PostgreSQL 19 has been released, about two months later than usual because of a flurry of AI-reported vulnerabilities the team had to address. A headline feature is read-your-writes consistency on standbys via the new WAIT FOR LSN command.
Kubernetes SIG Network plans to remove the kube-proxy IPVS backend in Kubernetes 1.40. It was deprecated in 1.35, and users should migrate to the nftables backend (GA since 1.33) or iptables.
WordPress 7.2 Release Candidate 1 is available for testing on non-production sites. The release notes in Make WordPress Core detail REST API, jQuery Migrate, and Interactivity API changes.
Plugins that bundle their own copies of @wordpress/ui, dataviews, admin-ui, grid or components must update to the fixed versions (for example components 44.0.0) and release before WordPress 7.2, to fix CSS module class name collisions.
Google reports threat actors now use autonomous multi-agent frameworks such as ARTEX, CAIRN, CyberStrikeAI and PentAGI to run multi-stage intrusions, including building exploits from software patches.
PostgreSQL updates released on 12 November fix 52 flaws, including CVE-2026-6460 (CVSS 8.8), a type confusion bug that lets a user who can create objects run code on the server. Full protection also requires rebuilding installed C-language extensions.
Amazon's threat team captured attacker AI agents hunting exposed AI apps such as LiteLLM, One API and Dify to steal stored AI provider keys, then using the keys to power more searching.
The Trigram campaign impersonates about 50 AI and blockchain brands to trick developers into running an infostealer themselves, per Stormshield. It targets Europe and the US.
CISA added four exploited vulnerabilities to its KEV catalog, including BerriAI LiteLLM (CVE-2026-40217) and Apache Axis (CVE-2026-90436), plus Synology DSM and REDCap.
CISA urges immediate patching of Oracle Java vulnerability CVE-2026-114938 ('Java4Shell'), a memory-corruption bug in the x86-64 HotSpot Base64 decoder that can allow unauthenticated remote code execution.
OpenAI's GPT-Live 1.1 full-duplex voice model is now on Vercel AI Gateway; it can be interrupted mid-reply and transcribes more accurately than the previous version. Existing GPT-Live integrations only need the new model ID, and client delegation lets it hand tasks to any text model on the gateway.
GitHub updated the new pull request Overview page, in public preview since November 5, based on feedback. Changes include a clearer merge status and next steps, and better handling of failed optional checks.
Claire Vo, who runs the AI product ChatPRD, reports organised token fraud: groups cycle free trials and promo codes to resell tokens or run commercial workloads through her agent. Simon Willison notes the abusers also probe which underlying model is in use. ChatPRD is ending free trials entirely.
Researchers from Waterloo, Zhejiang University and KTH found that AI agents retrained on larger sets of prompt-injection attacks became easier to hijack in three of four tests. The automated red-teaming tool that found more working attacks produced the weaker defence, because it kept finding near-duplicates of the same injection.
GitHub published a practical guide to agent skills: reviewable files that capture repeatable engineering work and run on any agent host, while tools, project context and permissions stay host-specific.
In the Codename MDASH pilot, GitHub and Microsoft turned thousands of raw AI-found issues into 99 checked, private reports with suggested fixes for six widely used open source projects. Maintainers have already fixed more than half.
Simon Willison and Prime Radiant describe an external sandbox pattern, where code from coding agents runs in a separate sandbox while the agent stays outside. Three server implementations are released, with an HTTP allow-list proxy and credential replacement, usable with Claude Code, Codex and Pi.
WordPress 7.2 adds a data-wp-html directive to the Interactivity API and gives the router navigation status. Three behaviours deprecated earlier are removed.
A critical unauthenticated deserialization flaw in Apache Roller's XML-RPC endpoint (CVE-2026-129448, CVSS 9.8) is being actively exploited, per the Apache Software Foundation.
Drupal released fixes for a critical core flaw in the installer. Under specific, uncommon conditions an unauthenticated attacker can attempt to reinstall a site, leading to takeover or code execution. Fixed releases for all supported branches shipped on 18 November.
A researcher published proof-of-concept exploits showing that low-privilege accounts can reach memory-corruption bugs in PostgreSQL, MySQL and ClickHouse and run code on the database server. A separate Oracle chain needs no credentials.