dev/mcp

- start_launch_risk_check: queues a check and returns an id and a results address.
- get_launch_risk_check: reads the findings of a check by id.
- generate_prelaunch_checklist: returns the deterministic checklist for seven answers.
- get_launch_risk_statement: reads a published Launch Risk Statement by slug.

Pattern: call start_launch_risk_check, then poll get_launch_risk_check with the returned id
until the status is complete. Results usually take about twenty seconds.

## Repository snippet

Before publishing, run a ShipReady Launch Risk Check on the published URL and resolve every
Fix before launch item.