46 casos · pág. 1/3
Security firm V12 reported on October 22 that the x402 SDKs, an open standard for AI agents paying for API calls over HTTP 402, contained 25 flaws, three rated critical. The most notable was that the default client set no spending limit, so a malicious API could get an agent with an automatic signer to sign a payment for its full balance …
The Wiz Customer Incident Response Team (CIRT) reports a rising pattern of LLMjacking in which threat actors compromise privileged Google Cloud Platform (GCP) users solely to offload the cost of large-scale LLM inference onto the victim. Attackers link external, attacker-controlled GCP projects to the victim's billing account, so all Gemi…
On September 8, 2026, an attacker added a malicious source dependency to Deep-Live-Cam, a Python face-swapping app with about 96,600 GitHub stars. Commit 7895c547 ("chore: update requirements.txt") rewrote 18 dependency entries and added requests @ git+https://github.com/pypls/requests.git. This points installation to a lookalike reposito…
FortiGuard Labs describes an LLMjacking incident in which an AWS account was compromised through a leaked, long-lived IAM access key with AdministratorAccess permissions. The operator subscribed to one or more foundation models through AWS Marketplace and invoked them through Amazon Bedrock, generating inference charges against the victim…
Forescout's Vedere Labs identified 12,976 internet-exposed service observations tied to refrigeration management systems across 6,017 public IP addresses in 99 countries and territories. In October 2026, CERT-FR confirmed that incidents at more than 700 French supermarkets involved Danfoss AK-SM and AK-SC, Copeland and Dixell XWEB systems…
The seller of the StealC infostealer began advertising Stadia on October 16. It is a subscription-based modular remote access trojan that pairs credential theft with hidden desktop control. Varonis Threat Labs obtained a build, reverse engineered it and found StealC code inside the agent. The panel changelog records an upgrade to the Stea…
Researchers at Huntress uncovered a global business email compromise (BEC) staging campaign against Microsoft 365 environments. Between July and September 2026 it affected 79 Huntress customer organizations. The actor used fake sign-in pages, device-code phishing, adversary-in-the-middle (AiTM) techniques and social engineering to steal s…
This weekly roundup covers a cluster of breaches, critical vulnerabilities and law-enforcement actions. Bluesky confirmed more than 4,500 accounts compromised in a phishing campaign using a fake government payment offer. Zeabur will shut down after a second breach this year that exposed over 3.2 million environment variable records and da…
Forescout's Vedere Labs published research on internet-exposed refrigeration control systems after reported October 2026 cyberattacks that disrupted refrigeration at hundreds of French supermarkets, causing temperature fluctuations and spoiled stock. The researchers observed 12,976 exposed service observations across ten product groups, t…
Between July 14 and September 22, 2026, Huntress detected 497 Microsoft 365 sign-in events across 79 victim organizations in a token replay campaign. The attacker replayed post-MFA session tokens stolen through AiTM relays, device-code phishing and social engineering, a pattern Huntress calls "token spraying". Replay traffic came from the…
Straiker reports that a criminal group it tracks as AiPT-67 (operating as Tr4nsHack Team under the name Hexagon Security Labs Mexico) used a multi-agent system called CARLIY0S to coordinate attacks against more than 400 financial institutions and payment providers in Latin America. The system combined commercial AI models, Anthropic's Cla…
ThreatFabric has documented Hirudo, an Android banking remote-access trojan that steals credentials and gives operators remote control of the phone. Its distinguishing feature is replacing selected genuine banking apps with modified copies that interfere with security checks from inside the app. Victims first install a lure app (Shopora, …
Symantec's Threat Hunter Team used an AI honeypot to capture the full instruction set of a malicious autonomous AI agent in summer 2026. The agent is designed to run up to 100,000 malicious actions without human approval: scanning for targets, exploiting vulnerabilities, stealing credentials, deploying cryptocurrency miners and reinfectin…
A ClickFix malvertising campaign uses sponsored Google results for "amazon business" to lure victims to a fake Amazon login page on Google Sites and then to a fake CAPTCHA. The CAPTCHA pastejacks a command into the clipboard that the victim is told to run in the Windows Run box. The command abuses the legacy finger.exe utility to fetch a …
A Chinese-speaking threat actor used AI agents coordinated through the open-source Cairn platform to attempt intrusions against more than 1,200 websites, reportedly compromising sites of over 100 companies, including Deere, SharkNinja and Cummins. Researcher Snorre Fagerland of NorForge found the activity while monitoring internet-exposed…
CrowdStrike reports that the cybercriminal group SLIM SPIDER breached a Brazilian financial institution in late March 2026 to obtain credentials and information tied to digital assets. The group did not attack Pix or the Central Bank directly. It targeted the cloud environments that banks and fintechs use. CrowdStrike also recorded a 171%…
In July 2026, 77 large healthcare data breaches (500+ individuals) were reported to the HHS Office for Civil Rights (OCR), a 10% month-over-month increase and the third consecutive monthly rise. At least 33,485,970 individuals were affected, up 642% from June and 154% above the 12-month average. January-July 2026 totals 67.4 million affec…
OpenCode, the open-source AI agent framework developed by Anomaly, confirmed on October 17 a security breach that occurred the night before and affected its Console service. Console stores each workspace's third-party AI provider credentials centrally. About 1.4% of Console users, those with external providers linked, may have had OpenAI,…
Threat actors are abusing the open source AI platform Cairn, an agent-coordination engine usable as an autonomous penetration testing tool, to automate attacks on hundreds of websites worldwide. AI security firm Norforge found the activity after threat actors left their Cairn panels exposed, following a Gambit Security report on attackers…
Zscaler ThreatLabz's 2026 Ransomware Report describes ransomware operators using generative AI to sharpen phishing, speed up reconnaissance and pick out the most sensitive corporate data to steal. The report records 896.2 TB exfiltrated by the top 10 ransomware groups between April 2025 and March 2026. The article cites a 275.8% annual in…
