The fundamental scale of vulnerability discovery has shifted. Frontier AI models and agentic analysis harnesses are now surfacing bugs across large code bases at a rate that the traditional, ad-hoc disclosure-and-patch model was never designed to absorb, not by Cisco, and not by the operators who run our gear. At the same time, the window between disclosure and exploitation has effectively closed. Manual, one-off advisories at unpredictable intervals are no longer the right tool for the job.
Starting in July, and for the foreseeable future, we are moving to a scheduled, twice-monthly security disclosure model, paired with seven days of advance notification of which technologies will be covered in each release. This is a deliberate, engineered response to a structural change in the threat landscape, not a reaction to any single incident. This is a hardening program run at scale, with the discipline customers expect from infrastructure they depend on.
Scheduled disclosures — 1st and 3rd Wednesdays. Beginning in July, Cisco is reserving the first and third Wednesday of each month for security hardened software publications.
Seven-day advance notice. Seven days before each release, PSIRT will publish the list of technologies and platforms included in that drop. If nothing is planned, there will be no communication. You will know what is coming, on which products, before it lands — so you can pre-stage change windows, lab validation, and maintenance approvals. Cisco is committed to thoughtfully bundling products to minimize overlap in upgrades.
Our core Network Operating Systems products (NOS) are being scheduled as the first products to be released. Core operating system products include Cisco IOS XE, IOS XR, NX-OS, Firepower/ASA, and SD-WAN. Our plan is for the NOS products to be released quarterly. Cisco will not release multiple core NOS products on the same day. Other products may be released more often.
Systemic fixes, not just point patches. Our agentic discovery framework — multiple specialized agents covering static code analysis, live system testing, configuration review, and exploit simulation — runs portfolio-wide. That breadth lets us identify recurring architectural patterns and remediate the underlying class of defect across products, not just the instance that was reported. Security engineers remain in-the-loop for validation, prioritization, and verification.
Bundled and streamlined CVEs. The security hardened releases will not have individual CVEs assigned to each bug as they have pervasive fixes and should be qualified and deployed urgently. Individual CVE assessment and corner-case workarounds will not be manageable. Cisco PSIRT will provide ‘bundled’ CVEs (Common Vulnerability Exposures) tied to CWE categories (Common Weakness Enumerations). For example:
This change to how we assign CVEs is not about sweeping issues away or reducing transparency; it reflects a shift in what keeps customers secure. Assessing security risk CVE-by-CVE and applying point mitigations is no longer fit for purpose. Any release predating our security-hardened versions carries materially higher risk, and that gap will only widen as adversaries use AI to develop exploits at machine speed. The most effective protection is running a current, hardened release, not patching individual findings across older ones.
We remain committed to disclosure and transparency. When a vulnerability warrants an individual CVE assignment, (e.g., requiring compensating controls, known exploitation, or otherwise demands defender action), Cisco will assign a CVE and provide robust details. We recognize this shifts emphasis from per-issue detail toward release-level assurance, but this is where the infrastructure industry must move towards defending against this new landscape.
We’ve listened to and understood the concerns: more findings, more patches, more operational load, and the fear of being exposed to the gap between discovery and deployment. The new model is designed specifically to reduce that pressure, not add to it.
For each release window, PSIRT will provide:
Our disclosure principles, our coordination with the broader security community, and our obligations to customers under existing support contracts are unchanged. The Cisco PSIRT is the gold standard for vulnerability disclosure and will drive this revolution – with significantly expanded tooling and cadence built for the new rate of discovery.
Emergencies will happen. Our process will remain unchanged for responding and working out of our normal release cycles to address security incidents, active exploitation and external discovery of zero-day vulnerabilities.
We are explicitly placing focus on key AI-discovered findings and the resulting systemic hardening ahead of new feature work in the affected platforms. That is a direct trade-off, and it is the right one. Resilient, well-maintained infrastructure is the product. Hardening our software is, for this period, the highest-value engineering work we can deliver to customers.
Furthermore, we are integrating advanced agentic capabilities in secure and responsible ways into our development and testing environments. By leveraging AI-driven testing and automated patching workflows (with security engineers firmly “on-the-loop”) we are accelerating our ability to identify, validate, and deploy fixes with greater speed and precision.
We continue to prioritize efforts to make patching easier across our product portfolio. Our controller platforms include capabilities to deploy patches at scale. Our investment in Live Protect is specifically designed to help organizations bridge the gap between when a vulnerability is uncovered, and the organization can patch.
Cisco IQ provides organizations with the information necessary to understand the security state of their installed base – CVE exposure, hardening, and provides guidance to allow organizations to address the related risks. Additionally, Cisco Services is available to assist organizations in evolving security processes for the AI-era.
This is a transition we have prepared for. The engineering teams, the PSIRT organization, the release infrastructure, and the customer-facing tooling are aligned behind it. The goal is straightforward: get fixes into your hands faster, on a schedule you can plan against, with enough advance notice to deploy them on your terms.
We will continue to refine the cadence, the notification format, and the supporting tooling based on what we hear from you in the first several cycles. Direct feedback from operators has shaped this model, and it will continue to shape how it evolves.
Thank you for the partnership. The work ahead is substantial, but it is the right work, and we are ready to partner with our customers to drive a new standard of security and resiliency.
此内容由惯性聚合(RSS阅读器)自动聚合整理，仅供阅读参考。 原文来自 — 版权归原作者所有。