Changelog

A new GA release for the Windows Cloudflare One Client is now available on the stable releases downloads page.

This hotfix resolves a regression that caused a large increase in DNS-over-TCP queries to fallback and internal DNS servers. The client now sends fallback DNS queries over UDP first, falling back to TCP only when a response is truncated, instead of querying both protocols in parallel.

A new GA release for the macOS Cloudflare One Client is now available on the stable releases downloads page.

This hotfix resolves a regression that caused a large increase in DNS-over-TCP queries to fallback and internal DNS servers. The client now sends fallback DNS queries over UDP first, falling back to TCP only when a response is truncated, instead of querying both protocols in parallel.

A new GA release for the Linux Cloudflare One Client is now available on the stable releases downloads page.

This hotfix resolves a regression that caused a large increase in DNS-over-TCP queries to fallback and internal DNS servers. The client now sends fallback DNS queries over UDP first, falling back to TCP only when a response is truncated, instead of querying both protocols in parallel.

The Account Roles API is deprecated and is being replaced by the Permission Groups API. An end of life date has not yet been established.

Review the Permission Groups API documentation; the response schema differs from the legacy Roles response.

• Integrations migrating to the Permission Groups API must obtain Permission Group IDs from that API and use them in the Account Members API policies request shape. Integrations that persist legacy Role IDs will need to remap their assignments.
• The legacy Role response includes a top-level description and a permissions object keyed by resource type with edit/read flags.
• The PermissionGroup response replaces those with a meta object containing label and scopes. Individual permissions are not returned as part of the permission group.
• The new API supports the API Token authorization scheme. The legacy Email + API Key authorization schema is provided for backwards compatibility.

For more information, refer to API deprecations.

Devin Outposts ↗ lets you run Devin agents on Cloudflare. Each Devin session runs in its own isolated sandbox backed by Cloudflare Containers, so agents can execute code and use development tooling in an isolated environment.

Use Devin Outposts when you want Devin sessions to run on Cloudflare managed infrastructure, with each session isolated from the others.

To get started, refer to Run Devin on Cloudflare using Devin Outposts.

This release introduces new rules and updates existing threat signatures to provide targeted protections for vulnerabilities in Adobe ColdFusion and WordPress, alongside enhanced generic protections against enhanced generic protections against Server-Side Request Forgery (SSRF), Local File Inclusion (LFI), and Cross-Site Scripting (XSS) obfuscation techniques. To strengthen overall detection capabilities across emerging threat vectors, new emergency detection rules have also been deployed for Generic Rules - Unauthenticated Remote Code Execution (RCE), Generic Rules - Authentication Bypass (Auth Bypass - 2) and Generic Rules - Information Disclosure.

Key Findings

• CVE-2026-48276: A path traversal vulnerability in Adobe ColdFusion file upload mechanisms allows unauthenticated attackers to write or upload files to arbitrary locations outside designated directories on the origin server.

• CVE-2026-48282: A path traversal vulnerability in Adobe ColdFusion enables unauthenticated attackers to manipulate directory sequences and access restricted system files on the host filesystem.

• CVE-2026-60137: An unauthenticated SQL injection vulnerability affecting WordPress. Threat actors exploit unsanitized input parameters to execute arbitrary SQL queries, leading to unauthorized database access, record manipulation, or data exfiltration.

• CVE-2026-63030: A remote code execution vulnerability affecting WordPress core and plugin components. Remote, unauthenticated attackers can execute arbitrary system commands to gain unauthorized access or establish backdoors on host servers.

Cloudflare Access now uses the standard browser-based login flow for private applications served over plaintext HTTP on port 80.

Previously, plaintext HTTP private apps fell back to the same session flow used for SSH, RDP, and other non-HTTP protocols: users got an Authentication required pop-up from the Cloudflare One Client, then had to select the notification to open a browser and log in. Now, users hitting an HTTP private app see the Access login page directly in the browser and receive a standard Access application token on success.

This brings the HTTP experience in line with HTTPS apps (with Gateway TLS decryption turned on). No configuration change is required. The Cloudflare One Client is still required to route traffic to the private network, but it no longer manages the Access session for HTTP apps.

Other non-HTTP protocols (SSH, RDP, arbitrary TCP/UDP) continue to use the Cloudflare One Client notification flow.

You can now monitor the total SQLite storage used by a Durable Object namespace over time in the Cloudflare dashboard. The new Total storage chart shows the maximum storage reported during each hour. This helps you identify storage growth, validate data cleanup, and investigate unexpected usage.

The chart appears only for SQLite-backed Durable Object namespaces. It does not appear for namespaces that use the legacy key-value storage backend. Viewing storage for individual Durable Objects by ID or name is not supported.

For more information, refer to Metrics and analytics.

You can now preview the content of sent emails directly from the Email Service Activity log. Expand a sent email and open the new Preview section to inspect the message as it was sent, across tabs for the rendered HTML body, the Text body, the Headers, the Attachments, and the full Raw RFC 5322 ↗ source.

Previously, the Activity log surfaced delivery and authentication metadata but not the message content, making rendering and content issues harder to debug. Message preview closes that gap.

To make messages previewable, turn on Email preview in your sending domain's settings. Previews cover messages sent while the setting is turned on and are retained for about seven days. Sending domains onboarded on or after 2026-07-02 have Email preview turned on automatically.

Refer to Email logs for more information.

This emergency release adds a new managed rule to block active exploitation of a critical remote code execution (RCE) and SQL injection (SQLi) vulnerability found in popular web frameworks.

Key Findings

• Generic Frameworks - Unauthenticated RCE: Attackers can execute arbitrary system commands with web server privileges by sending malicious input containing invalid path sequences during request processing.

• Generic Frameworks - SQLi: Attackers can execute unauthorized database queries due to a failure to sanitize input values within request parameters.

Users in browser-based RDP sessions can now print multiple PDF files as a single print job. Copy the files to your clipboard on the remote machine, then select Print all PDFs in the clipboard panel. The files are combined into one PDF and sent to your local printer.

Bulk print is available in Chromium-based browsers and Firefox. For more information, refer to Print PDFs for browser-based RDP.

Wrangler now includes wrangler flagship, a command suite for managing Flagship apps and feature flags from your terminal.

Create an app and, if you use it from a Worker, add it to your wrangler.json or wrangler.jsonc file as a binding:

wrangler flagship apps create "My Worker App" \
  --binding FLAGS \
  --update-config

Then create flags for the behavior you want to control. Flags can be booleans, strings, numbers, or JSON values:

wrangler flagship flags create <APP_ID> new-checkout

wrangler flagship flags create <APP_ID> checkout-flow \
  --variation control=old-checkout \
  --variation treatment=new-checkout \
  --default control \
  --type string

After a flag exists, change its default variation or use enable and disable commands as kill switches. Existing targeting rules continue to apply unless you change or clear them explicitly:

wrangler flagship flags update <APP_ID> checkout-flow --default treatment
wrangler flagship flags disable <APP_ID> checkout-flow
wrangler flagship flags enable <APP_ID> checkout-flow

For release workflows, use rollout, split, and rules to change exposure without redeploying your Worker:

wrangler flagship flags rollout <APP_ID> new-checkout \
  --to on \
  --percentage 25 \
  --by user_id

wrangler flagship flags split <APP_ID> checkout-flow \
  --weight control=80 \
  --weight treatment=20 \
  --by user_id

wrangler flagship flags rules update <APP_ID> checkout-flow \
  --priority 1 \
  --when "country equals US"

These commands can also be used from CI/CD pipelines, scripts, and AI agents to inspect Flagship state, update flag behavior, or roll back changes through Wrangler.

Refer to the wrangler flagship command reference for the full command guide.

Internal DNS is now generally available. Internal DNS provides authoritative and recursive DNS for private networks on the same global network and control plane you already use for public DNS, Zero Trust, and application services.

• Consolidate DNS operations. Public and private DNS run on one platform, with one API, one audit trail, and one place to set policy.
• Simplify split-horizon DNS. Internal and external resolution are defined as separate views over shared zones, managed from a single control plane — so there is no drift to chase down.
• Extend Zero Trust to DNS. Resolver policies decide which users and devices resolve against which view, enforced by the same Gateway that already governs the rest of your traffic.

Setting up Internal DNS takes three steps: create a zone, create a view, and define a resolver policy.

POST /zones
{
  "account": {
    "id": "<ACCOUNT_ID>"
  },
  "name": "corp.internal",
  "type": "internal"
}

Internal DNS is included with Cloudflare Gateway for Enterprise customers. To get started, refer to the Internal DNS documentation.

You can now subscribe to Email Sending events through Queues event subscriptions and receive outbound transactional email lifecycle events on a queue. Each subscription is scoped to one sending domain — either the zone apex, such as example.com, or a verified sending subdomain, such as send.example.com.

Six event types are published: message.delivered, message.deferred, message.bounced, message.failed, message.rejected, and message.complained. Use them to track deliverability, react to bounces and complaints, and drive suppression or retry logic. Email Routing events are not published on this source.

Each event includes the message details, delivery status, and SMTP response:

{
	"type": "cf.email.sending.message.delivered",
	"source": {
		"type": "email.sending",
		"zoneId": "023e105f4ecef8ad9ca31a8372d0c353",
		"domain": "example.com"
	},
	"payload": {
		"messageId": "0101018f7d0c4d9a-msg-deadbeef",
		"recipient": "user@example.net",
		"terminal": true,
		"delivery": {
			"status": "delivered",
			"smtpStatusCode": "250"
		}
	}
}

Refer to Event subscriptions to see all event types and example payloads.

The legacy Workers KV API routes under /accounts/{account_id}/workers/namespaces/* are deprecated as of July 15, 2026, and will stop working on October 15, 2026. Migrate to the documented Workers KV API routes under /accounts/{account_id}/storage/kv/namespaces/* before that date.

The legacy and replacement routes are interchangeable. They accept the same request parameters and return the same response payloads. To migrate, update the URL path from /workers/namespaces/ to /storage/kv/namespaces/.

Update any integration that calls a route under /accounts/{account_id}/workers/namespaces/ to use the equivalent route under /accounts/{account_id}/storage/kv/namespaces/. The migration is a direct URL path substitution — request parameters and response payloads are identical:

• GET and POST /accounts/{account_id}/workers/namespaces → GET and POST /accounts/{account_id}/storage/kv/namespaces
• GET, PUT, and DELETE /accounts/{account_id}/workers/namespaces/{namespace_id} → GET, PUT, and DELETE /accounts/{account_id}/storage/kv/namespaces/{namespace_id}
• GET /accounts/{account_id}/workers/namespaces/{namespace_id}/keys → GET /accounts/{account_id}/storage/kv/namespaces/{namespace_id}/keys
• GET /accounts/{account_id}/workers/namespaces/{namespace_id}/metadata/{key_name} → GET /accounts/{account_id}/storage/kv/namespaces/{namespace_id}/metadata/{key_name}
• GET, PUT, and DELETE /accounts/{account_id}/workers/namespaces/{namespace_id}/values/{key_name} → GET, PUT, and DELETE /accounts/{account_id}/storage/kv/namespaces/{namespace_id}/values/{key_name}

For more information about the deprecation timeline, refer to API deprecations.

This release introduces new rules targeting critical infrastructure vulnerabilities. These include an unauthenticated memory disclosure flaw in Citrix NetScaler ADC and Gateway (CVE-2026-8451) and a high-severity pre-authentication remote code execution (RCE) vulnerability in Progress Kemp LoadMaster (CVE-2026-8037).

Key Findings

• CVE-2026-8451: An insufficient input validation vulnerability affects Citrix NetScaler ADC and NetScaler Gateway appliances configured as a SAML Identity Provider (IdP). Remote, unauthenticated attackers can exploit this flaw by sending malformed requests to trigger a memory overread, allowing them to leak chunks of sensitive data from adjacent appliance memory.

• CVE-2026-8037: A critical OS command injection vulnerability in Progress Kemp LoadMaster load balancers allows unauthenticated remote attackers to achieve remote code execution (RCE).

Cloudflare Web Analytics (Real User Monitoring) has rolled out performance optimizations to significantly improve the stability and loading speed of account-wide dashboards.

For larger accounts (with >100 Web Analytics sites), loading the aggregate account-wide view would often fail, running into timeouts or unexpected interface errors due to the massive scale of parallel query processing. This update optimizes how high-volume multi-site data is queried to reduce errors and provide a snappier dashboard experience.

Accounts with up to 1,000 sites will now be able to load this account-wide aggregate view without experiencing misleading errors.

If you have an account with over 1,000 sites, we cannot currently aggregate over this volume due to processing constraints but you will now be presented with a clear error and instruction to filter to the relevant site(s) you wish to see the data for.

Platforms can now create temporary preview accounts through the Cloudflare REST API. This lets your platform deploy a live Worker before the user signs in to Cloudflare.

With the Temporary Accounts API, coding agents, AI app builders, and other platforms can build a similar flow for generated Workers and supported resources.

Your platform can keep users in its onboarding flow while they generate, deploy, and test an application. Users do not need an existing Cloudflare account, and your platform does not need write access to one.

The API returns a claim URL that lets the user make the temporary account and its resources permanent.

Cloudflare Drop ↗ demonstrates this preview-and-claim pattern for static sites. Someone can upload a site, test and share it for one hour, then sign in or create an account only when they want to keep it.

This API expands the flow first introduced with wrangler deploy --temporary. Your backend now controls the provisioning and deployment experience directly:

1. Show Cloudflare's Terms of Service and Privacy Policy in your product, and require the user to accept them.
2. Request and solve a proof-of-work challenge.
3. Create a temporary preview account.
4. Deploy with the returned temporary account ID and API token.
5. Show the deployed Worker URL and claim URL to the user.

curl "https://api.cloudflare.com/client/v4/provisioning/previews/challenge" \
  -X POST \
  -H "Content-Type: application/json" \
  --data '{}'

curl "https://api.cloudflare.com/client/v4/provisioning/previews" \
  -X POST \
  -H "Content-Type: application/json" \
  --data '{
    "termsOfService": "https://www.cloudflare.com/terms/",
    "privacyPolicy": "https://www.cloudflare.com/privacypolicy/",
    "acceptTermsOfService": "yes",
    "challengeToken": "<CHALLENGE_TOKEN>",
    "solution": {
      "checkpoints": "<BASE64_CHECKPOINTS>"
    }
  }'

For the complete API flow, proof-of-work requirements, supported products, and limits, refer to Claim deployments (temporary accounts). For the background and design goals behind this flow, refer to Temporary Cloudflare Accounts for AI agents ↗.

Agents connected to Model Context Protocol (MCP) servers with addMcpServer can now handle elicitation ↗ requests.

Elicitation lets an MCP server request user input while it handles a tool call. Form mode collects structured, non-sensitive data. URL mode asks for consent before opening an out-of-band flow, such as third-party authorization or payment.

sequenceDiagram
    participant User
    participant Agent as Agent (MCP client)
    participant Server as MCP server
    participant Browser

    Server->>Agent: elicitation/create
    Agent->>User: Show server, reason, and input or URL
    User->>Agent: Submit, open, decline, or cancel
    Agent->>Browser: Open URL after consent (URL mode)
    Agent->>Server: accept, decline, or cancel
    Server-->>Agent: Optional URL completion notification

Register a handler for each mode your Agent supports in onStart():

import { Agent } from "agents";

export class MyAgent extends Agent {
	onStart() {
		this.mcp.configureElicitationHandlers({
			form: (request, serverId) => this.forwardToUser(request, serverId),
			url: (request, serverId) => this.forwardToUser(request, serverId),
		});
	}

	forwardToUser(request, serverId) {
		// Show the request in your UI and resolve after the user responds.
		throw new Error(
			`Implement elicitation for ${serverId}: ${request.params.message}`,
		);
	}
}

import { Agent } from "agents";
import type { ElicitRequest, ElicitResult } from "agents/mcp";

export class MyAgent extends Agent<Env> {
	onStart() {
		this.mcp.configureElicitationHandlers({
			form: (request, serverId) => this.forwardToUser(request, serverId),
			url: (request, serverId) => this.forwardToUser(request, serverId),
		});
	}

	private forwardToUser(
		request: ElicitRequest,
		serverId: string,
	): Promise<ElicitResult> {
		// Show the request in your UI and resolve after the user responds.
		throw new Error(
			`Implement elicitation for ${serverId}: ${request.params.message}`,
		);
	}
}

Connections advertise only the modes with configured handlers. An Agent without handlers advertises no elicitation capability, which lets the server use its fallback. The SDK stores the advertised modes with each MCP server registration so they survive Durable Object hibernation. Callback functions remain in memory and reattach when onStart() runs.

For implementation details and a browser forwarding pattern, refer to MCP client elicitation. The mcp-client ↗ and mcp-elicitation ↗ examples implement both sides.

To update to this release:

npm i agents@latest

yarn add agents@latest

pnpm add agents@latest

bun add agents@latest

Precursor is rolling out to all customers starting today. Precursor is client-side JavaScript that enables session-based bot detection.

You can read the announcement blog ↗ for background on why we built Precursor and how session-level behavioral detection works.

With Precursor enabled, Cloudflare can:

• Continuously evaluate behavioral signals across a session
• Re-validate challenge clearance as behavior changes
• Update bot scores with session context
• Provide client-side visibility where none previously existed

It integrates with existing protections, including Security Rules, and can be enabled directly from the Cloudflare dashboard with configurable modes to balance security and user experience.

To learn more, refer to the Precursor documentation.

Markdown for Agents now preserves security- and cache-relevant response headers from your origin when converting HTML to Markdown:

• Markdown for Agents preserves security headers such as Strict-Transport-Security (HSTS), Content-Security-Policy (CSP), X-Frame-Options, Set-Cookie, and CORS headers (for example, Access-Control-Allow-Origin) on the converted response.
• Caching headers (Cache-Control, Expires, Age) continue to pass through.

Your origin's Content Signals ↗ policy is now authoritative. If your origin sets a content-signal header, Markdown for Agents preserves it. When the origin does not send one, Cloudflare adds the default Content-Signal: ai-train=yes, search=yes, ai-input=yes.

This release also fixes relative link resolution for directory-style base URLs (those ending in a trailing slash). Previously, relative links such as ../page/ could resolve one path segment too high and return a 404. Links are now resolved correctly per RFC 3986 ↗.

Refer to our developer documentation for more details.

R2 Data Catalog now accepts read-only API tokens, so query engines and clients that only read data no longer need a read-write token. Previously, every catalog operation required an Admin Read & Write token, which meant read-only clients were granted more access than they needed.

You can now authenticate your Iceberg engine based on your workload:

• Read-only operations (such as listing namespaces, loading tables, and querying data) work with an Admin Read only token (R2 Data Catalog read and R2 storage read).
• Write operations (such as creating or dropping tables and committing transactions) continue to require an Admin Read & Write token.

This lets you follow the principle of least privilege — for example, using a read-write token for the pipeline that writes to your tables and read-only tokens for engines like R2 SQL, DuckDB, or PyIceberg that query them.

Note that credentials vended by the catalog inherit the R2 storage permissions of the token used to authenticate. To ensure read-only access to your underlying data, scope the R2 storage permission to read-only as well.

For details on choosing and creating the right token, refer to Authenticate your Iceberg engine.

R2 Data Catalog, a managed Apache Iceberg ↗ catalog built into R2, now automatically optimizes manifest files as part of compaction.

Manifest files track the data files that make up an Iceberg table. As a table accumulates many small or fragmented manifests, query engines must read more metadata during query planning, which slows down queries even before any data is scanned.

When compaction runs, R2 Data Catalog now rewrites and clusters manifest files by partition as a best-effort pre-step. This consolidates fragmented manifests, reduces the number of manifests a query engine must open, and lowers metadata I/O overhead. Tables that are already well-clustered are skipped, so the operation only runs when it provides a benefit.

This happens automatically for tables with compaction enabled — no configuration changes are required.

For more information, refer to Table maintenance.

Data Loss Prevention (DLP) source code detection now focuses on identifying whole source code file uploads and downloads. Previously, source code detection performed partial scans resulting in a higher rate of false positives. Since only whole source code files are evaluated, code embedded in other content — such as chat messages, documentation, or code samples — is no longer flagged as source code, removing a common source of false positives.

Source code detection requires a minimum of 500 characters to evaluate a file. Files below this threshold are not flagged to reduce noise. This threshold filters out small fragments that lack enough context for reliable classification.

Enable and set confidence levels to tune match sensitivity. A higher confidence level reduces false positives by requiring stronger signals that the content is truly source code. A lower confidence level catches more files at the cost of additional noise.

Source code detection applies to standalone source code files in Gateway HTTP policies. It does not detect source code embedded within other file types or payloads, such as .docx files or chat messages.

For more information, refer to Source Code predefined profiles.