LIVE NEWSROOM · --:-- · May 26, 2026
A LIBRARY FOR SECURITY RESEARCHERS

Cisco CVE-2026-20223: CVSS 10.0 Flaw Gives Unauthenticated Attackers Full Secure Workload Access

Post on X LinkedIn
Cisco CVE-2026-20223: CVSS 10.0 Flaw Gives Unauthenticated Attackers Full Secure Workload Access

Cisco has patched CVE-2026-20223, a maximum-severity vulnerability in Cisco Secure Workload (formerly known as Cisco Tetration) that allows an unauthenticated remote attacker to access all sensitive data in the system and make arbitrary configuration changes with Site Administrator privileges. The flaw carries a perfect CVSS v3.1 score of 10.0 — the highest possible rating — and has no available workaround. Patching is the only mitigation.

// 01 CVE-2026-20223: Technical Details

Cisco Secure Workload is a workload protection platform designed for hybrid and multicloud environments. It provides microsegmentation (the practice of dividing a network into fine-grained security zones to limit lateral movement), workload visibility, and zero-trust policy enforcement across on-premises data centres, private clouds, and public cloud deployments. Enterprises use it to monitor east-west traffic between applications, enforce least-privilege network policies, and detect anomalous workload behaviour.

CVE-2026-20223 (classified as CWE-306 — Missing Authentication for Critical Operation) arises from insufficient validation and authentication on internal REST API endpoints. REST API (Representational State Transfer Application Programming Interface) endpoints are interfaces that software components use to communicate over HTTP. When Cisco Secure Workload's internal API endpoints fail to require authentication, any attacker who can reach those endpoints over the network can issue administrative API calls without credentials.

The CVSS v3.1 vector is CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:C/C:H/I:H/A:H, which breaks down as:

  • AV:N (Attack Vector: Network) — exploitable from anywhere over the internet or internal network
  • AC:L (Attack Complexity: Low) — no special conditions, race conditions, or unusual configurations required
  • PR:N (Privileges Required: None) — no authentication or account needed
  • UI:N (User Interaction: None) — fully automated, no victim action required
  • S:C (Scope: Changed) — the vulnerability allows access to resources beyond the vulnerable component, including cross-tenant data
  • C:H / I:H / A:H — complete compromise of confidentiality, integrity, and availability

This vector combination produces a perfect 10.0 and represents a textbook worst-case scenario: any attacker on the network can fully compromise the system with a single unauthenticated HTTP request.

A successful exploit grants the attacker Site Admin privileges — the highest level of access in Cisco Secure Workload. At that level, the attacker can:

  • Read all workload metadata, application communication patterns, and security policies across all tenant environments
  • Modify or delete microsegmentation rules, effectively removing network isolation between application tiers
  • Access sensitive data that would normally be siloed to individual tenant contexts (cross-tenant data breach)
  • Alter workload security policies to permit attacker-controlled traffic or block legitimate communication
  • Potentially use the platform's agent deployment capabilities to push configuration changes to all monitored workloads

// 02 Exploitation Status and Threat Landscape

No PoC (Proof-of-Concept — working exploit code) has been publicly released as of May 22, 2026, and Cisco has not confirmed active exploitation in the wild. The vulnerability is not currently listed in CISA's Known Exploited Vulnerabilities catalog.

However, the absence of active exploitation today does not reduce the urgency to patch. CVSS 10.0 flaws with no workarounds routinely attract weaponized exploit development within days to weeks of disclosure. The simplicity of the vulnerability — unauthenticated API access — means exploit code is likely straightforward to develop. Security researchers and attackers who can access even a test Secure Workload instance can likely reconstruct a working exploit from the vendor advisory description alone.

Cisco Secure Workload is primarily deployed in large enterprise and government environments, making it a high-value target for both ransomware operators seeking to disable security controls before deploying ransomware, and nation-state threat actors conducting espionage. The ability to silently remove microsegmentation policies without triggering alerts is particularly attractive for adversaries attempting lateral movement through otherwise well-segmented networks.

// 03 Who Is Affected

The affected versions are:

| Version Branch | Vulnerable Through | Patched Version | |—|—|—| | 3.9 and earlier | All releases | Upgrade to 3.10.8.3 or 4.0.3.17 | | 3.10.x | Before 3.10.8.3 | 3.10.8.3 | | 4.0.x | Before 4.0.3.17 | 4.0.3.17 | | SaaS deployment | Not affected | No action needed |

Cisco Secure Workload SaaS customers have already been patched on the backend and do not need to take action. Only organisations running the on-premises or private cloud (self-hosted) versions of Cisco Secure Workload are affected.

There are no configuration-based workarounds. The vulnerable API endpoints cannot be disabled or restricted through Secure Workload configuration without losing core functionality.

// 04 What You Should Do Right Now

  • Identify your deployment type. Log into Cisco Secure Workload and check the version displayed in the system settings. Determine whether you are on the SaaS offering or an on-premises cluster. If on-premises or private cloud, you are affected.
  • Upgrade to the patched release immediately.
  • Customers on 3.10.x: upgrade to 3.10.8.3
  • Customers on 4.0.x: upgrade to 4.0.3.17
  • Customers on 3.9 or earlier: upgrade to 3.10.8.3 or 4.0.3.17 (contact Cisco TAC for upgrade path guidance)
  • Consult the official Cisco advisory. The full advisory cisco-sa-csw-pnbsa-g8WEnuy contains the official upgrade path matrix and checksum verification instructions.
  • Network-layer restriction as a temporary measure. While not a complete mitigation, placing Cisco Secure Workload management interfaces behind a firewall or VPN so they are not reachable from untrusted networks reduces exposure while a patch is being scheduled. This does not protect against lateral movement from an already-compromised internal network segment.
  • Monitor for anomalous API activity. Until patched, review Secure Workload API access logs for unexpected calls from unfamiliar source IPs or service accounts. Investigate any unexplained changes to segmentation policies.
  • Rotate credentials. If your Secure Workload cluster has been internet-accessible without network-layer restrictions, assume credentials and API tokens stored in the platform may have been exposed. Rotate all API keys and service account passwords after patching.

// 05 Background: Understanding the Risk

Cisco has issued multiple maximum-severity (CVSS 10.0) advisories in recent years, including vulnerabilities in IOS XE, NX-OS, and now Secure Workload. This pattern reflects the complexity of Cisco's product portfolio and the difficulty of securing software systems that sit at the intersection of network control and data access.

Missing authentication on critical API operations (CWE-306) is a design-level flaw, not a coding mistake. It typically arises when internal APIs intended only for service-to-service communication are inadvertently exposed to external or unauthenticated contexts — a common failure mode in complex distributed systems where component boundaries blur during development. The fact that the vulnerable endpoints are described as "internal REST API endpoints" suggests this may be a case where internal microservices were not properly isolated from the external API surface.

From an attacker's perspective, workload security platforms are uniquely valuable targets. Their function — monitoring and controlling communication between workloads — means a compromised Secure Workload cluster gives an attacker visibility into the entire application topology of an organisation's data centre or cloud deployment. An attacker can map every application, understand what talks to what, and then selectively disable segmentation rules to enable lateral movement without triggering the alerts that Secure Workload is designed to generate.

The cross-tenant capability (S:C in the CVSS vector) is particularly alarming for organisations using Secure Workload in shared infrastructure or managed service provider (MSP) contexts where multiple tenants share a cluster. A single exploit against a shared cluster could simultaneously expose data from multiple unrelated organisations.

Cisco's Secure Workload product page indicates the platform is used across financial services, healthcare, and government verticals — all high-value targets for both financially motivated and nation-state adversaries.

// 06 Conclusion

CVE-2026-20223 is the worst class of vulnerability possible: unauthenticated, remote, zero-interaction, with full administrative impact and no workaround. Organisations running on-premises Cisco Secure Workload must upgrade to version 3.10.8.3 or 4.0.3.17 immediately. Network-layer restriction can reduce exposure in the short term, but patching is the only resolution. SaaS customers need not act.

For any query contact us at contact@cipherssecurity.com

    TE
    Team Ciphers Security

    The Ciphers Security editorial team — practitioners covering daily threat intel, CVE deep-dives, and hands-on cybersecurity research. About us →

    Previous CVE-2026-9111 and CVE-2026-9110: Critical Chrome Bugs Enable Remote Code Execution — Update Now Next Drupal CVE-2026-9082: Highly Critical SQL Injection Flaw Actively Exploited in the Wild

    Latest News

    PyTorch Lightning PyPI Backdoor: ML Supply Chain Audit and Credential Stealer Detection The PyTorch Lightning PyPI backdoor (versions 2.6.2–2.6.3) deployed a credential stealer targeting AWS keys and bro… Dual Ransomware Gang Attack: When ShinyHunters and Qilin Hit the Same Enterprise ShinyHunters and Qilin separately hit Cushman & Wakefield. Learn why dual ransomware gang attack enterprise in… Adversary-in-the-Middle Phishing MFA Bypass: Detecting the 35,000-User Microsoft 365 Campaign AitM phishing bypassed MFA for 35,000 Microsoft 365 users across 26 countries in 48 hours. Sentinel KQL queries and… Iran UAE Cyberattacks Triple: APT34, Mint Sandstorm, and the Critical Infrastructure Defense Playbook UAE breach attempts tripled to 600K/day after Iran conflict escalation. Map APT34, Mint Sandstorm & MuddyWater… Google GTIG: Chinese-Language PhaaS Ecosystem Rivals Russian Underground in Credential Theft Scale Google's Threat Intelligence Group analyzed a dozen Chinese-language phishing-as-a-service platforms now matching R… Anthropic Mythos Finds 23,000 Vulnerabilities in 1,000 OSS Projects — Patching Bottleneck Grows Anthropic's Mythos AI security scanner has identified over 23,000 potential vulnerabilities across 1,000 open-sourc… Underminr: DNS Bypass Flaw Lets Attackers Hide C2 Traffic Behind 88M Trusted Domains The Underminr vulnerability exploits SNI mismatches in shared CDN infrastructure to hide C2 connections behind trus… Project Glasswing: Claude Mythos AI Finds 10,000 Critical Flaws in Widely Used Software Anthropic's Project Glasswing reports Claude Mythos AI found 10,000+ high/critical vulnerabilities in 1,000+ open-s…
    Scroll to Top
    Ad