Guides

Three pages we already hand people on a first call.

Practical guidance on access control, AI evaluation, and repository ownership, based on recurring delivery work.

Access control before the first screen

Adding access control to a live product is how data breaches happen. We write who can read which record before the first screen exists.

A model wired to a button is easy. The hard part is knowing whether it helps, what it costs, and who is allowed to let it act.

The work happens in a repository you own. If the engagement ended tomorrow, you would still have the code, the history, and the issues.