AI disclosure

What the AI-assisted seal means, what it does not, and how to check the work instead of taking it on trust.

What the seal means

The AI‑assisted mark in the corner of every page is an honesty notice, not a badge. It says that AI tooling was involved in producing or editing text on this site — drafting, rewriting, tightening prose, generating code, or checking it.

It is on every page rather than buried on this one because a disclosure you have to go looking for is not a disclosure. If a page here was written without assistance, the mark still appears; the mark describes the site, not the individual article.

What it does not mean

The seal is not a quality rating, an accuracy guarantee, or a substitute for reading the thing. Concretely:

Not verified by definition

AI output is fluent and frequently wrong. Anything technical on this site should be treated as a draft until you have tried it.

Not a substitute for sources

Where a claim matters, the underlying source is linked. Follow the link rather than trusting the summary above it.

Not a claim of independence

Nothing here is generated to push you anywhere. There is no telemetry, no tracking pixel and no advertising on this site.

How to check instead of trusting

This is the part that matters. Every claim above is checkable without taking anyone's word for it:

• The hardening guides cite control references. STIG and CIS identifiers are given so a claim can be checked against the published control rather than against this page's interpretation of it.
• The code is public. Everything described here lives in public repositories under github.com/neohiro, with public history. Read it, fork it, or run it.
• The corrections are public too. An error that has been fixed leaves a trace in the history. If something here is wrong, the fix is a pull request, not a quiet edit.

Report a problem

If a page states something false — a broken mitigation, an invented control reference, a security claim that does not hold — that is a bug and it is worth reporting. Security findings go through the private advisory channel described in SECURITY.md; everything else is an issue on the relevant repository.