Search: “Oracle”

56 stories in the last 30d

Searchkeyword smart Sortlatest hot importance relevance

DomainAll Cyber AIWindow24h 3d 7d 30d AllIn the wild Important only Show duplicates

CategoryAll Vulnerability 12 AI research 10 Exploit / PoC 9 Advisory 8 AI industry 5 Research 4 AI safety & security 4 Model release 3 AI tools & infra 3 Data breach 3 Malware 2 Threat actor 2 Policy & legal 1 Ransomware 1

SeverityAll critical 5 high 15 medium 15 low 3 info 30

Oracle September 2026 Critical Security Patch Update addresses 672 CVEs

Oracle's September 2026 CSPU fixes 672 CVEs across 673 patches, including 104 critical updates, with E-Business Suite receiving the most patches (159).

Oracle released its September 2026 Critical Security Patch Update fixing 672 unique CVEs via 673 security updates across 17 product families, with 104 patches (15.5%) rated critical and 503 rated high. Oracle E-Business Suite received the most patches at 159 (23.6%), followed by Fusion Middleware at 153, of which 78 are remotely exploitable without authentication. The CSPU is a monthly release cycle Oracle introduced in May 2026 between larger quarterly CPUs. Tenable will publish plugins to identify affected systems.

Tenable BlogSep 15, 21:00 UTC · 12h agoVulnerability

U.S. CISA adds maximum-severity Oracle flaw to its Known Exploited Vulnerabilities catalog

CISA added actively exploited CVE-2026-21962, a CVSS 10.0 unauthenticated flaw in Oracle HTTP Server and WebLogic Proxy Plug-in, to KEV with an August 27 deadline.

CISA added CVE-2026-21962 (CVSS 10.0), an improper access control vulnerability in Oracle HTTP Server and the Oracle WebLogic Server Proxy Plug-in, to its Known Exploited Vulnerabilities catalog and ordered federal agencies to remediate by August 27, 2026. The unauthenticated flaw allows remote attackers with network access to create, delete, or modify critical data, potentially gain broad access, and cause a scope change to other systems; affected versions are 12.2.1.4.0, 14.1.1.0.0 and 14.1.2.0.0. CloudSEK honeypot data from January-February 2026 showed widespread exploitation of the flaw alongside older WebLogic RCEs including CVE-2020-14882/14883, CVE-2020-2551, and CVE-2017-10271.

Security AffairsAug 25, 08:50 UTC · 22d agoExploit / PoC in the wildCVE-2026-21962 CVE-2020-14882 CVE-2020-14883+2 CVEs

Actively Exploited Oracle WebLogic Flaw Lets Unauthenticated Attackers Access Critical Data

CISA added actively exploited Oracle WebLogic flaw CVE-2026-21962 (CVSS 10.0) to its KEV catalog, letting unauthenticated attackers access or modify critical data.

CISA added CVE-2026-21962, a CVSS 10.0 improper access control flaw in Oracle HTTP Server and Oracle WebLogic Server Proxy Plug-in, to its Known Exploited Vulnerabilities catalog, citing evidence of active exploitation. Oracle shipped patches in January 2026, and GreyNoise, CloudSEK, and SOCRadar have since reported exploitation attempts, including a lone IP scanning multiple WebLogic, Ivanti, GNU InetUtils, and GLPI vulnerabilities. The flaw is also among several exploited by a China-linked actor delivering the SNOWLIGHT downloader to government and commercial infrastructure in more than 100 countries. Federal civilian agencies must apply fixes by August 27, 2026 under BOD 26-04.

The Hacker NewsAug 26, 05:07 UTC · 21d agoExploit / PoC in the wildCVE-2026-21962 CVE-2020-14882 CVE-2020-14883+2 CVEs1

Oracle Patches 800+ Vulnerabilities in September 2026 Security Update

Oracle's September 2026 Critical Patch Update fixes 800+ vulnerabilities, including over 100 critical flaws and 240+ remotely exploitable without authentication.

Oracle released 673 new security patches in its September 2026 Critical Security Patch Update, resolving 672 unique CVEs across 17 risk matrices plus 130+ additional CVEs. More than 100 flaws are critical severity and over 240 are remotely exploitable without authentication. Oracle E-Business Suite received the largest batch with 159 patches, followed by Fusion Middleware (153, including 78 unauthenticated remote flaws) and Hyperion (102). Oracle reports no exploitation of these specific flaws but warns attackers routinely exploit unpatched Oracle products.

SecurityWeek08:06 UTC · 1h agoAdvisory

Oracle Critical Patch Update, August 2026 Security Update Review

Oracle's August 2026 Critical Patch Update fixes 943 vulnerabilities; Oracle Fusion Middleware and Hyperion received the most patches at 262.

Oracle released its August 2026 Critical Patch Update, addressing 943 security vulnerabilities across multiple product families, including third-party components bundled in Oracle products. Oracle Fusion Middleware and Oracle Hyperion received the highest number of fixes with 262 patches. Several of the addressed vulnerabilities impact more than one product.

Qualys ThreatPROTECTAug 19, 13:49 UTC · 27d agoAdvisory2

Telerik UI Padding-Oracle Bug Chained to Unauthenticated RCE — Public Exploit Released

TantoSec published a working exploit chaining a Telerik UI padding oracle to unauthenticated RCE; Progress patched the flaws in July.

TantoSec's Marcio Almeida released a full exploit chain and tooling for the RadAsyncUpload control in Telerik UI for ASP.NET AJAX, turning a padding oracle (CVE-2026-13182) into unauthenticated RCE via an unguarded type-resolution flaw (CVE-2026-13181, CVSS 8.1). Exploitation requires a rendered RadAsyncUpload handler and an explicit non-default encryption key, and took roughly 127,000 oracle requests (about an hour) in lab testing. Versions 2010.1.309 through 2026.2.519 are affected; Progress fixed the chain in 2026.2.708 on July 8, 2026, and no in-the-wild exploitation has been confirmed. A separate RCE chain in RadPersistenceManager and RadDockLayout (CVE-2026-13185, -13186, -13190) was also patched but has no public exploit.

The Hacker NewsSep 7, 11:20 UTC · 8d agoExploit / PoC CVE-2026-13181 CVE-2026-13182 CVE-2026-13183+5 CVEs1

Oracle security advisory – January 2026 quarterly rollup (AV26-042) – Update 2

CISA added actively exploited Oracle flaw CVE-2026-21962 to the KEV catalog; it allows remote unauthorized access to Oracle HTTP Server and WebLogic Proxy Plug-in.

Canada's Cyber Centre updated advisory AV26-042 on Oracle's January 2026 quarterly rollup, reporting that CISA added CVE-2026-21962 to the Known Exploited Vulnerabilities catalog on August 24, 2026. The flaw affects the Oracle HTTP Server and Oracle WebLogic Server Proxy Plug-in in Oracle Fusion Middleware and may allow a remote attacker to obtain unauthorized access. A public proof of concept has been available since January 21, 2026, and fixes shipped in Oracle's January 20, 2026 advisory covering multiple products.

Canadian Centre for Cyber SecurityAug 24, 18:42 UTC · 22d agoExploit / PoC in the wildCVE-2026-21962

EFI Pairs Without One-Way Puzzles: Oracle Separations from Communication Complexity

Theorists build a classical oracle where one-way puzzles fail yet EFI pairs survive, separating two candidate minimal assumptions of quantum cryptography.

The paper constructs a single classical oracle relative to which one-way puzzles do not exist, even with an unbounded verifier, while an EFI pair survives every classical-query distinguisher holding advice, making one superposition query at the end. Security is proven by reducing adversary knowledge to communication complexity for Vector-in-Subspace, with the superposition query bounded using random matrix theory. Relative to the oracle, quantum polynomial time offers no advantage on tasks with classical inputs and outputs and there is no proof of quantumness, separating the leading minimal assumptions of quantum cryptography.

arXiv cs.CRSep 10, 17:54 UTC · 5d agoResearch

Oracle August 2026 Critical Security Patch Update Addresses 925 CVEs

Oracle's August 2026 CSPU fixes 925 CVEs across 943 patches, including 154 critical fixes; Fusion Middleware gets 262 patches.

Oracle released its August 2026 Critical Security Patch Update on August 18, addressing 925 unique CVEs with 943 security updates. 154 issues (16.3%) carry a critical severity rating. Oracle Fusion Middleware received the most patches at 262, accounting for 27.8% of the total. Oracle introduced the monthly CSPU cycle in May 2026 as an interim release between quarterly Critical Patch Updates.

Tenable BlogAug 19, 00:41 UTC · 28d agoAdvisory

ZDI-26-635: Oracle Outside In Technology PDF File Parsing Integer Overflow Remote Code Execution Vulnerability

ZDI disclosed CVE-2026-60392, an integer overflow in Oracle Outside In PDF parsing enabling remote code execution, rated CVSS 7.8.

The Zero Day Initiative published advisory ZDI-26-635 describing an integer overflow vulnerability in PDF file parsing within Oracle Outside In Technology. Successful exploitation allows remote code execution but requires user interaction, such as opening a malicious file or visiting a malicious page. ZDI assigned the flaw a CVSS rating of 7.8.

ZDI Published AdvisoriesSep 9, 05:00 UTC · 7d agoAdvisory 4 sources CVE-2026-603921

ZDI-26-639: Oracle VirtualBox VMSVGA Heap-based Buffer Overflow Local Privilege Escalation Vulnerability

ZDI disclosed a heap-based buffer overflow in Oracle VirtualBox's VMSVGA component (CVE-2026-71116) enabling local privilege escalation.

Zero Day Initiative published ZDI-26-639, a CVSS 7.5 heap-based buffer overflow in the VMSVGA component of Oracle VirtualBox. Local attackers who already execute high-privileged code on the guest system can leverage the flaw to escalate privileges on affected installations. The vulnerability is tracked as CVE-2026-71116. No exploitation is reported.

ZDI Published AdvisoriesSep 9, 05:00 UTC · 7d agoVulnerability 6 sources CVE-2026-71116

CVE-2026-34486: Apache Tomcat EncryptInterceptor Fail-Open Bypass

Technical analysis shows CVE-2026-34486 is a one-line fail-open regression in Tomcat Tribes enabling unauthenticated RCE via Java deserialization.

CVE-2026-34486 is a fail-open regression in Tomcat's Tribes EncryptInterceptor, introduced while fixing the CVE-2026-29146 padding-oracle flaw; Apache rated it Important and NVD scores it 7.5 High. Decryption failures are now forwarded to an unfiltered ObjectInputStream, and public PoC repositories demonstrate unauthenticated RCE on default port 4000 when clustering, the EncryptInterceptor, a reachable receiver and deserialization gadgets are all present. Affected releases 9.0.116, 10.1.53 and 11.0.20 are fixed in 9.0.117, 10.1.54 and 11.0.21; Tomcat 8.5 is unaffected.

Web discovery (articles for new exploits & KEV entries)Sep 8, 20:09 UTC · 7d agoExploit / PoC CVE-2026-34486 CVE-2026-291461

The Oracle of Delphi Will Steal Your Credentials

Akamai honeypots reveal attackers brute-forcing RDP credentials to download and execute a previously undetected malware family named Trojan.sysscan.

Akamai's deception technology rerouted intruders into honeypots, revealing attacks that brute-force passwords for RDP credentials to access victims. Once connected, the attackers download and execute a previously undetected malware family that Akamai named Trojan.sysscan. The intrusions target exposed RDP services through opportunistic credential guessing.

Akamai BlogSep 8, 15:13 UTC · 7d agoMalware in the wild

Coupled Calibration and Learning: Mitigating Teacher Bias in LLM Distillation without Target-Domain Reward Feedback

CCL couples teacher calibration with student updates via token-level branching, provably removing teacher bias in LLM distillation.

The paper proposes Coupled Calibration and Learning (CCL), an LLM distillation algorithm that alternates teacher calibration using source-question reward feedback with student training on target questions under covariate shift. Each iteration calibrates the teacher on source feedback, trains the student on target questions, and lets the updated student inform subsequent calibration. The authors prove the student's expected KL divergence to the oracle student converges to zero at a polynomial rate, and show regularized direct matching error can remain bounded away from zero.

arXiv cs.AI / cs.LG / cs.CLSep 15, 17:15 UTC · 16h agoAI research

USN-8769-1: phpseclib vulnerability

Ubuntu patches phpseclib non-constant-time padding validation enabling remote padding oracle timing attacks against AES-CBC.

Ubuntu security notice USN-8769-1 addresses a vulnerability in phpseclib where padding validation was not performed in constant time when using AES in CBC mode. A remote attacker could leverage this timing side channel to conduct a padding oracle attack and decrypt sensitive information. Users are advised to update the phpseclib package.

Ubuntu Security NoticesSep 15, 15:59 UTC · 17h agoAdvisory

CVE-2026-60163: MySQL Group Replication unauthenticated remote arbitrary SQL execution

Oracle reclassified MySQL Group Replication flaw CVE-2026-60163 from Local to Adjacent Network; researcher says it enables unauthenticated remote arbitrary SQL execution.

CVE-2026-60163 in MySQL Group Replication permits an unauthenticated remote attacker to execute arbitrary SQL on the destination server. Following discussion after the July disclosure, Oracle revised the CVSS attack vector rating from 'Local' to 'Adjacent Network'. The researcher posting to oss-security clarifies that, in the extreme, the attack surface is reachable fully remotely.

oss-securitySep 15, 01:49 UTC · 1d agoVulnerability CVE-2026-601631

Beyond the Perimeter: Building Resilience Against Cloud and SaaS Supply-Chain Attacks

ShinyHunters exploited an Oracle PeopleSoft zero-day to steal data and extort roughly 100 organizations, including the Council of Europe, for up to $2.3M.

Between May and early June 2026, the ShinyHunters group exploited a critical zero-day in Oracle PeopleSoft across about 100 organizations and 300 instances worldwide, per reports cited by The Register. Stolen records included employee and student personal data, payroll, tax, financial and health information, plus immigration and passport documents. AgentCypher.ai estimates extortion demands of $400,000 to $2.3 million per victim, typically in Bitcoin; the Council of Europe refused to pay. The article uses the incident to argue for Zero Trust, supply-chain risk management, rapid patching, encrypted distributed backups and defined recovery-time objectives.

Cyber Security NewsSep 12, 10:46 UTC · 3d agoData breach in the wild1

When LLM Decompilers Recompile More and Preserve Less

Researchers show LLM decompiler outputs can recompile yet diverge behaviorally, proposing the Decompile-Diverge fuzzing oracle to catch hidden changes.

The paper demonstrates that LLM-based decompilers can produce code that recompiles and passes all shipped tests yet diverges on other legitimate inputs—4.9% overall and up to 13% for one system—and can make disclosed vulnerabilities vanish without a visible crash. Across 300 real GitHub functions and 287 CVE-grounded functions, a refinement LLM lifted Ghidra's build rate from 75% to 90% while Matched rate fell from 74% to 62%, with up to one tenth of vulnerabilities showing Crash Absence. Decompile-Diverge detects these gaps by synthesizing drivers, growing fuzzing corpora from the reference, and rerunning decompiled code on identical inputs.

arXiv cs.CRSep 4, 17:16 UTC · 11d agoResearch

Weekly Cybersecurity Newsletter – Top 50 Biggest Cybersecurity Stories of the Week

GBHackers weekly digest rounds up 50 stories including Microsoft's 973-CVE patch drop, exploited Cisco FMC flaws, and Claude agent attacks.

GBHackers' September 7-12, 2026 newsletter summarizes the 50 biggest cybersecurity stories of the week. Highlights include Microsoft patching a record 973 CVEs with two exploited zero-days, active exploitation of Cisco FMC, Check Point VPN and Ivanti flaws, China-linked crews chaining Chrome and Windows zero-days, AI agents mass-exploiting PaperCut to compromise 440 servers, and the emergence of Panzer cross-platform ransomware. It also covers Anthropic and OpenAI agentic AI incidents and CrowdStrike's SafeMind launch.

GBHackersSep 15, 11:44 UTC · 21h agoIndustry in the wild

Mind2Dialogue: Training Human-Aware Language Models by Simulating User Mental States

Mind2Dialogue simulates users' mental states to generate privileged supervision, boosting personalization and preference-following in Qwen, Llama, and OLMo assistants.

The Mind2Dialogue framework uses a psychology-guided simulator that preserves personal characteristics while updating user mental states through interaction, driving coherent conversations and an Oracle assistant's responses. Privileged distillation trains models on the Oracle's well-informed responses so they can assist users without direct access to mental states at deployment. Training on the full corpus improves every reported personalization metric over Qwen, Llama, and OLMo instruction-tuned baselines, including 26.6 to 40.9 percentage point gains in preference-following generation.

Hugging Face daily papersupdated Sep 14, 17:55 UTC · 1d agofirst Sep 13, 20:00 UTC · 2d agoAI research 2 sources

Beyond Solver Verdicts: Generative Reward Models for Autoformalization

Researchers introduce Generative Verification (GenV), a generative reward model achieving 0.961 AUROC in detecting unfaithful autoformalization that preserves solver verdicts.

The paper formalizes Verdict-Preserving-Unfaithfulness (VPU), a failure mode in neurosymbolic autoformalization where an incorrect encoding executes successfully and matches the expected solver verdict, and proves verdict-only verification is bounded to chance-level detection. The proposed Generative Verification (GenV) distills an offline Z3-equivalence oracle into a reference-free, continuous reference-equivalence score within the language model's vocabulary space. The oracle-mined verifier (GenV+HN) achieves 0.961 AUROC, generalizes zero-shot across unseen translators and formal styles, and yields an 11.3-point downstream accuracy gain in agentic test-time compute allocation. Mechanistic analysis with decision-projected logit lenses and sparse autoencoders shows the generative readout extracts precise spatial error coordinates without explicit localization training.

Hugging Face daily papersSep 9, 20:00 UTC · 6d agoAI research1

August 2026 CVE Landscape

Insikt Group catalogs 73 high-impact August 2026 CVEs (43 Very Critical), including PaperCut, Zimbra, and Metabase flaws actively exploited or weaponized.

Recorded Future's Insikt Group identified 73 high-impact vulnerabilities in August 2026, 43 rated Very Critical, spanning 45 vendors with Microsoft accounting for roughly 11%. 31 vulnerabilities surfaced via CISA's KEV catalog, with others validated via open sources, vendor telemetry, and honeypot data. New Nuclei detection templates were released for CVE-2025-62593 (Ray), CVE-2026-72898 (Metabase), and CVE-2026-9198 (IBM Langflow). The report also highlights two AI-assisted operations: UAT-10147 exploited Zimbra, AjaxPro, Nacos, and Telerik servers before using DeepAudit and PentestGPT post-compromise, while a separate Chinese-speaking actor weaponized Hermes Agent and DeepSeek in a failed attempt.

Recorded FutureSep 8, 00:00 UTC · 8d agoVulnerability in the wildCVE-2025-62593 CVE-2026-72898 CVE-2026-9198+4 CVEs

Meta FAIR Introduces AI Research Preference Models (RPMs): Ranking ML Experiments Before Spending GPU Hours

Meta FAIR, Oxford and UCL introduce Research Preference Models that rank unexecuted ML experiments, lifting AIRS-Bench scores from 0.684 to 0.729 and cutting compute ~1.6×.

Researchers from Meta FAIR, Oxford, and UCL introduce Research Preference Models (RPMs), which use frozen pretrained LLMs (Qwen3.6-27B backbone, no fine-tuning) to rank unexecuted experiment candidates and execute only the winner of a pairwise knockout tournament. Two variants shipped: an inference-only LLM-as-a-judge and an agentic variant that runs small pilot experiments in an H200 sandbox. On AIRS-Bench (20 tasks, 24 hours on one H200, 10 seeds), scores rise from 0.684 (random) to 0.711 and 0.729 versus a 0.748 validation oracle, and both variants reach the baseline's 24-hour score in roughly 15 hours. The team reports new SOTA on WinoGrande (94.1% with Agentic RPM) and SVAMP (95.7% with inference-only).

MarkTechPostSep 6, 20:25 UTC · 9d agoAI research

Hack One Robot, Reach the Next: Unitree G1 Security Flaws

Researcher chained two Unitree G1 flaws (CVE-2026-76639, CVE-2026-76640) to gain unauthenticated root access, with compromised robots able to infect others via Bluetooth.

Security researcher Olivier Laflamme chained a path-traversal file-write flaw (CVE-2026-76639) in the G1 chatbot service and a Bluetooth server buffer overflow (CVE-2026-76640) to achieve unauthenticated root access remotely. The chain abuses Unitree's cloud API as a decryption oracle for the AES-128 key that unlocks BLE and WebRTC channels, and a compromised robot can propagate the attack to nearby G1 units. Unitree patched the cloud ownership-check flaw, paid a $5,000 bounty, and firmware fixes for BLE pairing and the buffer overflow remain harder to deliver.

Security AffairsAug 29, 11:55 UTC · 17d agoVulnerability CVE-2026-76639 CVE-2026-76640

U.S. CISA adds Gitea flaw to its Known Exploited Vulnerabilities catalog

CISA added the actively exploited Gitea RCE flaw CVE-2026-60004 (CVSS 9.8) to its KEV catalog, with federal patching due by August 28, 2026.

CISA added CVE-2026-60004, a critical remote code execution flaw in Gitea (CVSS 9.8), to its Known Exploited Vulnerabilities catalog. The flaw affects Gitea versions since 1.17, fixed in 1.27.1, and lets attackers with repository write access execute shell commands via the diffpatch API and Git hooks. Open registration allows unauthenticated attackers to create accounts and exploit exposed instances, with a reported attack deploying a cryptocurrency-miner-like payload.

Security AffairsAug 26, 08:44 UTC · 21d agoExploit / PoC in the wildCVE-2026-60004

Risky Bulletin: Russia starts blocking DoH and DoT

Russian users report blocks on DoH and DoT servers, including Cloudflare 1.1.1.1 and Google 8.8.8.8, in an apparent censorship crackdown.

Russian internet users began reporting failures connecting to DNS-over-HTTPS and DNS-over-TLS servers, suggesting a government crackdown on the two privacy protocols. The blocks reportedly cover Cloudflare's 1.1.1.1 and Google's 8.8.8.8 resolvers; Roskomnadzor has not officially confirmed the action. The agency tested a similar block in March on Beeline's network and had named DoH for blocking as early as 2021. The bulletin also briefly notes state-sponsored phishing of EU officials, a DDoS against Norway's Digdir, the ReliaQuest/ShinyHunters dispute, and older ransomware and breach disclosures.

Risky Business NewsAug 26, 03:07 UTC · 21d agoPolicy & legal1

What must happen for AI’s trillion-dollar gamble to pay off

Hyperscalers need 2.7x productivity gains by 2030 to justify nearly $1.1 trillion in AI data center spending, or risk bankruptcy and capital misallocation.

Wharton finance professor Jessica Wachter estimates hyperscaler AI expenditure will reach nearly $1.1 trillion through 2027 and that a 2.7x productivity increase is needed to break even by 2030. AI revenues of roughly $150-200 billion this year fall far short of about $750 billion in annual spending, with total investment from Alphabet, Microsoft, Amazon, Meta, and Oracle potentially exceeding $5 trillion over four years. Alphabet reported its first free cash flow deficit (about $5.9 billion) since its 2004 IPO due to AI infrastructure costs. Researchers warn that failed demand could make the buildout the largest capital misallocation in history, with depreciating GPU chips risking stranded assets.

MIT Technology Review · AISep 15, 10:00 UTC · 23h agoAI industry

Pick Your Poison: Learning to Select Poison Sets for Stronger LLM Backdoor Attacks

Poison set selection swings LLM backdoor attack success from 3% to 80%; SAILS boosts held-out success by 30 points.

The paper shows that random poison set selection severely underestimates worst-case backdoor vulnerability: across three LLaMA-3-8B settings with fixed model, clean data, and poison count, attack success ranges from 3% to 80% depending only on which poison set is chosen. The authors formalize poison selection as oracle-budgeted set optimization and introduce SAILS, which learns a set scorer from a few hundred finetune-and-evaluate runs, ranks millions of candidate sets, and audits a shortlist. SAILS improves held-out attack success by 30 percentage points over the strongest influence baselines, transfers from small-scale to full-scale finetuning, and extends to code-generation, agentic, and API-only backdoors.

Hugging Face daily papersupdated Sep 14, 04:44 UTC · 2d agofirst Sep 13, 20:00 UTC · 2d agoAI safety & security 2 sources1

E2A-Bench: Benchmarking Evidence-to-Action Reliability in Financial Chart Reasoning

E2A-Bench, a 969-query financial chart reasoning benchmark, finds VLMs fail evidence-to-action consistency, with fine-tuning amplifying BUY:SELL bias 4-6x.

E2A-Bench is a 969-query benchmark built from 323 HS300 constituents across three input modalities with deterministic OHLCV-derived evidence anchors, evaluating grounding, reasoning-action consistency, evidence-confidence calibration, and directional coverage via UCR, RCI, ECI, and NDR metrics. Testing 20 VLMs showed the lowest-hallucination model ranked near the bottom on coverage with only 6.4% directional coverage, and oracle-aided verification reduced unsupported claims but could collapse coverage. Financial fine-tuning amplified the BUY:SELL ratio by factors of 4.21 to 4.68 across base-fine-tuned pairs.

Hugging Face daily papersSep 12, 20:00 UTC · 3d agoAI research

OpenAI Launches the Agents API in Public Beta, Putting the Codex Harness Behind One API Call

OpenAI released its Agents API in public beta, exposing the managed Codex harness with hosted or self-hosted sandboxes, MCP tools, and subagents.

The Agents API is a managed service built on the open-source Codex harness, handling context compaction, tool search, programmatic tool calling, and multi-agent orchestration. Agents run in OpenAI-hosted sandboxes, self-hosted environments, or partner sandboxes from Blaxel, Cloudflare, Daytona, DigitalOcean, E2B, Modal, Oracle, Runloop, and Vercel. Data residency is US-only and Zero Data Retention is unsupported. Examples use model gpt-6-astra; vendor-reported results include SafetyKit cutting case review cost 60% and Ciridae achieving 4x lower subagent latency.

MarkTechPostupdated Sep 11, 08:11 UTC · 5d agofirst Sep 10, 21:21 UTC · 5d agoAI tools & infra 4 sources1

Now everyone can put data to work

OpenAI launched a Data agent in ChatGPT Work that connects to enterprise warehouses and builds shareable analysis dashboards without SQL.

OpenAI introduced a Data agent in ChatGPT Work that connects to approved sources including Snowflake, BigQuery, Databricks, Redshift, ClickHouse, MongoDB, and Datadog, plus files from Google Drive and SharePoint. It investigates metric changes, builds interactive dashboards, and integrates with BI tools such as Power BI, Tableau, Omni, Sigma, and ThoughtSpot using semantic layers from dbt, Databricks Genie Ontology, and Snowflake Horizon. Queries enforce the connected account's existing table, row, and column permissions, with administrators controlling access via Workspace settings. OpenAI says nearly all of its product team and over two-thirds of its GTM organization use it internally, and NTT Data, Thermo Fisher, and ServicePiston are Alpha customers.

OpenAI NewsSep 10, 15:00 UTC · 5d agoAI industry

IdeaAMBIG: Benchmarking Implementation-Critical Gaps in Research-Idea Specifications

IdeaAMBIG benchmark with 660 instances measures whether LLMs can spot and fix underspecified research-method details for faithful implementation.

Researchers introduce IdeaAMBIG, a benchmark of 660 evidence-grounded instances (163 real-world gaps from reproducibility reports and GitHub issues, 497 controlled synthetic gaps) built from papers, codebases, and reproduction artifacts. It evaluates codification-readiness assessment, defect localization, and clarification action generation. Across 13 LLMs, the best model achieved only 9.6% Macro Defect Recovery Rate on real-world instances but 80.6% clarification success when given the annotated defect. An oracle study showed gold resolutions raise the codification-ready rate from 14% to 98%, identifying defect localization as the main bottleneck.

arXiv cs.AI / cs.LG / cs.CLSep 9, 17:59 UTC · 6d agoAI research2

GPT-6 Astra: The next generation in intelligence for work

OpenAI launched GPT-6 Astra, its most capable and aligned model, in ChatGPT Work, Codex, and the API, claiming frontier performance and cybersecurity gains.

GPT-6 Astra is now available in ChatGPT Work, Codex, and the API, with OpenAI claiming state-of-the-art performance in computer use, browsing, professional work, software engineering, cybersecurity, and science. Astra is billed as the first model to reach the Critical cybersecurity capability threshold under OpenAI's Preparedness Framework, with 89% fewer unintended outcomes than GPT-5.6 Sol and 74.7% fewer than Claude Fable 5.1 on OpenAI's internal computer-use safety benchmark. Pricing starts at $10 per million input tokens and $50 per million output tokens, with claims of occupying most of the cost-efficiency frontier on Terminal Bench 4.0 and the Artificial Analysis Intelligence Index. New enterprise admin controls and plugins from Oracle Analytics, Power BI, Navan, and Avalara launched alongside the model.

OpenAI Newsupdated Sep 12, 13:10 UTC · 3d agofirst Sep 9, 11:00 UTC · 6d agoModel release 10 sources1

What Makes Adversarial Examples Transfer Across Deepfake Detectors?

A controlled study of 60 deepfake detectors shows adversarial example transfer depends heavily on source-target compatibility, with source averaging understating vulnerability.

The study evaluates adversarial example transferability across 60 deepfake detectors spanning six backbones, two pretraining regimes, and five training-data configurations, using AutoAttack (AA) and Carlini-Wagner with Expectation over Transformation (CW-EOT). Transfer rises sharply when source and target share an exact backbone, architecture family, pretraining regime, or training data, with the dominant factor depending on the attack. Mean attack success rate is 7.21% under AA and 19.52% under CW-EOT for single sources, while a multi-source oracle reaches 64.48% after excluding exact matches, showing source averaging can substantially understate target vulnerability. The authors release 240,000 adversarially perturbed images, pairwise transfer results, detector configurations, and evaluation code.

arXiv cs.CRSep 9, 10:28 UTC · 6d agoAI safety & security

Proximity Gaps for Gabidulin Codes and Applications

Researchers prove proximity-gap bounds for rank-metric and Gabidulin codes, enabling the first polynomial commitment scheme framework based on rank-metric error-correcting codes.

The paper proves every linear rank-metric code admits a proximity gap for deltas up to (d-1)/(3n) with error at most q^(e+1)/q^m, and improves the gap to (d-1)/(2n) for Gabidulin codes with error at most 10q^(n-1)/q^m, matching bounds for Reed-Solomon codes. A constructed infinite family of constant-rate Gabidulin codes shows the (d-1)/(2n) bound is tight, and a counterexample establishes a lower bound on the error at the d/(3n) gap. Applications include an IOPP for interleaved Gabidulin codes adapted from the Ligero IOPP and a q-linearized polynomial commitment scheme adapted from Ligero-based PCS, reportedly the first PCS framework based on rank-metric codes.

arXiv cs.CRSep 9, 07:45 UTC · 7d agoResearch

Microsoft Patches Record 974 Flaws, Including Two Exploited Windows Zero-Days

Microsoft's September Patch Tuesday fixed a record 974 flaws, including two Windows privilege-escalation zero-days actively exploited and added to CISA's KEV catalog.

Microsoft's September 2026 Patch Tuesday addressed a record 974 vulnerabilities (999 including 25 non-Microsoft CVEs), with over 110 rated critical; 723 affect Windows and 111 affect Office. Two Windows privilege-escalation zero-days are actively exploited: CVE-2026-85880, an ALPC heap-based buffer overflow, and CVE-2026-81963, an improper link resolution flaw in the Windows Update Stack, both allowing attackers to gain SYSTEM privileges. CISA added both flaws to its KEV catalog, giving federal civilian agencies until September 22, 2026 to apply fixes. Volexity, Proofpoint, MSTIC, and independent researchers were credited with the reports; notable additional fixes include network-reachable RCEs in Exchange, SharePoint, SQL Server, Remote Desktop Services, DNS, and DHCP.

The Hacker NewsSep 9, 04:41 UTC · 7d agoExploit / PoC in the wildCVE-2026-85880 CVE-2026-81963 CVE-2026-55007+9 CVEs

September 2026 Patch Tuesday roundup: Plugs for two zero day holes among almost 1,000 fixes in Windows

Microsoft's September 2026 Patch Tuesday ships 964 fixes including two exploited Windows zero-days (CVE-2026-85880, CVE-2026-81963) and a wormable DNS RCE.

Microsoft's September 2026 Patch Tuesday includes 964 Microsoft vulnerabilities requiring customer action, a record attributed to AI-assisted bug discovery, plus 174 third-party/open-source and 23 Chromium/Edge CVEs. Two zero-days are exploited in the wild: CVE-2026-85880, a Windows ALPC heap overflow enabling AppContainer sandbox escape and privilege escalation, and CVE-2026-81963, a Windows Update Stack escalation to SYSTEM. CVE-2026-69730, an unauthenticated Windows DNS RCE, is not yet exploited but Microsoft expects exploitation, and roughly 20 bugs could be wormable. Separately, SAP issued a critical CVSS 10.0 fix for the EPP component used in S/4HANA and NetWeaver.

CSO OnlineSep 9, 03:27 UTC · 7d agoVulnerability in the wildCVE-2026-85880 CVE-2026-81963 CVE-2026-69730+2 CVEs1

Microsoft Plugs Nearly 1,000 Security Holes

Microsoft's September Patch Tuesday fixes a record 974 flaws, including two actively exploited Windows zero-days and critical DNS and Windows Shell bugs.

Microsoft released fixes for 974 vulnerabilities, its largest-ever monthly patch batch, bringing the 2026 total above 2,600. Two zero-day privilege elevation flaws, CVE-2026-81963 and CVE-2026-85880, are being actively exploited. Critical bugs include CVE-2026-69730, an unauthenticated DNS weakness in Windows Server 2012 onward and Windows 10 deemed likely to be exploited, and CVE-2026-69829, a CVSS 9.8 Windows Shell remote code execution flaw requiring no privileges or user interaction. Microsoft and other vendors credit AI-assisted discovery for the growing patch volumes.

Krebs on SecuritySep 8, 21:44 UTC · 7d agoVulnerability in the wildCVE-2026-81963 CVE-2026-85880 CVE-2026-69730+1 CVEs

IdeaAMBIG: Benchmarking Implementation-Critical Gaps in Research-Idea Specifications

IdeaAMBIG benchmark of 660 specification-gap instances shows LLMs localize implementation-critical research gaps poorly, with best model at 9.6% defect recovery.

IdeaAMBIG is a benchmark of 660 evidence-grounded instances evaluating whether research-method specifications provide enough information for faithful implementation: 163 real-world gaps from reproducibility reports and GitHub issues plus 497 controlled synthetic gaps. It tests codification-readiness assessment, defect localization, and clarification action generation across 13 LLMs. The best model achieves only a 9.6% Macro Defect Recovery Rate on real-world instances, though 80.6% clarification success when given the annotated defect, and an oracle study shows gold resolutions raise codification-ready rates from 14% to 98%. Defect localization emerges as the main bottleneck across all evaluated models.

Hugging Face daily papersSep 8, 20:00 UTC · 7d agoAI research

ExecCritic: Learn to Test, Test to Improve for Coding Agents

ExecCritic separates test generation from patching for coding agents, lifting SWE-bench Verified resolution to 72.6%.

ExecCritic pairs a test-verify-revise scaffold with role-specific reinforcement learning: a Test agent writes repository-native tests and a Repair agent fixes code from execution feedback, both using Qwen-3.5-35B-A3B backbones. Post-trained Qwen agents compose to 72.6% on SWE-bench Verified, an 11.4-point gain over the 61.2% no-test baseline, without stronger-model or oracle feedback at evaluation time. The work shows test quality is the key variable: base-agent tests lowered resolution to 57.3% while GPT-5.6-sol tests raised it to 65.3%.

arXiv cs.AI / cs.LG / cs.CLSep 8, 17:53 UTC · 7d agoAI research1