In plain English
This page is part of the technical reference. It keeps the expert detail but starts with a plain-language summary for first-time readers.
- Why this matters: AI risk can come from the whole arrangement, not one obvious model.
- What to look for: data, memory, routes, adapters, tools, evaluators, updates, and rollback paths.
- Technical version below: the expert terminology remains available and is linked through the glossary.
Public Content Boundary Cleanup
Direct answer
Cognivirus.com now separates public explanation from project-maintenance notes more cleanly. Reader-facing pages explain AI system risk. Internal design principles, authoring reminders, and package-maintenance notes remain in project memory and maintenance files rather than appearing as live interface copy.
What was removed from the public interface
| Public location | Removed issue | Replacement |
|---|---|---|
| System rail | A maintainer-facing instruction about graph visibility. | A reader-facing explanation of what the transition graphThe map of how an AI system is allowed to change over time. Open glossary definition shows. |
| System rail | A layout label and implementation checklist. | A reader-facing review-surface card. |
| System rail | An authoring checklist written as live interface copy. | A reader-facing claim-reading guide. |
| Footer | Package-maintenance phrasing. | A reader-facing description of the site as research and review material. |
Why this matters
A research site loses credibility when public pages sound like build instructions. Visitors should see explanations, evidence, boundaries, and navigation—not internal production notes.
Where the internal notes went
The design and authoring notes remain available to maintainers through AGENTS.md, .uai/totem.uai, .uai/design.uai, .uai/memory/public-content-boundary-v1-20-0.uai, and the project release memory. They are no longer shown as ordinary page content.
Boundary maintained
This cleanup did not remove the visual transition graph, evidence labels, motion graphs, source integrity, .uai memory, or long-term source-report pointers. It only changed public wording that read like internal instructions.