Our editorial commitments
We separate three things: what a tool suggests, what a test observes and what an authorised reviewer decides. The distinction is central to every guide. We do not describe an unexecuted path as tested or a missing source in an excerpt as proof that no source exists.
Our examples are clearly labelled and deliberately small. The brass-key story is an instructional example, not a report about a released game. Its purpose is to show how ownership, reachability and knowledge can produce different failures around the same line.
How we use references
We link to primary project documentation where practical and describe general-purpose modelling tools as such. A link is not an endorsement, affiliation or guarantee of compatibility. We do not claim to have benchmarked tools simply because we discuss how they might fit into a workflow.
Relevant foundations include ink, Yarn Spinner, Twine, Z3, Alloy, TLA+, Hypothesis and the NIST AI Risk Management Framework. Teams should check current documentation, licensing and data terms before adopting any external tool.
What this site does not provide
There is no script upload, hosted model, automatic certification or production issue tracker here. The contact page is a local note-preparation tool. It cannot analyse a game script or send a message without a separate action in your own email application.
We do not promise exhaustive coverage or a particular defect-detection rate. Those claims would require a defined dataset, method and evaluation that this site does not present.
Corrections and constructive disagreement
If an explanation conflates a runtime fact with an editorial assumption, tell us which passage needs attention and why. A public documentation link or a non-confidential minimal example is more useful than an unpublished production script.