PublishedGuidanceLAST REVIEWED · 18 NOV 2026
Security operations and release
Qualify infrastructure, retention, recovery and operating scope before enabling the engineering workflow.
Runtime scope
Default-off controls
- MC1_SECURITY_ENABLED controls the preview API and Console.
- MC1_SECURITY_RESPONSE_ENABLED separately controls response dispatch and application.
- MC1_SECURITY_VERIFICATION_ENABLED separately controls independent gate verification.
- Live mode requires strict existing human identity. Workers require exact expiring scopes and current active tenant/project authority.
- Configure Trust encryption, explicit dedicated targets, approved AATS mapping and current identity-provider delegation before use.
Required production evidence
- Exact artifact and reviewed threat model.
- Qualified workload/database/evidence isolation and narrowly scoped machine credentials.
- Full migration chain, backup/restore and reversible application rollback.
- Real protected-gate smoke tests using separate dispatcher and verifier identities.
- Retention and erasure execution, integrity monitoring, bounded retry and emergency-stop exercises.
- Measured performance, infrastructure cost, accountable support and incident-response owners.
- Approved commercial scope and entitlements; no new price or customer charge follows source implementation.
Broader capabilities require separate qualification
The original local MCP configuration analyzer is a separate passive, explicitly authorized file assessment. This workflow does not launch untrusted engines or scan remote targets from the control plane.
Managed assessment execution, remediation/rehearsal/retest, vendor-specific integrations, physical endpoint or network containment, MDR staffing and general availability remain separate capabilities. Do not advertise them based on these component tests.
Was this page helpful?
