Status
What’s genuinely solid
Section titled “What’s genuinely solid”Over half the EXO cmdlet surface is wired and live-verified end-to-end: directory (incl. shared mailboxes & mail-enabled security groups), permissions, RBAC, full mail-flow, Defender/threat + quarantine, compliance/retention, public folders, mailbox-job reads, and message trace. Full CRUD on the high-traffic families has been exercised create→read→update→delete against a real tenant. The codebase is correctness- and security-audited (the 15-agent endpoint-parity audit is fully remediated), with ~5,700 tests including behavioral coverage of the fragile seams.
What’s not done yet (tracked, not hidden)
Section titled “What’s not done yet (tracked, not hidden)”- Async-operations model - the single biggest remaining feature (the operations project
is a placeholder). Long-running jobs (mailbox-move execution, large audit-log searches, wide
reports) need a
202 + status resourcepattern under the App Service request-timeout cap. Tier 8 Reports (2/30) is largely blocked on this. - Coverage tail - Tier 9 Hybrid (26/75) + niche/diagnostic surfaces; most of the unrepresented ~20% is deliberately skipped (on-prem-only, deprecated, or Graph-preferred).
- Observability - console-only; OpenTelemetry / App Insights needed before production.
- Production deployment - Bicep / Container Apps / CI-CD / multi-tenant is intentionally on hold until the feature surface settles.
$metadatacompleteness - alternate keys are declared; OData bound-actions + containment are deferred until a generic-OData-tooling consumer (Excel / Power BI / codegen) exists.- Automated license fulfillment: manual in phase 1, an order triggers a human to run
LicenseSigner issueand email the key; see License key.
Readiness summary
Section titled “Readiness summary”| For… | Readiness |
|---|---|
| Code correctness + security review | ~90% - audited, tested, defects closed |
| An internal pilot against one tenant | ~80% - mainly needs a deployment + basic observability |
| A general multi-tenant production service with an SLA | ~60% - needs async-ops, observability, deployment, multi-tenant |
The honest one-liner: the hard engineering is in good shape; the “make it a running service” work is what is left, and most of that is deliberately deferred, not unfinished by accident. That is why exonaut is in early access rather than on sale: the waitlist exists so that the first customers arrive when the running-service work is done.

