Cursor disclosed that attackers stole an access key stored at third-party provider Fireworks AI and used it to pull data from a shared storage bucket containing historical AI coding-agent traces. Some traces had unredacted API credentials, which attackers then used to breach other Cursor systems and push malicious versions of Anthropic's Claude Code extension through Cursor's marketplace on October 29, exposing up to 5,000 developers to malware installing SSH backdoors. Cursor says fewer than 0.05% of users had traces exposed and fewer than 0.01% had credentials compromised.
bleepingcomputer.com
· 2026-11-14
A newly highlighted QR code creation tool positions itself as a simple, privacy-focused alternative to existing generators. The tool reportedly works without ads, user tracking, or complicated setup steps, letting people create QR codes for uses like payments, menus, or business cards.
fastcompany.com
· 2026-11-14
A developer published a workaround that uses FreeBSD's Linux binary compatibility layer to run Zed's remote-server component, which currently has no native FreeBSD build. The method enables the Linux kernel module, installs a Linux base package, and spoofs the platform identifier so Zed deploys its Linux server binary to a FreeBSD host over SSH. The author also publicly asked the Zed team to add official FreeBSD remote-server support and offered to help with CI.
hayzam.com
· 2026-11-13
Reports on ResetEra and X claim the download code on some international Grand Theft Auto VI retail boxes is printed on the inside cover and can be read from outside using a strong flashlight, rather than being hidden under a scratch-off coating or paper insert. The reports suggest US copies may be unaffected because their boxes include a backing that blocks the code from view. Engadget says it has asked Take-Two for comment and has not yet received a response.
engadget.com
· 2026-11-13
Speakers and attendees of the AIE Code Summit 2026 released a manifesto outlining values meant to curb low-effort AI-generated code, documentation and messages, dubbed 'slop.' The document, inspired by the 2001 Agile Manifesto, lists five value pairs such as 'outcomes over output' and 'understanding over speed,' and argues slop occurs when AI output is passed along without taking responsibility for its quality.
slopmanifesto.com
· 2026-11-13
Anthropic confirmed that Claude experienced elevated errors across claude.ai, its API, Claude Code, and connectors including Claude Docs and Sheets. The company said it is investigating degraded performance and later reported recovery on claude.ai and the API, though it warned errors could recur while a fix is finalized.
androidauthority.com
· 2026-11-13
On Oct. 29, attackers drained roughly $6.34 billion in cryptocurrency from at least 11 Binance-owned wallets by exploiting a feature of Ethereum's EIP-7702 smart-account standard to install malicious "sweeper" code. Binance's internal systems detected unusual activity within minutes, but the malicious authorization kept functioning for about four hours, ultimately emptying a treasury wallet of roughly $2.5 billion across dozens of cryptocurrencies down to a single ETH coin.
darkreading.com
· 2026-11-13
Buyers who received the newly released Apple TV 4K found the YouTube app malfunctioning out of the box. Some saw a black screen with a QR code and a message stating YouTube was not yet available on the device, while others saw the app running but overlaid with a red bar warning that the app was approved only for testing, not production use.
theverge.com
· 2026-11-13
Anthropic's consumer AI agent Claude Duet, currently a research preview limited to Pro, Max and Ultra subscribers in the US with a waitlist, was tested by a reviewer who connected it to email, calendar, Google Docs and Slack. The agent presents suggestions as swipeable cards and is aimed at professionals and small business owners rather than general consumer tasks like ordering food.
theverge.com
· 2026-11-13
Microsoft's FORGE Lab used an AI-powered vulnerability scanner called Hyenas, built on its MDASH research system, to find several memory-safety issues in Rust's core and std libraries. The team identified three recurring patterns where unsafe code wrongly trusts other code not to panic, trusts user-supplied implementations to behave correctly, or trusts the underlying system, allowing undefined behavior even in programs written entirely in safe Rust.
forge-blog.github.io
· 2026-11-13
A security researcher discovered a previously unknown memory-corruption vulnerability in PostgreSQL's core code, dating back 22 years, that can be triggered through type confusion in composite data types. The flaw was uncovered while investigating a separate exploit in the pg_clickhouse extension, and was developed into a reliable remote-code-execution exploit. The work is detailed in the third post of a series on systemic risks in managed PostgreSQL services.
mehmetince.net
· 2026-11-12
AgentRun is a newly released open-source project that layers self-improvement, durable execution and a management console on top of existing agent frameworks like Claude Code, Codex, OpenCode or pi. It lets agents retain notes, skills or workflows from past runs so they avoid repeating the same research, and it persists run state so interrupted sessions can resume rather than restart. The project is licensed under Apache-2.0 and can run locally via a single command, in a container, or as a server with runners.
github.com
· 2026-11-12