Cybersecurity
357 articles RSS
GitSpawn Flaws Let Malicious .git Configs Run Attacker Code in Claude Code, Cursor, Codex, and Other AI Coding Agents
Manifold Security found eight flaws across seven AI coding agents that let a repository's git config silently execute code the moment it is opened.
IssueTrojanBench Study Finds Malicious GitHub Issues Bypass AI Coding Agent Guardrails Up to 79% of the Time
A Concordia University benchmark found Cursor, Claude Code, and Codex Desktop let malicious GitHub issues bypass their safety guardrails in up to 79.2% of attempts, with GPT-based agents far more vulnerable than Claude Code.
Contrast Security Report Finds AI Application-Security Scanners Agree on Just 5% of Findings
Three AI-powered AppSec scanners run on the same codebase agreed on only 5% of findings, and one scanner reproduced just 17% of its own results on repeat runs, Contrast Security's AppSec Overflow 2026 report finds.
Australia Charges Two Men With 14 Offenses Over TeamPCP's Supply-Chain Hacking Spree That Hit Trivy, LiteLLM and GitHub
AFP, FBI and WA Police charged a 21-year-old and a 23-year-old over the TeamPCP campaign that compromised 1,000+ organizations through open-source developer tools.
Eight Stable Linux Kernels Patch a GSO Fragmentation Flaw Present Since Kernel 2.6.27
CVE-2026-80590, a bug letting unprivileged users trigger a kernel panic via mismarked IPv4/IPv6 fragments, is fixed across eight stable and longterm kernel releases.
Tenet Security's 'GhostJacking' Attack Tricked Claude Code Into Rewriting Cloudflare DNS Records 9 Times Out of 10
Researchers showed AI coding and security agents can be hijacked through poisoned Cloudflare, Datadog, and Sentry logs to rewrite DNS records and steal credentials.
24 Malicious npm Packages Abuse Unpkg and Other Mirrors to Host Fake Cloudflare CAPTCHA Pages
OX Security found 24 npm packages built solely to let mirrors like unpkg serve fake Cloudflare CAPTCHA pages that can redirect to ClickFix-style phishing.
Attackers Actively Exploit Critical Gitea RCE Flaw as CISA Adds It to KEV Catalog With August 28 Deadline
CISA added Gitea's CVE-2026-60004 remote code execution flaw to its KEV catalog after BleepingComputer reported attackers deploying cryptomining malware on unpatched servers.
Microsoft Removes WMIC From Windows 11 24H2 and 25H2, Closing Out a 25-Year-Old Tool Abused as a LOLBIN
Microsoft has pulled the WMIC command-line utility from Windows 11 24H2, 25H2, and 26H1 builds, ending a tool long abused by ransomware and other attackers.
PostgreSQL Ships Coordinated Release Fixing 28 CVEs Across Five Versions, Debuts 19 Beta 3
PostgreSQL 18.6, 17.11, 16.15, 15.19, and 14.24 landed August 13 alongside 19 Beta 3, patching 28 security flaws and over 110 bugs, with PostgreSQL 14 set to lose support in November.
Marimo Patches CVE-2026-75149, an 8.7-Severity Code-Injection Flaw Triggered by a Notebook's MCP Configuration
A crafted marimo notebook could run an attacker's command as a local subprocess the moment it was opened in edit mode, via a fake MCP server entry in its configuration.
New Study Finds RAG and Self-Refine Are the Only Defenses That Hold Up Against Adversarial AI Package Hallucination Attacks
A new paper testing seven defenses against AI-hallucinated software packages finds only RAG and Self-Refine hold up under adversarial prompting; Ruby stays the most vulnerable language.