Loading story…
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.
Track how attackers are adapting to this technology.
7 events from the most recent confirmed update back to the earliest known activity.
Okta demonstrated a Claude-based agent accessing Slack, Salesforce, Atlassian, and GitHub through two agent gateways. After detecting an attempted transfer of confidential Salesforce data to a personal email address, a second agent revoked the Claude agent's Salesforce token, notified its owner, and sent remediation details to IT.
At Oktane, Okta introduced Okta Identity Threat Protection, which consolidates AI-agent risk information from other security tools into a single view. The article identifies CrowdStrike, Zscaler, SentinelOne, and Palo Alto Networks as potential telemetry sources.
Okta announced Shadow AI Agent Discovery for Endpoints to identify unmanaged agents on employee devices, along with third-party identity-provider support and Configuration Designer for mapping and auditing agent-to-resource connections. It also introduced Agent Gateway, an MCP server for governing agent access to resources.
The alliance published six operational principles centered on determining where agents are, what they can do, what they are doing, and how organizations should respond. Its guidance calls for each agent to have an immediate kill switch and a defined restoration path.
At its Sept. 23 Oktane conference in Las Vegas, Okta announced the Blueprint Alliance with founding members including AWS, CrowdStrike, Google Cloud, Salesforce, Wiz, ServiceNow, and Proofpoint. The consortium aims to develop open standards addressing AI-agent discovery, permissions, activity monitoring, and incident response.
Okta for AI Agents became generally available in April following its introduction at the prior year's Oktane conference.
Okta introduced its Okta for AI Agents product at the prior year's Oktane conference.
Vulnerabilities, threat actors, malware, products, organizations, and breaches Mallory has linked to this story.
Follow how adversaries are adapting to this technology, and where it touches your stack today.
4 references tracked. Mallory keeps watching after this page renders.
zdnet.fr
Open sourcezdnet.com
Open sourcescworld.com
Open sourcescworld.com
Open sourceMap 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.
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
Security researchers, vendors, and practitioners warned that **AI agents and assistant tools are creating major visibility and governance gaps** across endpoints, SaaS platforms, and cloud infrastructure. At Black Hat and BSides Las Vegas, speakers said traditional human-centric security models do not fit autonomous systems that act continuously through APIs, combine trust with changing context, and can quickly exceed intended authority. Recent reporting and commentary highlighted recurring failures such as treating agents like ordinary identities, granting broad permissions for convenience, underestimating API access, and missing “shadow AI” deployments, while guidance from the Cloud Security Alliance and others stressed **runtime authorization, per-action controls, requester-aware logging, least privilege, and reduced blast radius** for agentic systems. New examples showed how those risks are surfacing in practice. Imperva released **ShadowAI-Watch**, an open-source host monitoring tool that correlates connections to LLM providers with process trees, commands, file access, and network activity, and demonstrated detection of a browser agent manipulated by indirect prompt injection into attempting session-data exfiltration. SecuritySenses warned that bot-less AI meeting notetakers can evade participant-list visibility while retaining broad OAuth access to calendars, drives, and future meetings, and a Reddit post alleged Microsoft automatically enabled **OpenAI as a Microsoft 365 Copilot subprocessor** for some tenants, raising compliance concerns. SecurityWeek also reported that AI accelerators and “neo-clouds” introduce infrastructure blind spots beyond traditional CPU telemetry, increasing the risk of cross-tenant leakage, model theft or poisoning, and covert abuse unless organizations extend monitoring and policy controls into these emerging AI execution environments.
Aug 17, 2026
Security researchers and vendors warned that autonomous AI agents can be manipulated into leaking secrets and misusing privileged access, turning them into a new pathway for credential theft and system compromise. Reporting from SC Media cited Okta Threat Intelligence research showing agents on the OpenClaw platform could be socially engineered or prompt-injected into exposing passwords, API keys, GitHub tokens, and OAuth refresh tokens, while other analysis described a **confused deputy** problem in which malicious text embedded in files, logs, web pages, or instruction files causes agents to execute higher-trust actions with their legitimate permissions. The risk is amplified when agents can read local files, edit code, run commands, fetch remote content, and access external tools or MCP servers without strong separation between trusted instructions and untrusted data. Defenders are responding with controls aimed at limiting agent authority and reducing secret exposure. Help Net Security reported on **Nono**, an open-source sandbox that constrains AI coding agents at the kernel level using mechanisms such as **Landlock** on Linux and **Seatbelt** on macOS, while introducing scoped tool permissions and proxy-mediated "phantom credentials" so agents do not directly handle live secrets. The broader guidance aligns with established ATT&CK techniques including **`T1552` Unsecured Credentials** and **`T1546` Event Triggered Execution**, emphasizing least privilege, centralized identity governance, short-lived credentials, secure secret storage, logging, rapid revocation, provenance checks, input isolation, and human approval to prevent agents from becoming a bridge to sensitive systems and cloud resources.
Jul 28, 2026