OWASP GenAI LLM Top 10 2026

OWASP Top 10 for LLM Applications 2026 is the latest community-driven guide to the most critical security risks facing applications powered by large language models. Developed by hundreds of AI security experts, this edition introduces updated rankings, expanded threat coverage, and new research grounded in thousands of real-world AI security incidents. The guide provides practical […]

genai.owasp.org

Keyv and friends compromised in npm supply chain attack

Mini Shai-Hulud malware was injected into keyv and eight related npm packages on August 4, 2026 after an attacker compromised the maintainer's GitHub account

aikido.dev

CosmosEscape: Taking Over Every Azure Cosmos DB | Wiz Blog

Wiz Research details CosmosEscape, a critical vulnerability in Azure Cosmos DB that granted full read/write access to every database. Now fully remediated.

wiz.io

Coding Agent Horror Stories: The 29 Million Secret Problem | Docker

Learn how AI coding agents can expose credentials in supply chain attacks and how Docker Sandboxes keep secrets out of an agent's reach.

docker.com

VMware fixes three critical flaws allowing auth bypass, VM escapes

Broadcom has released security updates to fix five vulnerabilities in VMware vCenter, ESX, Workstation, and Fusion, including three critical flaws that allow attackers to bypass authentication, execute arbitrary code, or escape from a virtual machine to the host.

bleepingcomputer.com

Russian Hackers Exploit Microsoft OWA Flaw to Keep Mailbox Access After Credential Rotation

Russian hackers exploit CVE-2026-42897 in OWA to deploy OWAReaper, a browser implant that persists through credential rotation and device re-imaging.

thehackernews.com

RufRoot: The MCP Bridge Vulnerability That Turns Agents Into Rogue Admins (CVE-2026-59726) - Noma Security

TL;DR Noma Labs found a critical (10 CVSS) vulnerability in Ruflo, an open source AI agent orchestration platform with more than 67,000 GitHub stars at the time of this writing and ranked #2 on MCPMarket. Ruflo ships with a chat UI, agent swarms, persistent memory, and MCP-based tool calling. The platform’s MCP Bridge, the Express.js […]

noma.security

Comparing Open-Source AI Code Security Harnesses

A guide to open-source AI tools for finding code vulnerabilities, comparing exploit generation, skill-boosted auditing, and SAST+LLM hybrid approaches.

semgrep.dev

Claude AI Shared Chats Reportedly Exposed in Google Search Results

Anthropic’s Claude share links appeared in public search results, raising fresh privacy concerns for users who shared sensitive conversations.

cybersecuritynews.com

Pope Official Prayer App Data Leak Exposes 700K+ User Records

An IDOR flaw in the Pope official prayer app data leak exposed the names, plaintext emails, and admin roles of over 700,000 Click to Pray users.

techstory.in

Certighost Exploit Lets Low-Privileged Active Directory Users Impersonate a Domain Controller

Certighost exploit lets a domain user obtain a Domain Controller certificate and reach DCSync through a vulnerable AD CS chase.

thehackernews.com

Top AIs invent same fake PyPl and npm package names

Research reveals that slopsquatting remains a threat to developers using AI to aid coding.

infoworld.com

AgentForger, Part 1: ChatGPT Cross-Site Agent Forgery

Zenity Labs reveals AgentForger, a ChatGPT Workspace Agents vulnerability that enables attackers to create autonomous agents within your organization.

labs.zenity.io

New wp2shell WordPress Core Flaw Lets Unauthenticated Attackers Run Code

WordPress 6.9.5 and 7.0.2 patch wp2shell, a pre-auth core RCE that lets anonymous attackers run code on default installs, even with no plugins.

thehackernews.com

Security Incident Postmortem

On July 11, 2026, an attacker used a stolen npm token to publish malicious versions of the Jscrambler package. We caught it within seconds.

jscrambler.com

OWASP London Chapter Meetup 16-Jul-2026 Live-Stream

YouTube video by OWASP London

youtube.com

Context bombs: stopping AI attackers in their tracks | Tracebit

We planted context bombs, short strings that trip an AI model's own safety guardrails, inside canary secrets in a live AWS environment. Across 5 frontier models and 152 runs, they cut successful attack paths from 91% to 15% - stopping attackers outright as well as detecting them.

tracebit.com

Coordinated AsyncAPI Supply Chain Attack: Miasma RAT Delivered via Compromised CI/CD Pipelines in Two Repositories - StepSecurity

On July 14, 2026 at 07:10 UTC, three packages in the AsyncAPI generator monorepo (@asyncapi/generator@3.3.1, @asyncapi/generator-helpers@1.1.1, and @asyncapi/generator-components@0.7.1) were published to npm carrying an obfuscated dropper that fires the moment the library is loaded, not on install. The packages were published through the project's own legitimate GitHub Actions release workflow and carry valid npm OIDC provenance attestations, because the attacker didn't steal an npm token: they gained push access to the repository's next branch and let the project's real CI/CD pipeline do the publishing for them.

stepsecurity.io

Telegram's t.me domain goes offline after registry suspension

Telegram's t.me domain has been suspended by the .me registry and removed from DNS, breaking all shared links and raising questions for crypto communities.

cryptobriefing.com

jscrambler npm Package Compromised in Supply Chain Attack - ...

A compromised jscrambler npm release added a malicious preinstall hook that runs hidden native binaries on Linux, macOS, and Windows.

socket.dev

OWASP SAMM vs OWASP DSOMM: An Overview And Comparison - Timo Pagel

"An overview and comparison of OWASP SAMM and DSOMM" - Timo Pagel Many practitioners struggle to understand how OWASP Software Assurance Mautiry Model (SAMM) and OWASP DevSecOps Maturity model (DSOMM) differ and where they overlap. In this session, I’ll provide a high-level overview of each model,

youtube.com

GitLost: How We Tricked GitHub’s AI Agent into Leaking Private Repos - Noma Security

TL;DR: Noma Labs discovered a critical prompt injection vulnerability within GitHub’s new Agentic Workflows, allowing an unauthenticated attacker to silently pull data from private repositories by posting a crafted GitHub Issue in a public repository belonging to the same organization as the private repositories. Noma Labs named the vulnerability GitLost.    Introduction GitHub recently launched […]

noma.security

https://lookup.disclose.io

lookup.disclose.io

Indirect Prompt Injection Targets AI Agents | ThreatLabz

ThreatLabz details indirect prompt injection hidden in malicious webpages meant to mislead AI agents performing tasks.

zscaler.com

16-Year-Old Linux KVM Flaw Lets Guest VMs Escape to Host on Intel and AMD x86 Systems

Januscape abuses KVM shadow MMU page reuse to panic x86 hosts, with Kim reporting a controlled full escape exploit behind the unreleased path.

thehackernews.com