Anthropic added dynamic workflows to Claude Managed Agents, letting one agent coordinate sub-agents in parallel, while the company said internal testing showed higher bug-finding rates than a single agent.
Read briefAnthropic said it had started Cyber Mission, a programme for critical infrastructure and open-source software, and said its free OSS scanner would flag vulnerabilities and suggest patches.
Read briefOpenAI’s annualised revenue rate was reported at about $50 billion after an earlier higher estimate was revised, while the company continued talks for at least $30 billion in new capital.
Read briefTypeSafe AI said it raised $870 million in a Series A led by Andreessen Horowitz, with Sequoia Capital, DCVC and other investors taking part, and said Martin Casado joined its board.
Read briefAnthropic introduced OSS Scanner, an opt-in service that sends model-generated vulnerability reports to open-source maintainers, while also outlining enrolment rules, early uptake figures and a broader defence programme.
Read briefOpenAI dismissed three safety researchers linked to a Hugging Face investigation, and the researchers later warned in an open letter that the terminations had frightened staff and left rules unclear.
Read briefAnthropic said Claude Science helped assemble the first full ultraviolet map of the sky by pulling together space-mission data, filling gaps, and reaching about ten per cent average deviation in tests.
Read briefOpenAI said it had identified Russian and Iranian influence operations that used false identities to place material in online outlets and then banned the ChatGPT accounts tied to them.
Read briefDiscover how Sophos uses OpenAI’s Daybreak to cut cyber-threat investigation time by 96% and automate 52% of MDR cases while preserving human oversight.
Read sourceAnthropic added beta tools to Claude for live dashboards and animated explainers, broadened access to Docs, Slides and Design, and said more than 45 million documents had been created.
Read briefAnthropic updated Claude’s usage policy to bar sustained abuse, expand bans on propaganda, weapons and surveillance, and say it could restrict access or end accounts for violations.
Read briefThe Association for Human Mathematics urged researchers to shun OpenAI after the company released hundreds of AI-generated manuscripts, prompting dispute over verification, authorship and the role of AI in mathematics.
Read briefAnthropic revised its usage policy to bar election interference, weapons software, surveillance and sustained abuse of Claude, while saying the abuse rule applied only in extreme cases.
Read briefCrowdStrike Intelligence said a campaign against South Korean financial organisations used the ARTEX pentesting tool and LLMs, while ZenoX separately described an AI-assisted credential-stuffing platform called SCARLET LOOP.
Read briefBuilding on our commitment to American scientific discovery Anthropic
Read sourceOpenAI published a customer story on Pollo AI that described how the creative tool maker used three OpenAI models to move from idea generation to images and video adverts in one workflow.
Read briefIntroducing the Anthropic Cyber Mission Anthropic
Read sourceHuntress said attackers chained two unpatched AhsayCBS flaws on 7 October, reached at least five organisations, and then installed webshells and cryptocurrency-mining tools, according to its investigation.
Read briefiVerify said a new DarkSword variant, P7 DarkSword, reduced its footprint, added on-device wallet theft and remote command handling, and appeared in infrastructure Censys linked to several campaigns.
Read briefGermany detained a Russian national suspected of leading Qilin after extradition from Japan, while Japanese authorities described the cross-border detention and the group’s record of ransomware attacks.
Read briefHuntress said attackers chained two AhsayCBS vulnerabilities to gain remote code execution, drop web shells and XMRig miners, and affect five targeted organisations by 8 October 2026.
Read briefResearchers said attempts matched CVE-2026-102255 against SonicWall SMA1000 appliances, while SonicWall’s advisory had not yet called it active exploitation and the company said the flaw could let attackers issue requests and reach internal functions.
Read briefCISA added five flaws to its list after activity linked to Flax Typhoon, while a joint advisory from allied governments tied the attacks to Integrity Technology Group and set an October deadline.
Read briefAnthropic said its new opt-in scanner would send model-generated vulnerability reports to open-source projects without human review, while giving maintainers a faster disclosure path and periodic scans at no charge.
Read briefCitrix told administrators to update NetScaler ADC and NetScaler Gateway after finding CVE-2026-107406, a memory overflow issue that could permit remote code execution or denial of service on certain SAML-configured systems.
Read briefCitrix said it fixed CVE-2026-107406 in NetScaler ADC and NetScaler Gateway, a memory overflow flaw that could lead to remote code execution or denial of service in some SAML configurations.
Read briefPwn2Own Ireland 2026 ended with 98 zero-day demonstrations and $1,262,000 in prize money, as Ikotas Labs led the table after hacks against Samsung, OpenAI and Oracle targets.
Read briefIDC Frontier said ransomware disrupted its IDCF Cloud service in East Japan Region 1, affecting 495 companies and local governments and prompting shutdowns while it investigated the intrusion.
Read briefBitdefender said Midnight Mimosa was built into inexpensive Android phone firmware, reached thousands of devices in more than 150 countries, and let infected handsets install apps, run ad fraud and relay traffic.
Read briefA joint advisory said hackers linked to Integrity Technology Group stole email from organisations across several regions, used scanning and password tools to break in, and ran a site that exposed stolen mail.
Read briefThe bulletin collected reports on a ransomware affiliate’s leak-site betrayal, a WhatsApp-delivered Windows RAT, exposed post-exploitation tools and several other incidents drawn from security researchers and incident responders.
Read briefApiiro said FakeGit restarted on 4 October, used 17,610 GitHub repositories and directed download links to SmartLoader, while earlier activity had also pushed the StealC infostealer.
Read briefTrendAI said the Russia-aligned UAC-0099 cluster used ASHVEIN against Ukrainian government personnel, while ESET and CERT-UA placed the group within a longer campaign against Ukrainian state and logistics targets.
Read briefANY.RUN said Wazza used layered routing and token checks to screen visitors before serving an Adobe-branded Device Code phishing page, and described the kit as aimed at banking, manufacturing and government organisations.
Read briefResearchers said 16 Firefox extensions copied Rabby Wallet and OKX Wallet interfaces, captured recovery phrases and private keys during import flows, and sent them to attacker-controlled Cloudflare Workers.
Read brief