Private betaThis is the SIGSHIP beta. Everything here is live for testers; the public launch is in January.SIGSHIP private betaOpen the app →
Readiness score

The score, with the sums shown. 

A number from 0 to 100 that shows how close your app is to Ready, and how it changes over time. Here is exactly how it is worked out.One published formula from the finding counts. No weights hidden in a model, no machine learning.

The formula

Every app starts at 100. Each problem the check finds takes points away, by how serious it is. The score can’t go below 0.Start at 100. Subtract per finding by severity, with a cap on how many findings of each severity count. Floor at 0.

FindingPoints offCounted at most
Fix nowSEV-130 each2
Fix soonSEV-212 each3
Nice to fixSEV-34 each5

So the lowest an app can score is 100 minus 60, 36 and 20, which is 0. A clean app scores 100.Minimum 100 − (2×30 + 3×12 + 5×4) = 100 − 116, floored to 0. A clean run scores 100.

What does not count

A problem you have marked as an accepted risk, with a reason and an expiry date, is left out of the score.Findings matched by an entry in sigship-accept.json are excluded. Expired entries count again.

What the score is for

The score is for the graph: it shows your app getting safer over time, and it lets you compare two days. It never decides the verdict.The score feeds the history graph and the badge record. GO and NO-GO are unchanged by it: any SEV-1 or SEV-2 finding is NO-GO, whatever the score.

A score of 100 means these checks found nothing that blocks launch. It does not prove the app is safe.Score 100 means zero findings from a pattern check. It is not a security audit and does not prove the app is safe.

Worked examples

FindingsScoreVerdict
None100GO
1 Nice to fix1 × SEV-396GO
1 Fix soon, 2 Nice to fix1 × SEV-2, 2 × SEV-380NO-GO
1 Fix now, 1 Fix soon1 × SEV-1, 1 × SEV-258NO-GO
3 Fix now (only 2 count), 4 Fix soon (only 3 count)3 × SEV-1 (cap 2), 4 × SEV-2 (cap 3)4NO-GO

Where you see it

The score and its history come with every Cloud plan, rolling out by January. The free check gives you the verdict and the list of findings today.Cloud plans store a score per run and draw the history; rolling out by January. The CLI, MCP and the in-browser check report findings and the verdict; the score is computed from the same counts.