Security model (short)
- RBAC: roles gate access.
- Approval gates: sensitive actions require explicit approvals.
- Audit trail: append-only events + verification.
- Exports: PDF/JSON/CSV designed for audits.
Security docs
What we log, how we gate sensitive actions, and how we export evidence safely.
How we're different
- Auditability is built in.
- Policy gates reduce risk.
- Exports are deterministic.
Trust rule: AI drafts, humans approve sensitive actions.