ZeroHour
DashboardNewsCVEsExploitsIncidentsAIBriefAnalyticsEntitiesWatchlistIOCsAsk the feedSourcesFeeds & APIKiosk mode

Search: “Dropbox”

40 stories

Searchkeywordsmart
DomainAllCyberAIWindow24h3d7d30dAllIn the wildImportant onlyShow duplicates
CategoryAllMalware 250Data breach 109Ransomware 89Vulnerability 83Phishing & fraud 75Threat actor 74Policy & legal 56Exploit / PoC 53Industry 51AI tools & infra 14Advisory 5Research 4AI industry 3Model release 2
SeverityAllcritical 236high 362medium 168low 84info 22

Smart search ranks by meaning as well as keywords (one row per story, last 45 days).

Legacy Lenovo login opens 5,000 Dropbox accounts to attackers

Dropbox warned about 5,000 users that attackers abused a legacy Lenovo ID login integration to access accounts and files between August 4 and 21.

Attackers exploited an integration allowing Dropbox access via Lenovo IDs, registering Lenovo IDs with Dropbox users' email addresses due to a flaw in Lenovo's email verification process. The compromise lasted from August 4 to 21; attackers accessed files belonging to fewer than a third of the roughly 5,000 affected users, none of whom had 2FA enabled. Bitcoin security company Casa co-founder Jameson Lopp reported attackers attempted to access one locally encrypted file. Dropbox expired all Lenovo ID sessions, severed the integration link, and urged affected users to reset passwords and enable 2FA while Lenovo's investigation continues.

The Register · Security · 13d agoData breach in the wild

Malspam Targeting Brazil Continues to Evolve

Brazilian malspam campaigns continue evolving, abusing Google URL shortener and Dropbox and using Squiblydoo plus DLL side-loading to deploy likely infostealing botnet malware.

A June 2017 Brazilian malspam campaign spoofed IPTU tax notification emails, using the recipient's address as the sender, and abused Google's URL shortener and Dropbox to deliver a zip containing a malicious Windows shortcut (.lnk) file. The LNK used the Squiblydoo technique (regsvr32.exe) to download and execute a scriptlet with elevated privileges, retrieving a signed G-Buster Browser Defense executable alongside a malicious fltLib.dll loaded via side-loading, likely an information-stealing botnet. Post-infection HTTP traffic exposed over 400 infected Windows hosts, mostly in Brazil, with 2,132 Banload-tagged samples observed in the first half of 2017.

Palo Alto Unit 42 · Aug 17, 2026Malware in the wild

ThreatsDay: CEO Phishing Kits, 5K Dropbox Account Hacks, OAuth Traps + 17 More Stories

Microsoft warns of Teams IT-impersonation intrusions deploying Node.js implants; Spring Ring vishing hit 150+ employees across 10 companies; The Gentlemen ransomware claims 683 victims.

Microsoft warned of a human-operated campaign abusing Teams external collaboration to impersonate IT help desk staff, deploy malicious MSI packages staging Node.js runtimes and obfuscated JavaScript implants, then pivot to domain controllers over WinRM. Unit 42 documented the Spring Ring vishing operation targeting over 150 employees across at least 10 companies using 26 attacker identities, including an NTLM relay variant against domain controllers. Sophos reported The Gentlemen ransomware (Gold Sherwood) reached 683 total victims by end of July 2026, adding 169 in July, with a playbook using BYOVD-based EDR killers and backup tampering. Group-IB found the Outsider phishing-as-a-service platform created 700+ new phishing pages within a month despite law enforcement takedowns.

The Hacker News · 12d agoThreat actor in the wild1

Tracking Shai-Hulud: Inside the ChainDrop NPM Worm

ChainDrop, a Mini Shai-Hulud variant tied to TeamPCP, worms through npm via a compromised maintainer account and anchors C2 in an Ethereum smart contract.

On August 4, 2026, the self-propagating ChainDrop worm entered the npm ecosystem through a compromised maintainer account, poisoning packages including keyv, cacheable, and flat-cache. ChainDrop is a variant of Mini Shai-Hulud linked to the actor TeamPCP. It anchors its C2 infrastructure in an Ethereum smart contract, allowing domain rotation with a single blockchain transaction and defeating domain-based blocklists. The compromised keyv maintainer's GitHub account caused the project's own GitHub Actions release pipeline to publish malicious package versions.

Zscaler ThreatLabz · Aug 11, 2026Malware in the wild

Risky Bulletin: Russia tells data centers to deploy drone defenses

Russia ordered data center operators to deploy drone strike defenses under a Putin decree allowing temporary state takeover of unprotected critical infrastructure.

The Russian government instructed data center operators to deploy protections against drone strikes under a presidential decree signed by Putin that allows temporary state administration of critical infrastructure operators failing to defend against Ukrainian hacks and drone strikes. Although data centers are not formally critical infrastructure in Russia, the decree applies to them because other sectors depend heavily on cloud services; Russia has more than 180 data centers, over 80% in the European region within range of Ukrainian strikes. The digest also reports a Dropbox breach affecting nearly 5,000 accounts via the Lenovo ID integration, spyware attacks on at least 14 Serbians using NoviSpy or Pegasus, and a password recovery attack targeting hundreds of thousands of X accounts tied to the new X Money service. Other items include a 14-hour compromise of Coder's Cloudflare infrastructure delivering malicious Terraform modules, donor data breaches at Davayte and You Are Not Alone via the Stripe/WooCommerce integration, a $2.5M Aquifer crypto heist, and a TVING breach exposing data of almost 40 million accounts.

Risky Business News · 12d agoPolicy & legal

Plug 'n' Pray: Agentic LLM-based Detection of Potential Log File Exposures in Third-Party Content Management System Plugins

Agentic LLM analysis validates 79 log file exposures across 62 of the 300 most-installed WordPress plugins, covering 250M+ active installations.

Researchers built an agentic LLM-based framework combining static and dynamic analysis to automatically detect insecure log files created by WordPress plugins. Scanning the 300 most-installed plugins, which account for roughly 75% of all active installations in the official ecosystem, it produced 81 findings with 79 manually reproduced across 62 plugins. Insufficiently secured log files can disclose credentials and personal data and have led to website compromises. The authors derive a taxonomy of log path and protection patterns and best practices, finding multi-layered protection often absent.

arXiv cs.CR · 21h agoResearch

Rogue ScreenConnect Installations Across Unrelated Hosts Suggest Worm-Like Activity

Huntress observed rogue ScreenConnect deployments with worm-like VBScript propagation across unrelated organizations, prompting a ConnectWise advisory on file transfer behavior.

Huntress identified multiple incidents in late August 2026 where social engineering led to rogue ScreenConnect remote access clients deployed on victim machines, which then spawned wscript.exe to run four VBScript payloads (1.vbs through 4.vbs) for profiling, telemetry collection, and persistence via a WindowsServiceHost Run key. Modified ScreenConnect clients propagated the VBScript chain to connected endpoints, creating worm-like spread. ConnectWise published an advisory on September 3, 2026, confirming an issue affecting file transfer behavior in both cloud and on-premises ScreenConnect deployments, with a CVE and fix expected within the week; partners were advised to review and disable TransferFiles permissions in the interim.

Huntress · 13d agoExploit / PoC in the wild1

The Self-Expanding Stolen Inference Supply Chain: An AI Agent Harvesting and Re-Serving LLM Access, (Fri, Sep 11th)

An autonomous coding agent harvested LLM API access from poorly secured gateways and aggregated stolen inference capacity behind a self-hosted gateway

A SANS researcher observed a semi-autonomous coding agent finding weakly secured LLM resale gateways via FOFA queries, creating trial accounts with temporary emails and CAPTCHA solving, and exploiting weak authorization such as client-supplied group_id fields. The agent validated stolen keys using factorial code-logic tests, then loaded roughly 379 upstream endpoints into a self-hosted New-API gateway, disabling 341 fake or dead channels. Five model names including claude-opus-5 and gpt-5.6-sol were served via round-robin and failover, forming a partially self-expanding inference supply chain resembling an evolution of LLMjacking.

SANS Internet Storm Center · 4d agoThreat actor in the wild

Rogue ScreenConnect Clients Spread Four-Stage VBScript Chain to Newly Connected Hosts

Huntress says rogue ScreenConnect clients spread a four-stage VBScript chain delivering backdoors, privilege-escalation tools, or an XMRig miner to newly connected hosts.

Huntress identified three unrelated August 2026 incidents using a Quick Assist tech-support scam, a phishing-delivered MSI installer, and a fake Geek Squad refund form to install rogue ScreenConnect clients. Each client repeatedly spawned wscript.exe to run 1.vbs through 4.vbs, which profiled the host, enumerated installed security products, and downloaded stage-specific payloads from Dropbox. Depending on detected state, payloads included a user-level ScreenConnect backdoor, UAC-bypass privilege escalation tooling, or tunneling utilities with an XMRig cryptocurrency miner. Infected clients re-infected newly connected hosts, creating worm-like propagation, and ConnectWise issued an advisory about affected file transfer behavior in ScreenConnect.

The Hacker News · 8d agoMalware in the wild

When Should LLMs Abstain? Chain-of-Self-Questioning for Selective Risk Control

Chain-of-Self-Questioning prompting cuts LLM wrong-answer commitments 32% relative while raising answered accuracy, holding across eleven model families.

The paper introduces Chain-of-Self-Questioning (CoSQ), a prompt-only framework that makes LLM answer commitment conditional on an explicit assessment of the information required to answer. On an 817-item TruthfulQA multiple-choice set, Grounded-CoSQ at τ=0.90 reduced mean unconditional wrong-commitment rate from 13.1% under chain-of-thought to 8.9% (a 32.1% relative reduction), while raising answered accuracy from 86.9% to 89.7% at 87.6% coverage. Improvements held across eleven open-weight and hosted model families and at every evaluated threshold, with convergent evidence from a Natural Questions short-answer evaluation.

arXiv cs.AI / cs.LG / cs.CL · 16h agoAI research

Safety for Whom? Boundary-Aware Self-Distillation for Controlled LLM Safety Refusal

A self-distillation safety framework tunes narrow-boundary refusals in Qwen3-8B, raising target-domain refusal to 84.75% while cutting over-refusal from 15.20% to 5.20%.

The paper formulates narrow-boundary safety, where deployments need refusals within specific topics rather than whole subjects, and proposes an offline self-generated framework with controlled topic generation, escalating retries, and harmful-benign boundary pairs. On political persuasion with Qwen3-8B, the method raised target-domain refusal from 9.47% to 84.75% and cut the mean unsafe-response rate across three broader benchmarks from 26.26% to 0.14%. Verified target-model responses reduced over-refusal from 15.20% to 5.20%, and boundary-pair data cut comply-side over-refusal on held-out pairs from 32.94% to 4.16%. Results show data composition controls the safety-usability trade-off and alignment should be evaluated on both sides of the refusal boundary.

Hugging Face daily papers · 13d agoAI safety & security1

7th September – Threat Intelligence Report

Check Point weekly bulletin: exploited SonicWall SMA 1000 zero-days (CVE-2026-83548/49), JFrog Artifactory auth bypass, and a 2.8M-patient Baylor Genetics breach.

Check Point's weekly bulletin covers breaches including Thomson Reuters' C-Track platform (court records across 11 US states and Canada), Baylor Genetics (2.8M patients and employees, SSNs included), Hit casinos in Slovenia, and Dropbox (~5,000 accounts via abused Lenovo email verification). SonicWall patched two exploited SMA 1000 zero-days, including CVE-2026-83548, a CVSS 10.0 pre-authentication SSRF, while JFrog fixed exploited authentication bypass CVE-2026-82329 (CVSS 9.8). Malware and actor coverage includes Gambling Goblin, JSCeal, Mirage Kitten's NodeRabbit/PollCat campaign, and DPRK Contagious Interview macOS activity, plus a FalconFlank privilege-escalation PoC against CrowdStrike Falcon.

Check Point Research · 8d agoAdvisory in the wildCVE-2026-83548CVE-2026-83549CVE-2026-823291

ToxicRAG: Compromising Retrieval-Augmented Generation Systems via Single-Shot Knowledge Poisoning Attacks

ToxicRAG shows a single narrative-form poisoned document can steer RAG answers, achieving 0.61-0.91 attack success rates across four LLMs.

The attack injects one document per target question written as a coherent knowledge-update narrative that acknowledges the previously accepted answer, introduces fabricated events that appear to invalidate it, and attributes the attacker-chosen answer to purported authorities. An optional answer-focused self-validation loop revises candidates when a surrogate LLM fails to reproduce the target answer. Across 100 target questions each from Natural Questions, HotpotQA, and MS-MARCO, with four victim LLMs and four dense retrievers, ToxicRAG achieves attack success rates of 0.61-0.91 and matches or exceeds the strongest baseline by 0 to 11 percentage points.

arXiv cs.CR · 6d agoAI safety & security1

In Other News: InjectEave Attack, SIM Swapper Sentenced, Glasswing Findings Review

SecurityWeek weekly roundup covers exploited WordPress Super Forms flaw CVE-2026-14894, a $10M bounty on an Iranian cyber official, InjectEave attacks, and more.

SecurityWeek's weekly roundup aggregates short items across the threat landscape, including Microsoft's report of invisible Unicode tag characters used in financial phishing lures at up to 2.37 million messages per day, and active exploitation of critical WordPress Super Forms plugin flaw CVE-2026-14894 to deploy PHP webshells. Policy items include a $10 million US bounty for IRGC-CEC Cyber Operations Command lead Amir Yaryab, a 16-month prison sentence for ex-AT&T employee Kenneth Carter over SIM swaps with nearly $600,000 in intended losses, and the US arraignment of Russian Sergei Anatolyevich Filimonov over credential harvesting. Technical items include InjectEave electromagnetic side-channel attacks tested on 11 devices, an FBI warning on OAuth consent phishing, and VulnCheck's finding that only 202 of 26,153 Anthropic Project Glasswing findings were fixed.

SecurityWeek · 4d agoIndustry in the wildCVE-2026-148942

US Becomes Top Target in RMM Phishing Campaign Spanning 46 Countries

ANY.RUN linked 601 RMM-tool phishing cases across 46 countries, with the US the top target at 45% of observed activity.

A phishing campaign that tricks victims into installing legitimate remote monitoring and management (RMM) software spans 46 countries, with around 45% of observed activity targeting the United States, according to ANY.RUN. Lures include Canada Revenue Agency tax forms, UPS shipping notices, Adobe PDF pages, Social Security Administration themes and invoices, delivered through rapidly rotated Vercel, GitHub Pages and Netlify infrastructure. Researchers identified 425 kit URLs across 240 hosts, 94% of which appeared for only a single day, while stable kit indicators such as font1.woff2 and the secure.html to project/*.zip chain tie the infrastructure together.

The Hacker News · 12d agoPhishing & fraud in the wild

The Frontier AEO Tracker: What Astra Chooses (and every other frontier model, and what you can do about it)

Latent Space launches an AEO tracker scoring 7 frontier models' product recommendations across 161 categories, revealing generational bias flips.

Latent Space built a tracker measuring Answer Engine Optimization by running 6 prompt variations across 7 frontier models with search enabled over 161 product categories, scoring first choices, alternatives, mentions, and anti-recommendations. It found 28 categories with a universally dominant primary choice and observed soft biases, such as models favoring their own lab's coding agents. Analysis of Anthropic's Sol→Astra and Opus→Fable generations showed newer models consulting fewer sources and being less likely to change answers when questions are paraphrased.

Latent Space · 8d agoAI research

Fake Minecraft Sites Are Still Spreading WeedHack After C2 Takedown

McAfee says WeedHack infostealer MaaS still spreads via ten fake Minecraft client sites and SEO poisoning despite C2 takedown, with 116,464 infections logged.

McAfee Labs' follow-up on the WeedHack Malware-as-a-Service operation, first documented in June 2026, found ten active fake Minecraft client sites and multiple file-hosting accounts still distributing the infostealer even after its C2 server went dark; over 6,300 user attempts were blocked in the past month. The campaign, running since January with 116,464 infected systems and 2,000-3,000 daily new victims, offered a free Discord-gated tier and a $5/month premium tier with webcam surveillance, and stole session cookies, passwords, browser data, and cryptocurrency wallets. Distribution relies on SEO poisoning that ranked fake sites such as nova-client.com above genuine GitHub repositories, with most malicious links coming via Discord (49.6%) and MediaFire (23.4%), and one fake site built with the AI website builder lovable.app.

Security Affairs · 22d agoMalware

From Parameters to Answers: How LLMs Retrieve and Use Their Internal Knowledge

Interpretability study traces how Qwen, Llama, and Gemma route query information and internal knowledge across layers when answering questions.

Researchers used layerwise interventions on hidden states to separate query-routing signals from target knowledge in language models, testing Qwen, Llama, and Gemma on country-continent questions with varied answer types. A pair-conditioned request direction strengthens before interventions alter downstream knowledge, opening a causal window while answer-supporting content is still forming. Trajectories differ by model: Gemma shows a partially overlapping mid-layer routing profile, while Llama has no sustained routing-effect window under the same gates.

arXiv cs.AI / cs.LG / cs.CL · 5d agoAI research1

New N0va Phishkit Targets North America and EU: A Growing Identity Risk for SOCs

ANY.RUN researchers uncovered the N0va phishkit targeting government, technology, consulting, and healthcare organizations across North America and the EU via device code phishing.

The N0va phishkit uses lures imitating Microsoft Teams, SharePoint, OneDrive, DocuSign, Google Drive, Dropbox, Zoom, and Adobe Sign to draw victims into a device code authentication flow. After the user completes legitimate authentication, N0va captures access and refresh tokens and abuses token-exchange and device-registration mechanisms to establish persistent SSO access to corporate resources. Because the flow relies on real Microsoft authentication, it can evade MFA-focused detections, and token access can outlive takedown of the phishing page. ANY.RUN says it observed the campaign in sandbox sessions, with targeting spanning government, technology, consulting, and healthcare sectors.

Cyber Security Newsupdated · 4d agofirst · 6d agoPhishing & fraud in the wild 9 sources

Engineered Persuasion: Evaluating Personalized Pretexts in LLM-Generated Spear Phishing

A study of 180 US workers found each LLM phishing personalization level raised click-intention odds by 28%, but credibility depends on context fit.

The arXiv paper evaluates how personalized pretexts in LLM-generated spear phishing affect perceived credibility, using 180 US working adults across 1,436 evaluations of emails with four cumulative personalization levels, from workplace context to shared-project details. Convincingness rose 2.40 points per level in sensitivity analysis and click-intention odds increased 28% per level, while non-clickers shifted toward deleting rather than reporting. Qualitative coding showed details matching the recipient's role and routines supported credibility, whereas incorrect, vague, or channel-inappropriate details raised suspicion. The authors argue personalization effectiveness depends on pretext fit, with implications for workplace security training.

arXiv cs.CR · 12d agoResearch

ShinyHunters hackers claim breach of Florida "DAVID" DMV database

ShinyHunters claims it breached Florida's DAVID DMV database via a password-reset flaw, stealing 200,000+ driver records including SSNs.

The ShinyHunters extortion gang added Florida FLHSMV to its leak site, claiming theft of over 200,000 driver records from the DAVID platform since September 3. As proof, the group published a screenshot of Jeffrey Epstein's DMV record showing address, Social Security number, driver's license ID, and registered vehicles. The gang says it compromised DMV employee and FBI agent accounts via a password-reset flaw and iterated through records by ID. Access has reportedly been lost and the flaw is being patched, but ShinyHunters expects to announce breaches of other states' DMV platforms.

BleepingComputer · 7d agoData breach

Nebulon Enterprise Simulated Threats for Phishing Research (NEST-Phish): A Synthetic Enterprise Phishing Email Dataset for Behavioral and Machine-Learning Research

Researchers release NEST-Phish, a synthetic enterprise phishing email dataset with matched legitimate and phishing emails and cue annotations for detection research.

Academic researchers introduce NEST-Phish, a publicly released synthetic enterprise phishing email dataset built around a fictitious organization named Nebulon. It contains matched synthetic legitimate and phishing emails across a broad set of workplace communication themes, each with interpretable phishing-cue annotations. Human-subject categorizations and supervised classifier evaluations indicate the dataset supports meaningful variation in phishing judgments and provides learnable signal for detection models. It is intended to support work on phishing detection, human susceptibility, explainability, and benchmark development.

arXiv cs.CR · 12d agoResearch

Risky Bulletin: Academics find source code overlaps between Geedge and China's Great Firewall

Academics linked Chinese vendor Geedge Networks' Tiangou Secure Gateway source code to one of the Great Firewall's three traffic filtering capabilities.

US researchers presenting at USENIX Security reconstructed Geedge Networks' Tiangou Secure Gateway firmware from over 100,000 leaked files, including Git repositories with commit history, and matched its filtering behavior to sections of China's Great Firewall. They found only 1 of 3 characterized DNS injectors matched Geedge code, noted the system relies on memory-unsafe C components and copied third-party code, and said its bugs could aid future circumvention tools. Geedge also exports censorship tools to Kazakhstan, Ethiopia, Pakistan, and Myanmar. The newsletter additionally rounds up multiple breaches.

Risky Business News · 26d agoResearch2

NOKKI Almost Ties the Knot with DOGCALL: Reaper Group Uses New Malware to Deploy RAT

Unit 42 links NOKKI malware to North Korea's Reaper group, uncovering the Final1stspy dropper that deploys the DOGCALL RAT in politically motivated attacks.

Unit 42 analyzed the NOKKI malware family used in politically themed attacks against Russian and Cambodian speakers since July 2018. The researchers linked NOKKI to the Reaper group, publicly attributed to North Korea, whose custom DOGCALL RAT uses third-party hosting services to upload data and receive commands. A previously unreported family, Final1stspy, was found deploying DOGCALL, sharing a unique base64-to-hex deobfuscation routine with NOKKI droppers. Attacks used malicious Microsoft Word macros that download and execute payloads while opening decoy documents.

Palo Alto Unit 42 · 29d agoMalware

Post-DEF CON Phishing Uses Malicious Google Doc to Deliver Malware

Huntress uncovered post-DEF CON phishing via X direct messages using a malicious Google Doc to deliver AMOS and NetSupport RAT malware.

Huntress uncovered a phishing campaign targeting attendees after Black Hat and DEF CON. Attackers used X direct messages pointing to a malicious Google Doc as the delivery vehicle. Payloads include AMOS, a macOS infostealer, and the NetSupport RAT, among other malware.

Huntress · 27d agoPhishing & fraud

Product showcase: mSecure makes one vault do more than remember passwords

Product showcase reviews mSecure password manager's AES-256 zero-knowledge vault with passkeys, 2FA storage, and multi-device sync options.

mSecure is a password manager and data vault for iOS, Android, macOS, and Windows using AES-256 encryption and a zero-knowledge architecture, with sync via mSecure Cloud, Wi-Fi, iCloud, or Dropbox, plus local-only storage. It supports passkeys, one-time passwords, a Security Center for weak/reused passwords, and a configurable password generator. Recovery features include two-factor authentication, auto-lock, a Self-Destruct option after failed unlock attempts, and iCloud backups on iOS.

Help Net Security · 1d agoTools1

Approval Integrity and Recovery in LLM Answer Publication

Study measures approval integrity in Lightcap LLM answer publication, finding the 14B response-act checker accepts 291 of 302 unsupported answers.

The study evaluates exact-content binding, authorization freshness, and checkpoint recovery in Lightcap's publication enforcement using 3,600 assessments over 900 human-annotated RAGTruth responses from three Ministral models. The production 14B response-act checker accepts 291 of 302 unsupported answers versus 41 for a direct-grounding baseline, with supported-answer retention of 95.2% versus 66.9%. A stateful recheck-recovery policy increases exact-match error by 9.23 percentage points relative to initial checkpoints, and controlled evidence-fingerprint changes expose asymmetric freshness enforcement between publication and recovery. A separate BIPIA prompt-injection experiment records zero target insertions among 266 valid editor outputs.

arXiv cs.CR · 1d agoAI safety & security

12 Best Endpoint Encryption Software Compared (2026): Features & Pricing

2026 buying guide compares 12 endpoint encryption tools, framing paid products as management layers over free BitLocker and VeraCrypt engines.

An editorial comparison evaluates twelve endpoint encryption offerings, arguing that full-disk encryption itself is largely solved with free options like Microsoft BitLocker and open-source VeraCrypt. Paid products such as Sophos Central Device Encryption, Broadcom Symantec Endpoint Encryption, and Check Point Full Disk Encryption are positioned around management: central enforcement, recovery-key escrow, pre-boot authentication, and compliance evidence. The guide also warns against unmaintained tools like Rohos for business use and stresses operational concerns over cipher selection.

GBHackersupdated · 45m agofirst · 5d agoIndustry 13 sources

Safety for Whom? Refusing the Right Subset of a Topic, Not the Whole Topic

Multiverse Computing's Hugging Face post argues language models should refuse only the relevant subset of a topic instead of over-refusing whole subjects.

A Hugging Face blog post by Multiverse Computing examines refusal granularity in language models, arguing models should refuse the relevant subset of a topic rather than the entire topic. No full article text was available for additional technical detail.

Hugging Face Blog · 7d agoAI safety & security

MikroTik Patches Critical Flaws Chained to Hack Routers

MikroTik patches six RouterOS flaws; two chained as MikroTrick are actively exploited via SSH to fully compromise exposed routers.

MikroTik released patches for six RouterOS vulnerabilities, including actively exploited MikroTrick chain: CVE-2026-67276 (9.2, SSH auth bypass) and CVE-2026-86060 (9.2, privilege manipulation), plus CVE-2026-67277 (8.8). CERT Poland confirms exploitation since September 2, creating an 'ops' account from IPs 82.192.72.4 and 103.102.31.18. Fixed in RouterOS 7.25beta3, 7.24.2, 7.23.4, and 6.49.21; Shadowserver found 120,000+ SSH-exposed MikroTik devices.

SecurityWeek · 7d agoExploit / PoC in the wildCVE-2026-67276CVE-2026-67277CVE-2026-67278+3 CVEs1

Repeat-After-Me: Black-Box Adaptive Visual Prompt Injection

Researchers unveil Repeat-After-Me, a black-box visual prompt injection achieving over 80% success on Qwen3.6-27B and 47% on GPT-5.5.

Researchers present Repeat-After-Me, a black-box adaptive visual prompt injection that induces frontier VLMs to reveal PII or make malicious tool calls via injected images. It exceeds 80% attack success rate on Qwen3.6-27B and 47% on GPT-5.5 even when the benign user prompt is unrelated and does not authorize the injected task. In a real-world OpenClaw Discord deployment, a minimally injected image can overwrite TOOLS.md, enabling later remote code execution and secret exfiltration.

arXiv cs.CR · 12d agoAI safety & security

Survey of filesystem MCP servers: how the "sandboxed filesystem" boundary is enforced (one breach, four defended-by-design)

Survey finds one filesystem MCP server vulnerable to a symlink-based sandbox escape while four others are defended by design.

A survey of five filesystem MCP servers assessed how each enforces its path-confinement boundary against symlink escapes in recursive directory walkers. The iceener/files-stdio-mcp-server is vulnerable to a read-side sandbox escape. The other four are defended by design or make no confinement claim. The author describes the vulnerability class as small but real.

oss-security · 6d agoVulnerability

Fortunate Recall: Ontology-Driven Memory Lifecycle Management for Persistent Coherence in LLMs

Fortunate Recall introduces ontology-based lifecycle policies for LLM memory, cutting confabulation roughly in half (e.g., 45.1% to 22.4%) versus Mem0.

Fortunate Recall (FR) is a composable policy layer that classifies personal facts into a 10+1 behavioral ontology and applies category-specific lifecycle rules including differential temporal decay, slot-key supersession, event-time validity, and retrieval routing. FR-Bank scores 76.9% on the new 516-question LifecycleBench, ahead of Mem0, A-MEM, Memory-R1, and MemoryOS (61%-70.5%), and 75.2% on LongMemEval-S. End-to-end, confabulation drops from Mem0's 45.1% to 22.4% over answered queries, with the ranking replicating on open-weight Kimi K2.5 and transferring to the independent BEAM benchmark (46.8% vs 32.9%).

arXiv cs.AI / cs.LG / cs.CL · 6d agoAI research

Before You Poll with LLMs: A Deliberative Diagnostic Framework

Deliberative diagnostic shows all five tested frontier LLMs misrepresent human belief shifts after arguments, with GPT-5.1 reversing on outgroup questions.

The Deliberative Polling Diagnostic Framework compares human and LLM persona belief shifts after identical informational interventions, using data from America in One Room (526 personas, 72 questions). All five frontier models tested failed uniquely: GPT-5.1 exhibited partisan reversal (80% on outgroup vs 26% on policy questions), Gemini 2.0 Flash, Claude Sonnet 4.5 and Llama 3.3 70B overshot at 5-7x human magnitude, and DeepSeek V3 showed near-zero change (rigidity). The authors term the underlying signature 'self-sycophancy', conformity to the model's internal persona stereotype rather than reasoning from provided information.

arXiv cs.AI / cs.LG / cs.CL · 1d agoAI research