AI-assisted development governance

This entry describes an earlier build of Ledgeron. A new build has since replaced it under the same name, and some areas this entry calls implemented are planned again. Current status is on the product page.

Ledgeron’s development repository now includes formal operating rules for AI coding agents, module boundaries, testing requirements and security constraints. AI-assisted changes follow the same issue, review and CI workflow as human contributions.

The rules are written down so they can be enforced:

Which modules an agent may change for a given issue, and which boundaries it may not cross.
Which tests must exist and pass before a change is reviewable, including PostgreSQL-native verification for critical behavior.
Security constraints that apply regardless of who or what wrote the code: authorization checks, audit events, fail-closed defaults.

Nothing merges without review and a passing CI run. That was already true; the change is that the expectations are now explicit.

Help shape Ledgeron before general availability.

We are building Ledgeron first for real hosting operations at GOZEN, then expanding into controlled early access. Hosting providers interested in evaluating the platform, giving structured feedback or following development can register interest now.