Mitra OS Wiki

Tag: ai safety

feature AI safety and auditability How Mitra OS limits what its language-model calls can do: read and draft only, grounded in live data, logged for audit, with failures isolated.