Cloudflare One

Workflow

1. Classify the ask: architecture, configuration, troubleshooting, migration, or review.
2. Gather context: account ID, users/sites/apps, identity provider, SCIM/group sync, device management, traffic path, compliance constraints, and rollout blast radius.
3. Retrieve only the current docs needed for the products involved: Access, Gateway, WARP/device client, Tunnel/Mesh, Cloudflare WAN, DLP, CASB, device posture, or identity.
4. If account access is available, inspect existing resources before proposing or making changes: Access apps/policies/groups/IdPs, Gateway rules/lists/categories, device profiles/posture checks, tunnels/routes, DNS/resolver settings, and locations/sites.
5. Propose the change set with prerequisites, validation, and rollback. For risky changes, stage disabled or scoped to a pilot group/site unless the user explicitly asks otherwise.

Assessment Prompts

Use these to avoid jumping straight to configuration. Ask only the prompts relevant to the user’s task.

Architecture and Current State

• Sites and users: offices, branches, data centers, VPCs, remote users, contractors, user counts, and current connectivity model.
• Applications and destinations: SaaS, public apps, private apps, APIs, infrastructure targets, protocols, ports, hostnames, and IP ranges.
• Connectivity: VPN, MPLS, SD-WAN, direct Internet breakout, centralized backhaul, site-to-site needs, and private DNS architecture.
• Security stack: current SWG, NGFW, VPN/ZTNA, DLP, CASB, email security, logging, and compliance requirements.
• Identity: IdP, SCIM/group sync, group naming, multi-IdP needs, service accounts, and contractor/partner access.
• Rollout: pilot users/sites, blast radius, rollback path, support owners, and success criteria.

Access and SaaS Federation

• App shape: web app, API, SSH/RDP/VNC, database, SaaS app, public hostname, private IP, or private hostname. Retrieve Access application type (opens in a new tab) docs before choosing.
• Access model: clientless browser access, private networking with device client, peer to peer connectivity, service connections with service tokens or mutual TLS, or SaaS SSO federation.
• Policy needs: user groups, device posture, session duration, mTLS, service tokens, and app launcher visibility. Retrieve Access policy (opens in a new tab) docs before configuring selectors or evaluation order.
• SaaS details: SAML vs OIDC support, ACS/redirect URLs, Entity IDs/client IDs, required attributes, and tenant-control requirements.

Tunnel and Private Networking

• Sites and segments: which data centers, VPCs, offices, or network segments need connectivity.
• HA: dev/test single connector, production multiple connectors, or advanced multi-tunnel/site redundancy.
• Runtime: where cloudflared or WARP Connector/Mesh will run: VM, container, Kubernetes, bare metal, or other target.
• Egress: whether connectors can reach Cloudflare over the required outbound ports/protocols. Retrieve Tunnel connectivity prechecks (opens in a new tab) before naming exact endpoints.
• Origin reachability: whether the connector can resolve and reach every private origin.
• Routing: required CIDRs/hostnames, overlapping IP spaces, virtual networks (opens in a new tab), Split Tunnels (opens in a new tab), and private DNS/resolver policy (opens in a new tab) needs.
• Management model: prefer remotely managed/token-based tunnels for new deployments unless there is a clear reason for local config.

Gateway, TLS, and DLP

• Traffic controls: DNS categories, HTTP URL/path inspection, L4 ports/protocols, egress IP requirements, custom lists, and allow/block exceptions. Retrieve Gateway traffic policy (opens in a new tab) docs for current selectors and order of enforcement.
• Identity: whether Gateway policies need user or group selectors, and whether users will be authenticated through WARP/IdP context. Check Gateway identity selectors (opens in a new tab) and SCIM provisioning (opens in a new tab) when groups are involved.
• TLS inspection: root CA deployment path, certificate-pinned applications, compliance exceptions, and FIPS requirements. Retrieve TLS decryption (opens in a new tab) docs before enabling.
• DLP: sensitive data types, channels to inspect, TLS inspection readiness, DLP profiles, payload logging requirements, and false-positive tolerance. Retrieve DLP (opens in a new tab) docs before creating enforcement.

CASB, Device Posture, and Risk

• CASB: SaaS vendors, admin access level, scan policy, org size, remediation owner, and whether inline protection is also required. Retrieve CASB findings (opens in a new tab) docs before recommending remediation.
• Device posture: required checks, third-party EDR/MDM integrations, enrollment rules, device profiles, and split tunnel alignment.
• Risk scoring: relevant behavior signals, false-positive sources such as VPNs or service accounts, and whether risk is for investigation or enforcement. Retrieve user risk score (opens in a new tab) docs before using risk in policies.

Cloudflare WAN / Site Connectivity

• Site topology, on-ramp type, route ownership, tunnel redundancy, static vs BGP-managed routes, network firewall needs, and appliance/profile ownership. Retrieve Cloudflare WAN (opens in a new tab) and Cloudflare Network Firewall (opens in a new tab) docs before proposing site connectivity changes.

Guardrails

• Access controls application authorization; Gateway controls traffic inspection/filtering. Use both when the requirement spans identity-aware app access and network/web security.
• For new Access deployments, create policies through the reusable policy API (/access/policies) and attach them to applications. Do not send inline policies in an application create/update request unless the current API documentation explicitly requires an app-scoped policy.
• Treat an app-scoped policy reported as reusable: false as legacy. Migrate existing policies with the documented make_reusable endpoint or replace them with reusable policies; do not create new legacy policies. Distinguish legacy policies from the deprecated legacy private-network application type.
• Public hostname Access apps can be clientless. Private destination apps require WARP/Device client or another network on-ramp plus routes and DNS resolution. Retrieve self-hosted private app (opens in a new tab) docs before configuring private destinations.
• Cloudflare Tunnel is an off-ramp from a private network to Cloudflare. Cloudflare WAN and Mesh are other off-ramps which can also be on-ramps.
• Group-based policies depend on IdP group claims or SCIM. If group sync is missing, do not invent group selectors.
• Private hostnames need explicit DNS routing/resolution; creating an Access app alone is not enough. Use resolver policies (opens in a new tab) and review Connect a private hostname (opens in a new tab)
• HTTP inspection and DLP for encrypted web traffic require TLS inspection and planned Do Not Inspect exceptions.
• Gateway DNS, Network, HTTP, and Egress policies have different evaluation semantics. Retrieve order of enforcement (opens in a new tab) docs before explaining precedence.
• Start broad block/allow/DLP/TLS policies disabled limited to a pilot with specific target users or groups unless the user approves a wider rollout.

Identity and Access

• Access Groups are Cloudflare objects; IdP/SCIM groups are identity claims. Gateway group selectors use synced IdP groups, not Access Groups.
• Group names and SAML/OIDC attributes are case-sensitive. Verify exact claim names and values before creating group-based rules.
• SCIM changes and group membership can be stale until sync and re-authentication complete. Troubleshoot with the user’s last authenticated identity, not just the IdP state.
• Access policies are default-deny. A private app with routes but no Allow policy still blocks access.
• Access policy selectors can use IP lists, not Gateway domain or URL lists.
• SaaS federation handles authentication into the SaaS app. SaaS authorization and tenant restrictions usually require SaaS-side roles and/or Gateway tenant controls.
• Browser Rendering for SSH/VNC/RDP is an Access capability. Browser Isolation renders general web content remotely. Do not conflate them.

Device Client Deployment

• The Cloudflare One device client (opens in a new tab) is the on-ramp for user devices. Two components control it: enrollment rules (who can connect) and device profiles (how the client behaves after enrollment).

• The enrollment rule is an Access application of type warp, not a device setting. It accepts reusable Access policies. Look in Access for enrollment debugging, not Devices.

• For headless or autonomous devices (services, kiosks, Linux hosts), use service token enrollment. Non-human devices authenticate as non_identity@[team-domain].cloudflareaccess.com and have no group membership - device profiles targeting IdP groups will not match them. Target headless devices explicitly with the non-identity email, specific conventions about the devices (OS information, etc.),or let them fall to the default profile.

• Device profiles control connection mode, split tunnel (opens in a new tab) configuration, user permissions (disable, switch lock), auto-reconnect, and captive portal behavior. Profiles are matched by user group or device attributes in precedence order - first match wins, default profile catches the rest.

• Split tunnel mode is the single most impactful client setting. Choose the mode based on the deployment goal:

  Goal	Mode	Rationale
  VPN replacement only (private apps)	Include	Route only specified private CIDRs and hostnames through the client. Everything else goes direct. Minimal blast radius.
  SWG only (internet security)	Exclude	All traffic through the client. Exclude only what breaks (local printers, certificate-pinned apps).
  VPN replacement + SWG	Exclude	All traffic through the client. Most common enterprise configuration.
  Coexistence with another VPN	Include	Avoids conflict with the other VPN’s tunnel interface and DNS control.
  DNS filtering only	DNS-only mode	Only DNS queries go to Gateway. No traffic proxying.

• Include vs exclude is per-profile, not per-entry. You cannot mix modes in the same profile. Switching modes mid-deployment requires re-evaluating every entry.

• Split tunnel entries must align with tunnel routes bidirectionally. A CIDR in the include list without a matching tunnel route causes a black hole. A tunnel route without a matching device profile entry means traffic never enters the tunnel.

• MDM parameters (mdm.xml / managed preferences) override dashboard-configured profile settings for any setting specified in the file. If dashboard changes appear to have no effect on managed devices, check MDM config. Retrieve MDM deployment (opens in a new tab) docs for platform-specific file locations and parameters.

• If another VPN client or agent controls DNS on the device, the device client’s DNS interception will conflict. In coexistence scenarios, use “traffic only” mode to avoid routing table and DNS conflicts.

• Captive portal detection temporarily disconnects the client when it detects a portal (hotel WiFi, airport). This is a common source of end-user friction and should be managed carefully.

Private Networking

• Split tunnel mode changes the meaning of every route decision: Exclude mode sends traffic to Cloudflare when removed from excludes; Include mode sends traffic only when added to includes.
• Virtual networks should be used primarily when IP subnets overlap and hostname-based routing is not used. It can be used to control other user connectivity behavior, but it is recommended to manage through security policies.
• A healthy tunnel only proves cloudflared can reach Cloudflare. The tunnel must have appropriate published application routes, network routes, or hostname routes for connectivity to function.
• Cloudflare Tunnel and Cloudflare Mesh can both be used to facilitate connectivity to internal networks. Cloudflare WAN can as well, but it is gated behind Enterprise subscriptions. Retrieve choose an on-ramp (opens in a new tab) when deliberating between Tunnel types.
• Run multiple cloudflared connectors for production HA, preferably on separate hosts. Token-based, remotely managed tunnels are the default for new deployments.

Gateway, TLS, and DLP

• dns.domains matches a domain and subdomains; dns.fqdn is exact-match only.
• DNS pre-resolution selectors and post-resolution selectors do not behave like a single strict precedence list. Retrieve current evaluation docs before changing rule order.
• HTTP Do Not Inspect rules run before HTTP Allow/Block/Isolate behavior. A later block rule will not override an earlier inspection bypass.
• Certificate-pinned apps need Do Not Inspect exceptions before broad TLS inspection. Deploy the Cloudflare root CA to managed devices before enabling inspection.
• DLP profiles are detection definitions only. They do nothing until referenced by Gateway HTTP policies or CASB scan settings. Rules with body inspection may be evaluated multiple times in a single pass.
• Start DLP with payload logging where appropriate, tune false positives, then block.
• Gateway Network policies are strict L4 controls. Identity-aware L4 matching requires authenticated device context.

CASB, Risk, and Operations

• API CASB is out-of-band and periodic. It does not provide real-time inline enforcement although some integrations support “remediation”; use Gateway granular application controls for inline CASB capability for supported applications. Retrieve Granular application controls (opens in a new tab) when creating security policies for specific actions in specific SaaS applications.
• CASB findings are tied to specific assets and instances. Drill into affected assets before recommending remediation.
• Use current Dashboard remediation guidance for CASB fixes. Most remediations happen in the SaaS admin console, not Cloudflare.
• Large SaaS integrations can take 24-48 hours for initial scans. Reauthorizing can restart scan state; check credential health before reconnecting.
• User risk scores are behavior-based and asynchronous. CASB findings do not automatically imply high user risk.

Infrastructure Access

• Zero Trust Infrastructure Access (opens in a new tab) (ZTIA) is the purpose-built offering for SSH access through the device client. It provides capabilities not available through self-hosted apps: keystroke logging, control over how users authenticate to the target machine, short-lived certificates (opens in a new tab) that replace static SSH keys with ephemeral certs tied to Access identity, and lightweight privileged access management. Use Infrastructure Access apps for SSH when the device client is deployed.
• Browser Rendering (opens in a new tab) provides clientless SSH, RDP, and VNC through the browser without requiring the device client. Clientless RDP includes session recording and file transfer controls. Use clientless access when a device client cannot be installed (contractors, partner access, unmanaged devices) - typically not as the default for managed users with the client installed.
• Audit SSH (opens in a new tab) is a Gateway Network policy action that logs SSH commands without blocking. It requires the session to be proxied through Cloudflare.
• Short-lived certificates require CA configuration on the target host and sshd configured to trust the Cloudflare CA public key. Retrieve short-lived certificate setup (opens in a new tab) docs before configuring.
• For kubectl and database access behind private networks, use the device client with private destination routing. There is no Infrastructure Access or browser-rendered equivalent for arbitrary TCP protocols today.

Logs, Analytics, and DEX

• Gateway activity logs (opens in a new tab) record DNS, HTTP, and Network policy decisions. Filter by rule name, user identity, destination, action, and time range. These are the primary troubleshooting tool for “why was this blocked/allowed.”
• Access audit logs (opens in a new tab) record authentication decisions per app - who authenticated, which policy matched, and session details. Use for verifying policy behavior and investigating access failures.
• Shadow IT discovery (opens in a new tab) uses Gateway HTTP logs to surface unmanaged SaaS applications. Requires TLS inspection for HTTPS visibility.
• DEX (Digital Experience Monitoring) (opens in a new tab) provides fleet-level and per-device connectivity diagnostics. Use DEX tests (opens in a new tab) (HTTP, traceroute) to proactively monitor reachability to critical origins and internal apps. Fleet status shows device client health, connection mode, and connectivity state across the enrolled population.
• Logpush (opens in a new tab) exports Gateway, Access, Network, and DEX logs to external SIEM or storage. Configure before go-live if the customer requires centralized log retention or compliance reporting.
• When troubleshooting, work from logs toward config: identify the log entry showing the failure (Gateway block, Access deny, tunnel error, DNS resolution miss), then trace back to the responsible rule, route, or policy.

Cloudflare WAN / Site Connectivity

• Cloudflare WAN is connectivity, not a security service. Apply inspection and policy with Gateway and Network Firewall where required.
• WAN firewall expressions are not the same language as Gateway wirefilter expressions. Retrieve the current syntax before editing.
• Generated IPsec PSKs and some OAuth/client secrets are returned once. Store them immediately.

Output Defaults

• Designs: current assumptions, target architecture, product responsibilities, rollout phases, validation, and open decisions.
• Configuration work: prerequisites, exact resources to inspect/create/change, test cases, and rollback.
• Troubleshooting: traffic path, likely failure point, evidence to collect, and next test.

Validation Prompts

• Access: test authorized, unauthorized, posture-failing, service-token, and multi-IdP flows when applicable; inspect logs and policy precedence. For new policies, verify they are managed through the reusable policy collection and not returned as reusable: false app-scoped policies.
• Private network access: verify route lookup, tunnel health, origin reachability, split tunnel behavior, DNS resolution, and end-to-end access from a device client test device.
• Gateway: verify rule type, action, traffic expression, precedence/evaluation phase, referenced lists, and Gateway settings before enabling broadly.
• TLS/DLP: test Do Not Inspect exceptions and root CA trust before enabling inspection; test DLP with known samples and monitor false positives before blocking.
• CASB/risk: confirm integration health, credential expiry, asset discovery, scan timing, finding instances, and risk-score signal latency before declaring remediation complete.
• Cloudflare WAN: verify tunnel health, route priority/ownership, traffic flow, firewall expression syntax, and connector/appliance telemetry where applicable.

API Safety

• Use fully qualified MCP tool names when MCP tools are available.
• Never guess category IDs, application IDs, wirefilter fields, or API request bodies. Retrieve the current schema/docs and existing account objects.
• Do not enable broad production policies without explicit approval.