Cloudflare is bringing its secure web gateway closer to its Workers development platform. With Forward, available in public beta, businesses can add code to the processing of certain authorised HTTP requests.[2] For IT teams, this development opens up a concrete possibility: adapting controls to business uses. It also requires this code to be treated as a sensitive component, since an error could disrupt users’ work.
The gateway becomes a point of customisation
A secure web gateway controls users’ access to web services. Its role differs from that of a WAF, which is usually placed in front of an application to protect it. The subject here is the traffic passing through Cloudflare Gateway and the processing the company wishes to apply to it.
The Worker can modify a request or a response, produce its own response or launch a background task. Gateway retains management of the proxy connection, TLS decryption and routing to the service.[2] The company therefore adds application logic to a network path that is already managed.
This flexibility may appeal to a team that needs to apply a specific exception to a business tool. But the need must be clearly defined: adding code also increases the number of behaviours to understand in the event of an incident.
Two mechanisms with different timelines
The announcement of 15 October distinguishes between Forward and Custom Functions. The latter are expected before the end of 2026: they will return a value that can be used in the conditions of a Gateway policy, based on logic specific to the organisation.[1] Their timeline should not be confused with the beta that is already available.
Cloudflare illustrates this approach with pages asking for an access justification, header transformations and notifications. Calls to Workers follow their usual pricing, with the quotas and execution limits associated with the chosen plan.[1] Customisation must therefore be assessed together with its operating cost.
Three questions for choosing the right control
This grid suggests decision criteria for a pilot, without assuming any gains measured in production.
| Need | Expected benefit | Useful check |
|---|---|---|
| Adapt an exchange | Transform a request or a response with Forward.[2] | Does the application still work after the change? |
| Enrich a decision | Introduce a calculated value with the upcoming Custom Functions.[1] | What decision should be taken if the data is missing or out of date? |
| Inform teams | Trigger a contextual notification.[1] | Is the alert useful without exposing confidential content? |
The code remains bound by the gateway’s scope
The documentation states that Forward does not bypass a block rule. Inspection may interrupt the Worker, and WebSocket traffic goes directly to the origin service without passing through it.[2] The presence of a Worker therefore does not prove that all of an application’s exchanges are covered.
The announcement states that Gateway and Workers share the same infrastructure, which avoids a network hop between them.[1] This does not eliminate the execution time of the code, nor that of any external services called. A slow dependency can affect the user experience: the full journey needs to be measured.
An SME can start with a single business use case
Take an illustrative scenario: an SME wants to display a reminder before users access a file-sharing service. A pilot could target one domain and a small test group. The team would check that the reminder appears at the right moment, that connections remain possible and that unaffected traffic continues normally.
Testing should include a standard request, an unexpected format and the unavailability of any service called. Response times should be compared, logs reviewed and a rollback to the previous configuration prepared. This scenario is a suggested test, not a customer deployment reported by Cloudflare.
Professionals design and maintain verifiable controls
This development draws directly on Cybersecurity. Professionals must translate a business requirement into precise behaviour: which traffic to control, which information to use, which action to allow and which record to keep. A rule must be explainable both to those it protects and to those who maintain it.
The work now includes reviewing the code and its dependencies. Permissions, secrets and data written to logs deserve particular attention. Tests must verify both the expected blocking and the continuity of legitimate uses.
Maintenance continues after deployment. A change in an application or an identity can invalidate an assumption. Versioning changes, appointing an owner and keeping a removal procedure make it possible to retain control over the mechanism over time.
Editorial note
A programmable gateway gives more latitude to teams able to maintain their extensions. For a small organisation, the right criterion remains the need being addressed: does customisation provide a useful, verifiable control with an acceptable maintenance burden?
Further reading
Four CCCLX reads to explore infrastructure security and control in more depth.
Open-source security scales up: IBM, Red Hat and Palo Alto join forces
Sandbox, terminal, policies: GitHub puts guardrails around its coding agents
Project Glasswing: Anthropic goes on the offensive against software vulnerabilities
AI voice agents: Oracle details a LiveKit architecture on OCI
Sources
- 01
Gerstenhaber, A. J. and Carryl, A., Cloudflare (15 October 2026). “A fully programmable secure web gateway”. Overview of Gateway customisation with Workers and of the announced Custom Functions.
- 02
Cloudflare One, documentation (updated 15 October 2026). “Integrate Workers”. Documentation on the Forward action, request and response processing, Gateway rules and known limitations.
AI propels Nokia: data centres are reshaping the networking market
The AI explosion is transforming the infrastructure market Nokia is reaping the full benefits of the rise of artificial intelligence and the proliferation of data centres that comes with it. In the second quarter of 2026, the company’s Network Infrastructure business...
Gemini 3.5 Pro delayed: Google under pressure in the AI race
The race for AI models is speeding up Google is going through a difficult period in the competition over advanced artificial intelligence models. According to several reports in the specialist press, the launch of Gemini 3.5 Pro has reportedly been pushed back by...
Project Glasswing: Anthropic goes on the offensive against software vulnerabilities
A new step for AI applied to cybersecurity Anthropic is pressing ahead with Project Glasswing, an initiative that makes its Claude Mythos Preview model available to specialist cybersecurity players in order to identify and fix vulnerabilities in critical software. The...
Cloud, AI, defence: AWS steps up on every front
AWS makes artificial intelligence a pillar of critical infrastructure At the AWS Summit Washington, D.C. 2026, Amazon Web Services unveiled several developments aimed at government departments, public bodies and critical sectors. The company is strengthening its...
MDASH: the AI helping Microsoft secure Windows
Microsoft hands part of the bug hunt to artificial intelligence Cyberattacks are evolving ever faster, forcing software vendors to detect vulnerabilities before they can even be exploited. To meet this challenge, Microsoft has unveiled MDASH (Microsoft Detection and...
Microsoft speeds up its cybersecurity with AI: defence moves to machine scale
Microsoft wants to regain the upper hand against cyberattacks Cybercriminals are now using artificial intelligence to discover vulnerabilities, automate their attacks and speed up their malicious campaigns. In response, Microsoft is pressing ahead with its Secure...
Free access to Claude Fable 5 extended as Anthropic holds off
A new generation of AI designed for specific uses OpenAI is changing its strategy with the launch of GPT-5.6, a new family of models made up of Sol, Terra and Luna. Rather than offering a single model capable of meeting every need, the company is now betting on a...
OpenAI bets on modular AI with Sol, Terra and Luna
A new generation of AI designed for specific uses OpenAI is changing its strategy with the launch of GPT-5.6, a new family of models made up of Sol, Terra and Luna. Rather than offering a single model capable of meeting every need, the company is now betting on a...
Anthropic’s AI is advancing at a pace that worries cybersecurity specialists
AI models cross a new capability threshold Barely a month after its launch, Claude Mythos Preview is already improving its performance. According to the latest tests published by the UK’s AI Security Institute (AISI), Anthropic’s model now beats its own results from...
Lumo AI 2.0 speeds up without sacrificing data protection
Proton wants to prove that a high-performing AI can also be private The artificial intelligence race is no longer played out solely on model performance. User trust and data protection are now becoming major differentiators. With Lumo AI 2.0, Proton is upgrading its...
