Alpha Omega / Trust center
AI limitations
Model output is proposed data, never trusted authority.
Accuracy
AI responses and generated specifications may be incomplete, stale, or wrong. External websites and APIs may change after a tool is built.
Hidden reasoning
The application does not request, display, or persist private chain-of-thought. User-visible summaries, structured plans, evidence, and tool results are stored instead.
Authority
The model cannot grant itself permissions, raise budgets, modify Alpha controls, declare its own tests successful, or claim an action occurred unless trusted server evidence confirms it.