# DriftGate

> Enterprise Agent Control Plane

A governance and learning control plane for agentic systems. Bring approved tools, identity, policy, execution and evidence into one system.

Revision: 2026-09-22.website-v1
Publication: public-platform-overview

## Availability

This site presents DriftGate’s full platform direction. Release and integration qualification are not established by this website build.

## Six pillars

### Tooling & Environment Control

Bring tools, skills and configuration into one approved environment. Know what should be there—and what actually is.

Boundary: Client adapters must be qualified. Installed, loaded and observed are different states; unknown stays unknown.

https://driftgate.ai/platform/tooling-environment-control/

### Identity & Policy

Connect every action to an identity, a scope and a current policy. Keep personal access and company authority separate.

Boundary: A connected account, a paid plan or a model-supplied agent name is not authorization.

https://driftgate.ai/platform/identity-policy/

### Execution Governance

Gate consequential actions. Apply approvals and limits. Reconcile uncertainty, and keep an attributable receipt.

Boundary: Controls apply at qualified execution boundaries. There is no blanket exactly-once guarantee.

https://driftgate.ai/platform/execution-governance/

### Observability & Assurance

Trace the path from configuration to decision to effect. See what changed, who acted and which evidence supports the outcome.

Boundary: Observation is not enforcement. Missing telemetry does not mean success or zero usage.

https://driftgate.ai/platform/observability-assurance/

### Lifecycle Management

Review, release, provision, rotate, suspend and retire. Give agent infrastructure a lifecycle, not a collection of loose ends.

Boundary: Revocation stops supported future use; it does not undo an external effect that already completed.

https://driftgate.ai/platform/lifecycle-management/

### Data & Context Governance

Control which knowledge reaches which agent. Preserve provenance, isolate context and make sharing a deliberate decision.

Boundary: Changing a profile does not remove information already present in a conversation. Keep sensitive sessions separate where required.

https://driftgate.ai/platform/data-context-governance/

## All 59 capability areas

- **T01 — Approved MCPs:** Discover and assign reviewed MCP servers, tools and versions. A catalog entry is not blanket action permission. Availability: qualification-pending.
- **T02 — Skills and plugins:** Review, release, pin and update skills and client-specific plugins with provenance and compatibility. Availability: qualification-pending.
- **T03 — Model management:** Manage supported model preferences and observed settings without pooling subscriptions or provider quotas. Availability: qualification-pending.
- **T04 — Environment sync:** Apply approved profiles across enrolled contexts while preserving unmanaged settings and independent authorization. Availability: qualification-pending.
- **T05 — Version control:** Pin artifact and profile revisions. Activate upgrades or rollbacks deliberately, not on every source edit. Availability: qualification-pending.
- **T06 — Infrastructure templates:** Reuse reviewed runtime and deployment patterns with explicit prerequisites, ownership and approval. Availability: qualification-pending.
- **T07 — Secrets and configuration:** Distribute managed settings and opaque credential references, never raw secrets in portable profiles. Availability: qualification-pending.
- **T08 — Drift detection:** Compare expected, approved, installed, loaded and observed state. Explain differences and precise unknowns. Availability: qualification-pending.
- **T09 — Tool sandboxing:** Bound approved execution by qualified filesystem, network, credential, time and resource controls. Availability: qualification-pending.
- **I01 — Agent identity:** Identify agents through authenticated, server-bound principals rather than self-reported names. Availability: qualification-pending.
- **I02 — Human identity:** Keep the authenticated human authorizer distinct from the client and executing agent. Availability: qualification-pending.
- **I03 — Service accounts:** Scope organization-owned identities for unattended workloads without permanent employee impersonation. Availability: qualification-pending.
- **I04 — Roles and scopes:** Limit operations and resources within the applicable tenant, project, application and environment. Availability: qualification-pending.
- **I05 — Policy engine:** Evaluate shared, versioned rules using trusted inputs. Distinguish observation, warnings, approval and enforcement. Availability: qualification-pending.
- **I06 — Permissions and entitlements:** Intersect policy, consent, delegated authority, resource permissions and plan eligibility. Availability: qualification-pending.
- **I07 — Credential vaults:** Protect independently selected accounts and credentials, with scoped reauthorization and disconnection. Availability: qualification-pending.
- **I08 — Delegation:** Grant bounded, expiring task or child-agent authority without widening the parent’s scope. Availability: qualification-pending.
- **I09 — Multi-tenant isolation:** Preserve tenant and personal/business boundaries across reads, writes, sessions, caches and exports. Availability: qualification-pending.
- **I10 — Compliance rules:** Apply organization-defined requirements through supported controls. Rules alone do not establish certification. Availability: qualification-pending.
- **E01 — Action authorization:** Check the exact actor, operation, target, connection and current authority before a protected effect. Availability: qualification-pending.
- **E02 — Approval workflows:** Bind approvals to material inputs, resource, revision and expiry. Re-evaluate changed requests. Availability: qualification-pending.
- **E03 — Capability controls:** Constrain eligible tools and actions, and enforce those constraints on direct invocation. Availability: qualification-pending.
- **E04 — Execution limits:** Bound supported operations by concurrency, duration, size and rate at the actual execution point. Availability: qualification-pending.
- **E05 — Budgets and quotas:** Track and constrain supported usage. Hard limits require enforceable reservations or provider controls. Availability: qualification-pending.
- **E06 — Runtime policy enforcement:** Apply shared decisions through qualified adapters, including revocation and queued-work rechecks. Availability: qualification-pending.
- **E07 — Effect isolation:** Bind effects to intended destinations and isolate owned execution resources from unrelated environments. Availability: qualification-pending.
- **E08 — Idempotency and safety:** Preserve operation identity, reject material conflicts and avoid unsafe duplicate effects. Availability: qualification-pending.
- **E09 — Execution receipts:** Retain scoped, attributable records of authority, relevant versions, execution state and bounded results. Availability: qualification-pending.
- **E10 — Reconciliation:** Resolve an uncertain outcome using the original effect identity before considering a retry or recovery. Availability: qualification-pending.
- **O01 — Tracing and telemetry:** Correlate requests, decisions, approvals, attempts, workflows and receipts with scoped telemetry. Availability: qualification-pending.
- **O02 — Audit logs:** Inspect governed changes and actions through tenant-scoped history and declared retention. Availability: qualification-pending.
- **O03 — Evidence capture:** Connect outcomes to source, release, permission basis and exercised enforcement boundaries. Availability: qualification-pending.
- **O04 — Risk detection:** Surface suspicious or unexpected patterns for review without turning risk scores into authority. Availability: qualification-pending.
- **O05 — Forensics:** Reconstruct source-to-decision-to-effect lineage while preserving original evidence and privacy. Availability: qualification-pending.
- **O06 — Replay and simulation:** Inspect compatible history and simulate policy without repeating live external effects. Availability: qualification-pending.
- **O07 — Verification:** Read back actual state. A written configuration alone is not proof of synchronization. Availability: qualification-pending.
- **O08 — Metrics and reporting:** Measure setup, adoption, drift resolution, action outcomes and approval burden. Availability: qualification-pending.
- **O09 — Incident response:** Link investigation, scoped suspension and evidence-led recovery to appropriate authority. Availability: qualification-pending.
- **O10 — Runtime health:** Distinguish reachability, expiry, callable state, client freshness and observed runtime identity. Availability: qualification-pending.
- **L01 — Agent registry:** Track ownership, approved definitions, environment bindings and observed running state. Availability: qualification-pending.
- **L02 — Versioning:** Separate editable drafts from approved, immutable revisions and their active consumers. Availability: qualification-pending.
- **L03 — Provisioning:** Establish supported resources and bindings through explicit ownership and verified enrollment. Availability: qualification-pending.
- **L04 — Change management:** Plan, review and activate changes with affected-consumer visibility and a rollback path. Availability: qualification-pending.
- **L05 — Reviews and approvals:** Review governed releases and lifecycle transitions without granting blanket execution authority. Availability: qualification-pending.
- **L06 — Rotation:** Replace supported credentials and bindings deliberately, accounting for consumer impact. Availability: qualification-pending.
- **L07 — Suspension:** Pause intended governed use without deleting history or obscuring in-flight work. Availability: qualification-pending.
- **L08 — Revocation:** Remove relevant authority, account for queued work and preserve independently authorized peers. Availability: qualification-pending.
- **L09 — Deprecation:** Identify affected consumers and supported replacements before retiring a version or integration. Availability: qualification-pending.
- **L10 — Retirement:** Retire scoped resources while preserving required lineage and reconciling uncertain deletions. Availability: qualification-pending.
- **D01 — Knowledge bases:** Connect selected sources and deliberately publish and bind approved knowledge revisions. Availability: qualification-pending.
- **D02 — Vector stores:** Use scoped, derived retrieval indexes where needed. Stale or failed indexing cannot satisfy activation. Availability: qualification-pending.
- **D03 — Data classification:** Carry sensitivity, ownership and intended-consumer labels into access and handling decisions. Availability: qualification-pending.
- **D04 — Context isolation:** Keep retrieval and context within declared boundaries. Separate sessions where isolation requires it. Availability: qualification-pending.
- **D05 — Redaction and PII protection:** Apply qualified redaction and data-handling controls without claiming perfect detection. Availability: qualification-pending.
- **D06 — Data-access policies:** Check current grants, connection state, source permissions and release approval at retrieval boundaries. Availability: qualification-pending.
- **D07 — Retention and deletion:** Apply defined retention to each data class and verify supported deletion outcomes. Availability: qualification-pending.
- **D08 — Provenance and lineage:** Trace released or retrieved context to its source, revision and transformation history. Availability: qualification-pending.
- **D09 — Egress controls:** Restrict supported outbound data paths and disclose what can actually be intercepted. Availability: qualification-pending.
- **D10 — Cross-agent knowledge sharing:** Share reviewed knowledge through explicit grants, purpose, audience and revision. Availability: qualification-pending.

## Solo pricing direction

USD 19/month. Status: selected launch target, not purchasable. Later plan prices and final commercial terms remain in review.

## More information

- [Human overview](https://driftgate.ai/platform/)
- [Architecture](https://driftgate.ai/architecture/)
- [Integration scope](https://driftgate.ai/integrations/)
- [Product manifest](https://driftgate.ai/agent/product.json/)
- [Capability manifest](https://driftgate.ai/agent/capabilities.json/)