MomentumOS

Security & Trust

Version 1.3 · Effective 11 July 2026 · Reviewed 6 November 2026

This Security & Trust page summarises the controls MomentumOS uses to protect customer workspaces, govern access and support enterprise security review.

Document control

Previous version: 1.2

Version: 1.3

Effective date: 11 July 2026

Last updated: 6 November 2026

Reviewed: 6 November 2026

Change summary: November 2026 cross-platform review: SourceMend data health, AI Connect MCP read-only access, tenant-gated Action Connect shadow proposals and approvals, independent ProProof execution boundary, updated UK privacy and cookies guidance, EU AI transparency, and clearer operational and commercial limitations.

Authentication and access

MomentumOS uses authenticated access, role-based permissions, company-scoped workspaces, workspace isolation and server-side access checks to help ensure users only access information they are authorised to view.

Workspace and company isolation

MomentumOS is designed around company and workspace boundaries. Database access patterns and Row Level Security policies support tenant isolation where implemented.

Encryption and credential handling

MomentumOS is intended to run over HTTPS. Platform vendors such as Supabase and Vercel provide managed infrastructure controls. Credentials should be handled through OAuth or Secure Credential Exchange and stored separately from onboarding workspace data. Credentials are not redisplayed after submission.

Auditability and evidence lineage

MomentumOS supports audit logging foundations, security event visibility, evidence snapshots, AI usage logging, report generation records, data-rights workflows and investigation history where available.

Connector governance

Integrations are governed through connector lifecycle status, validation, production-readiness checks, secure credential handling, OAuth where supported and approval boundaries for sensitive changes.

Read-only integration by default

MomentumOS uses read-only intelligence integrations by default. Controlled outbound actions, signed webhooks, embedded intelligence and headless API access are enabled only for approved customers or platform partnerships where they are configured, entitled, validated and governed through explicit permissions and approval controls.

Controlled outbound actions and field allowlists

Controlled outbound actions are Disabled unless feature flags, entitlement, configuration, validation and approval requirements are satisfied.

Allowed fields and allowed values are governed through policy records. Arbitrary JSON write-back is not supported as a customer-facing default.

Approval controls keep AI recommendations separate from operational execution.

OAuth and Secure Credential Exchange

OAuth is preferred where the provider supports it. API keys, client secrets, service accounts, certificates, private keys, webhook signing secrets and API credentials should use Secure Credential Exchange and credential references rather than ordinary workspace fields.

API credential management and headless API scopes

Headless API access is scoped, rate-limited and governed by entitlement and credential status where enabled. Credential rotation, revocation and expiry metadata should be tracked without exposing secret values.

Webhook security

Webhook delivery requires an approved HTTPS endpoint, signing configuration, replay protection, delivery monitoring and dead-letter handling before production use. Webhook replay controls help prevent repeated event acceptance. Capability states should be labelled Disabled, Configurable, Beta, Awaiting Validation or Validated.

Embedded intelligence security

Embedded widgets and restricted iframe delivery require signed short-lived embed tokens, allowed-origin configuration, frame restrictions, event bridge validation, scoped access and production approval gates.

Provider validation

Provider validation is tracked separately from connector code. A provider should not be labelled Production Ready for write-back, webhooks, embeds or headless API until sandbox and production validation evidence exists.

Monitoring and responsible disclosure

MomentumOS monitors platform reliability and security signals using available deployment and application controls. Security concerns can be reported to info@arysconsultants.com.

Certification status

MomentumOS documentation may describe security-ready or compliance-ready architecture. MomentumOS does not claim SOC 2 certification, ISO 27001 certification, GDPR certification or penetration-test completion unless such evidence is separately published.

SourceMend read-only review and export controls

SourceMend is designed to inspect authorised data sources, flag missing, invalid, duplicate or conflicting values and preview suggested corrections. Corrections are exported only after user review; source-system mutation is not part of the default scanning flow.

Data quality findings, uploaded samples and exported fixes remain subject to tenant access restrictions and the customer's data security responsibilities.

AI Connect MCP security controls

AI Connect uses tenant-bound credentials with read-only access scopes such as summary:read, assessment:read and evidence:read. Credential expiry, revocation and a connection kill switch support customer control over access.

An action:propose permission, where enabled, is a separately scoped proposal capability, not permission to approve or execute. Third-party AI applications remain outside MomentumOS's direct administrative control.

Action Connect shadow safeguards and ProProof boundary

Action Connect uses server-side tenant checks, explicit feature flags, per-tenant activation and shadow-mode controls. Its proposal state machine separates draft or proposed work, human approval, possible execution and recorded outcome.

Approved proposals bind the exact payload and evidence snapshot hashes, evidence and policy versions, actor and expiry. Independent ProProof assurance and an authenticated internal worker are required before any enabled execution can proceed.

As reviewed on 6 November 2026, Action Connect is deployed for controlled shadow-mode evaluation in an isolated test tenant. No tenant is authorised for live execution, and the end-to-end live ProProof gateway, receipt replay and destination execution path has not yet been accepted for general production use.

An internal synthetic backend test has verified proposal creation, human approval and refusal of an execution claim in shadow mode. This is not a substitute for independent penetration testing, provider acceptance or an authenticated browser-to-API end-to-end test.

Data health, connector readiness and runtime limitations

CRM and PRM connections are subject to provider authentication, data mapping and acceptance checks; a connector appearing in the product does not imply that production access, write-back permissions or provider approval has been granted.

Read-only assessment and evidence retrieval must remain separate from any future outbound actions. Feature availability and assurance coverage should be verified for each deployment, tenant and integration.