Loading story…
Auth0 published guidance for Cross App Access (XAA), describing how enterprise identity providers can centrally authorize AI agents and applications to reach APIs and MCP servers without repeated per-user OAuth consent prompts. In the documented flow, a requesting application authenticates a user with an enterprise IdP such as Okta, receives an ID-JAG identity assertion, and exchanges that assertion with Auth0 for a standard access token that can be used against the target resource.
The implementation aligns with the emerging IETF work on the OAuth identity assertion authorization grant, which defines how identity assertions can be exchanged for OAuth tokens. Auth0 said the model supports both OIDC and SAML environments through a SAML compatibility layer, with Auth0 acting as the resource application's authorization server and Okta serving as the enterprise IdP; the company also outlined setup steps for Okta, Auth0, and its XAA Inspector, and said organizations can scale deployments by publishing integrations through the Okta Integration Network.
See the reporting duties and controls this puts on the clock.
Vulnerabilities, threat actors, malware, products, organizations, and breaches Mallory has linked to this story.
See what this changes for your reporting obligations and which controls it puts on the clock.
2 references tracked. Mallory keeps watching after this page renders.
Map indicators from this story to your assets and identify affected systems in minutes.
Every observed campaign, victim, and pivot linked to actors named in this story.
Malware, exploits, and IOCs connected to the activity described here.
YARA, Sigma, and Snort rules deployed to your SIEM as soon as they’re published.
Get matching new stories delivered to your team as they break — not the next morning.
Ask questions about this story and take action on the answers.
Other stories Mallory is tracking that overlap on entities, sources, or topic labels.
Okta unveiled new **Okta for AI Agents** capabilities at its Oktane conference, including an Agent Gateway for real-time policy enforcement, agent-to-resource governance, and interaction logging; Shadow AI Agent Discovery for Endpoints to identify unmanaged agents on employee devices; third-party identity-provider support; and configuration mapping. Okta plans to add a kill-switch capability to the gateway by the end of 2026, enabling organizations to revoke an agent’s tokens, terminate sessions, or remove granular permissions when behavior becomes anomalous or prohibited. Okta also joined the 12-vendor **Blueprint Alliance** alongside organizations including AWS, Google Cloud, Salesforce, CrowdStrike, Wiz, ServiceNow, and Proofpoint. The alliance will develop an open, multivendor reference architecture focused on locating AI agents, defining their permissions, monitoring their activity, and responding to incidents. Its approach emphasizes centralized identity controls, OAuth token revocation, and broader security telemetry to prevent autonomous agents from retaining excessive or ungoverned access to sensitive enterprise systems.
Sep 25, 2026
Security operations centers (SOCs) are facing unprecedented challenges due to the overwhelming volume of alerts, the technical complexity of modern environments, and the need for rapid, precise responses to threats. Traditional solutions such as detection-as-code, deterministic response automation, and data lakes have helped address some of these issues, but scaling remains a persistent problem. The introduction of AI agents represents a significant shift in how security operations can be managed, offering the potential for intelligent, autonomous analysis that can keep pace with evolving threats and the scale of modern enterprises. AI agents are not only capable of automating routine tasks but also bring a new level of adaptability and decision-making to the SOC, potentially reducing human error and increasing the speed of incident response. However, the adoption of AI agents introduces new complexities, particularly in the realm of identity management. Unlike traditional machine identities, which are deterministic, or human users, who are non-deterministic but slow and easily identified, AI agents combine the speed and scale of machines with the unpredictability of human behavior. This hybrid nature challenges existing identity and access management (IAM) frameworks, requiring the development of new standards and practices. Okta, a leading identity provider, has recognized this challenge and is working on solutions such as the Cross App Access standard to enforce access restrictions on AI agents. The company emphasizes that every AI agent must have a human ultimately responsible for its actions, underscoring the need for accountability in automated systems. The rise of AI agents is driving a reevaluation of best practices in identity security, as organizations must now manage identities that do not fit neatly into existing categories. The unpredictability and autonomy of AI agents mean that traditional controls, such as passwords and periodic verification, may not be sufficient. Security teams must adapt their architectures, expectations, and foundational infrastructure to effectively integrate AI-powered operations. The scale and speed at which AI agents operate also increase the potential impact of any security incident, making robust identity and access controls even more critical. As AI agents become more prevalent in security operations, organizations must balance the benefits of automation and scalability with the risks associated with non-deterministic, autonomous systems. The evolution of SOCs toward AI-driven models is not just a technological shift but also a cultural and procedural one, requiring new approaches to risk management, accountability, and trust. Industry leaders are actively developing new frameworks and standards to address these challenges, recognizing that the future of security operations will be defined by the effective integration of AI agents into both technical and organizational processes. The transformation from traditional SIEM to AI-powered SOCs is underway, and the security community is responding with innovation in both technology and governance.
Jun 29, 2026
Major identity and access management vendors are introducing new platforms and making strategic acquisitions to address the security challenges posed by the rapid adoption of autonomous AI agents in enterprise environments. Ping Identity has launched "Identity for AI," a platform designed to register, manage, and monitor AI agents as non-human identities, providing features such as agent authentication, authorization, least-privilege enforcement, real-time monitoring, and threat detection. This solution aims to ensure accountability and compliance as organizations scale their use of agentic automation, with general availability expected in early 2026. Meanwhile, Twilio has acquired Stytch to develop an intelligent identity layer that verifies trust between humans and AI agents in real time. Additional industry developments include Nuggets' release of a privacy-preserving verification plugin for ElizaOS and Akeyless' expansion of its security suite with identity provider and privileged access management features tailored for AI agents. Industry leaders warn that unsecured AI agents could become a major source of enterprise breaches, highlighting the urgent need for robust identity and security controls as AI-driven automation becomes ubiquitous.
Jun 29, 2026