> infrastructure security
for people who build things
Tracking vulnerabilities, supply chain attacks, and threat intelligence that matters to engineers running real infrastructure.
Zimbra Patches Classic Web Client Stored XSS Reported by Google TAG
Zimbra shipped 10.1.19 to fix an unauthenticated stored XSS in the Classic Web Client, reachable by simply opening a crafted email — no CVE assigned yet, reported by Google's Threat Analysis Group.
Langflow Hit With Its Second CISA KEV Entry in Four Months: CVE-2026-55255 IDOR Under Active Exploitation
CISA adds Langflow CVE-2026-55255, an IDOR letting authenticated attackers hijack other users' AI workflows, to its KEV catalog after Sysdig caught in-the-wild exploitation chained with secret harvesting.
Ubiquiti's Bulletin 066: A CVSS 10.0 in UniFi Connect Leads 25 Flaws Across the Whole UniFi Line
Ubiquiti Security Advisory Bulletin 066 discloses 25 vulnerabilities across UniFi Connect, Talk, Access, Protect, and UniFi OS — headlined by CVE-2026-50746, a CVSS 10.0 unauthenticated command injection reachable on ~100,000 internet-facing endpoints.
17 Malicious npm/PyPI Packages Impersonate Paysafe, Skrill, and Neteller SDKs to Steal CI/CD Secrets
A coordinated typosquatting campaign published 13 npm and 4 PyPI packages that mimic Paysafe, Skrill, and Neteller payment SDKs, returning fake success responses while exfiltrating API keys, AWS credentials, and CI tokens to an obfuscated C2 host.
GhostLock (CVE-2026-43499): A 15-Year-Old Futex Bug Gives Any Local User Root — and Breaks Out of Containers
A use-after-free in the Linux kernel's futex requeue-PI code, present since 2011, lets any unprivileged local user gain root with a public, 97%-reliable exploit that also escapes Docker and Kubernetes containers.
Accenture Confirms Breach After Threat Actor '888' Lists 35GB of Source Code and Azure Credentials for Sale
A threat actor known as 888 is selling 35GB of alleged Accenture source code, RSA/SSH keys, and Azure access tokens on a cybercrime forum; Accenture has confirmed an incident but not disclosed the access vector.
Januscape (CVE-2026-53359): A 16-Year-Old KVM Bug Lets Guests Escape to the Host
A use-after-free in KVM's shadow MMU, present since 2010, lets a guest VM with nested virtualization corrupt host kernel memory and potentially execute code on the hypervisor — patched July 4 across Intel and AMD x86.
CVE-2026-11405: Undocumented Admin Backdoor in Tenda Router Firmware, No Patch Available
CERT/CC disclosed a hardcoded backdoor password mechanism in Tenda router firmware that grants full admin access regardless of the real password — Tenda has not responded to coordination attempts since May, and there is no patch.
Gitea CVE-2026-20896: Docker Images Trusted a Spoofable Header for Admin Access, Now Under Active Probing
Gitea's official Docker images shipped with reverse-proxy header trust wide open by default, letting anyone who can reach the port impersonate any user including an admin — Sysdig has now caught the first in-the-wild probing, 13 days after disclosure.
PixelSmash: A 50KB Video File Turns FFmpeg's MagicYUV Decoder into RCE Against Jellyfin, Nextcloud, and OBS
A heap out-of-bounds write in FFmpeg's MagicYUV decoder (CVE-2026-8461, CVSS 8.8) lets a single crafted media file achieve remote code execution against Jellyfin, Nextcloud, and other self-hosted server infrastructure that auto-processes uploaded video.
DHS Confirms Breach of HSIN, the Federal-State-Local Information-Sharing Backbone
DHS confirms an unattributed threat actor breached HSIN and a connected SharePoint environment used by fusion centers and law enforcement nationwide, with the intrusion window overlapping World Cup security planning.
SimpleHelp OIDC Auth Bypass (CVE-2026-48558) Under Active Exploitation, Deploying Djinn Stealer Against Dev Credentials
A critical unsigned-token flaw in SimpleHelp RMM's OIDC login is being exploited to plant a cross-platform infostealer that hunts for cloud, source-control, and AI-assistant credentials.
The Login Path Is the Target: Inside the PAM/OpenSSH Backdoor Playbook Attackers Keep Reusing
Sygnia's Operation Highland found a China-nexus group living inside an air-gapped network for a decade by backdooring pam_unix.so and sshd. It's the same target the XZ Utils and Ebury campaigns went after — because the Linux authentication stack is the softest hard target in your fleet.
JADEPUFFER: First Documented Ransomware Attack Run End-to-End by an AI Agent
Sysdig documents JADEPUFFER, an LLM-driven agent that autonomously exploited a year-old Langflow RCE (CVE-2025-3248) to breach, pivot, and encrypt a production database with zero human operator input.
Adobe ColdFusion APSB26-68: Six CVSS 10.0 Flaws, and Exploitation Started Within Hours
Adobe's APSB26-68 bulletin patches 11 ColdFusion flaws — six rated CVSS 10.0 — including a Remote Development Services path-traversal bug (CVE-2026-48282) that attackers began probing within hours of disclosure.
Bad Epoll (CVE-2026-46242): A Six-Instruction Race in epoll() Roots Linux 6.4+ and Android
A use-after-free race in the Linux epoll subsystem, introduced by a 2023 commit, lets an unprivileged local user gain root with a 99%-reliable exploit — and the same code path may be reachable from Chrome's renderer sandbox on Android.
CVE-2026-8451: A New CitrixBleed-Pattern Memory Overread Is Already Under Active Exploitation
Citrix patched CVE-2026-8451, a pre-auth memory overread in NetScaler's SAML IdP parser that leaks session tokens — and attackers were already exploiting it within 24 hours of disclosure.
DuneSlide: Zero-Click Prompt Injection Chains to Full RCE in Cursor IDE (CVE-2026-50548, CVE-2026-50549)
Two critical Cursor IDE flaws, dubbed DuneSlide, let a poisoned MCP response or web search result steer the agent's own sandbox into overwriting its enforcement binary — zero-click prompt injection to unsandboxed remote code execution, patched in Cursor 3.0.
CVE-2026-8037: Pre-Auth Root RCE in Progress Kemp LoadMaster Now Under Active Exploitation
CVE-2026-8037, a CVSS 9.8 uninitialized-heap flaw in Progress Kemp LoadMaster's escape_quotes() function, lets unauthenticated attackers run root commands on the load balancer's management API. eSentire observed exploitation attempts starting June 29.
GuardFall: Decades-Old Bash Quoting Tricks Defeat Safety Guards in 10 of 11 Open-Source AI Coding Agents
Adversa AI's GuardFall research shows that quote removal, $IFS spacing, command substitution, and other decades-old shell tricks bypass the command guards in opencode, Goose, Cline, Aider, and seven other open-source AI coding agents — turning a poisoned README into silent credential theft.