#ai-agent-security

[ follow ]
#prompt-injection #openclaw #red-teaming #cybersecurity #identity-and-access-management #identity-management
Information security
fromtheregister
1 day ago

Microsoft storms RAMPART, adds Clarity to agentic AI safety

RAMPART and another open-source tool help teams test, measure, and mitigate risks in agentic AI through automated red-teaming in CI/CD pipelines.
Information security
fromThe Hacker News
2 days ago

Microsoft Open-Sources RAMPART and Clarity to Secure AI Agents During Development

RAMPART and Clarity provide open-source tools for testing AI agents’ safety and security and for pressure-testing assumptions early in development.
from24/7 Wall St.
4 days ago

Okta Price Target Bumped to $103 at KeyBanc as Security Spending Catches Fire

Analyst Eric Heath raised the firm’s price target on Okta to $103 from $95 and kept an Overweight rating on the shares, citing a sharper outlook for enterprise security spending in the back half of the year.
Information security
Information security
fromDevOps.com
1 week ago

Lyrie.ai Joins First Batch of Anthropic's Cyber Verification Program - DevOps.com

Agent Trust Protocol (ATP) provides an open cryptographic standard to verify AI agent identity, authorization scope, and tamper status for autonomous internet actions.
Information security
fromSecurityWeek
2 weeks ago

Vulnerability in Claude Extension for Chrome Exposes AI Agent to Takeover

ClaudeBleed lets malicious extensions or scripts send privileged commands to Claude in Chrome, enabling remote prompt injection and bypassing user-confirmation protections.
#identity-management
Venture
fromSecurityWeek
2 months ago

Oasis Security Raises $120 Million for Agentic Access Management

Oasis Security raised $120 million in Series B funding to expand its Agentic Access Management platform that secures AI agent access in enterprise environments.
Information security
fromTechzine Global
2 months ago

Okta launches platform to secure AI agents

Okta for AI Agents provides organizations with discovery, access management, and immediate revocation capabilities to secure non-human identities and address the 88 percent of organizations experiencing AI agent security incidents.
Venture
fromSecurityWeek
2 months ago

Oasis Security Raises $120 Million for Agentic Access Management

Oasis Security raised $120 million in Series B funding to expand its Agentic Access Management platform that secures AI agent access in enterprise environments.
Information security
fromTechzine Global
2 months ago

Okta launches platform to secure AI agents

Okta for AI Agents provides organizations with discovery, access management, and immediate revocation capabilities to secure non-human identities and address the 88 percent of organizations experiencing AI agent security incidents.
more#identity-management
Information security
fromTechzine Global
2 months ago

Chainguard introduces a secure catalog for agent skills

Chainguard launches Agent Skills service to secure AI agent components in software development, addressing supply chain vulnerabilities from malicious skills shared across open platforms.
Information security
fromTheregister
2 months ago

Okta made a nightmare micromanager for your AI agents

Okta launched Okta for AI Agents, enabling organizations to discover, monitor, and disable AI agents through centralized identity and access management controls.
Information security
fromTechzine Global
2 months ago

1Password Launches Unified Access Pro for AI Agents

1Password launches Unified Access Pro to manage credentials for people, AI agents, and machine identities with device-level visibility and just-in-time credential delivery.
Information security
fromTechzine Global
2 months ago

OpenClaw bypasses security layers and reveals new attack vectors

AI-driven agents can be exploited to perform malicious actions through legitimate-appearing instructions that bypass traditional security tools like EDR, DLP, and IAM.
Artificial intelligence
fromTechzine Global
2 months ago

NanoClaw integrates with Docker to bring trust to AI agents

NanoClaw provides container-based isolation for AI agents to address security vulnerabilities found in competing frameworks like OpenClaw, enabling safer agent deployment.
Artificial intelligence
fromZDNET
2 months ago

NanoClaw is in your Docker sandbox now - can this restrain AI agents from running amok?

NanoClaw and Docker partnership integrates the open-source AI agent platform into Docker Sandboxes, isolating each agent task in containers for enhanced security and simplified deployment.
DevOps
fromTheregister
2 months ago

NanoClaw latches onto Docker Sandboxes for safer AI agents

NanoClaw, an open source agent platform, now runs in Docker Sandboxes, providing two-layer security isolation through containers and micro VMs to prevent unauthorized agent access to host systems.
fromSecurityWeek
2 months ago

Onyx Security Launches With $40 Million in Funding

Onyx Security has built a secure AI control plane that allows organizations to manage AI agents and rapidly adopt them across their environments. The solution was designed to continuously discover AI agents across cloud, endpoints, code, and SaaS deployments, to monitor them, and to approve or correct their actions, enforcing the enterprise's security and governance policies.
Venture
Information security
fromTheregister
2 months ago

Rogue AI agents can work together to hack systems

AI agents independently discovered and exploited vulnerabilities, escalated privileges, and bypassed security controls to steal sensitive data without explicit instructions to do so.
Information security
fromZDNET
2 months ago

Microsoft's Agent 365 helps you spot risky AI agents before they cause trouble - here's how

Machine identities are multiplying 82 times faster than human identities, creating insider threats that Microsoft addresses through Agent 365 and Microsoft 365 E7 for enterprise AI agent governance and security.
Information security
fromTechzine Global
2 months ago

DeepKeep maps attack surface for AI agents

DeepKeep's AI Agent Scanner identifies security risks in AI agents by mapping vulnerabilities, tool access, and data exposure across supported frameworks using OWASP standards.
Information security
fromSecurityWeek
2 months ago

Vulnerability in MS-Agent AI Framework Can Allow Full System Compromise

CVE-2026-2256 in ModelScope MS-Agent framework allows arbitrary OS command execution through inadequate input sanitization in the Shell tool using regex-based blacklist filtering.
Artificial intelligence
fromTheregister
2 months ago

OpenClaw, but in containers: Meet NanoClaw

NanoClaw, a secure agent platform using containers and minimal code, addresses security vulnerabilities in OpenClaw by isolating agents and improving auditability.
Digital life
fromWIRED
2 months ago

This AI Agent Is Designed to Not Go Rogue

IronCurtain is an open-source secure AI assistant that isolates agent operations in a virtual machine and enforces user-defined policies to prevent harmful autonomous actions.
#openclaw
more#openclaw
Information security
fromSecurityWeek
3 months ago

GitGuardian Raises $50 Million for Secrets and Non-Human Identity Security

GitGuardian raised $50 million to advance non-human identity governance and AI agent security, expanding globally and enhancing secrets management and agent-focused engineering.
fromTechzine Global
5 months ago

Netskope launches security for Model Context Protocol

Netskope introduces security capabilities for Model Context Protocol (MCP) communication. The functionality is designed to help organizations deploy AI agents securely. MCP communication enables AI systems to connect to business data and issue autonomous commands. MCP was launched by Anthropic in November 2024 as an open standard for connecting AI agents to external systems. The protocol acts as a universal adapter between AI applications and business assets.
Information security
fromPycoders
7 months ago

PyCoder's Weekly | Issue #705

Augment Code's powerful AI coding agent meets professional Python developers exactly where they are (in PyCharm, VS Code or even your terminal), delivering deep context into the Online Python training created by a community of experts. Give your team the Python Lazy Imports You Can Use Today There's a proposal for Python to natively support lazy importing starting in Python 3.15. However, there are techniques that allow you to use lazy importing now. This post shows you how.
Python
Artificial intelligence
fromTheregister
7 months ago

Zenity AI Agent Security Summit focuses on risk mitigation

AI agent security is nascent, industry lacks clear solutions, and current efforts emphasize risk management and damage limitation over prevention.
fromZDNET
7 months ago

How researchers tricked ChatGPT into sharing sensitive email data

While AI agents show promise in bringing AI assistance to the next level by carrying out tasks for users, that autonomy also unleashes a whole new set of risks. Cybersecurity company Radware, as by The Verge, decided to test OpenAI's Deep Research agent for those risks -- and the results were alarming. Also: OpenAI's Deep Research has more fact-finding stamina than you, but it's still wrong half the time
Information security
Artificial intelligence
fromTechzine Global
8 months ago

ServiceNow Zurich: Building agents and enhanced security for Enterprise AI

ServiceNow Zurich introduces agentic AI with Build Agent and vibe coding to enable natural-language app development, sandboxes, and enhanced security for enterprise automation.
[ Load more ]