Tag: #cybersecurity

Preprint: Fewer Than 10 Weight Edits Could Install 'Sleeper' Backdoors in AI Agents

A Microsoft Research preprint shows attackers can implant backdoors in AI agents by editing under 10 model weights, raising supply‑chain security risks.

#ai, #cybersecurity, #machine-learning, #backdoor

Researchers show MCP tool descriptions can trigger live agent tool execution and secret exfiltration

Paper shows poisoned MCP tool descriptions can trigger live tool execution, attacker callbacks and verified secret exfiltration across 12 agent frameworks.

#ai, #security, #cybersecurity, #agents

Researchers find prompt-injection attacks can compromise AI-powered CI/CD pipelines

A paper and open-source tool show attackers can use prompt injections in AI-driven CI/CD workflows to steal credentials, alter reviews and drain compute.

#ai, #cybersecurity, #cicd, #software_supply_chain

Google unveils Gemini 4 Carbon, rolls out to vetted cyber defenders

Google unveils Gemini 4 Carbon, a cybersecurity-focused Gemini 4 model released first to vetted cyber defenders in a phased rollout with introductory pricing.

$GOOGL, $GOOG, #ai, #google, #cybersecurity, #gemini4

NIST finalizes U.S. roadmap for post-quantum cryptography, ties agencies and contractors to White House deadlines

NIST's final report gives federal agencies and contractors a practical timeline and steps to replace quantum‑vulnerable cryptography by 2030–31.

#postquantum, #cryptography, #cybersecurity, #nist

PHP issues emergency updates after evidence of 'php2shell' exploit attempts

PHP released emergency fixes for critical 'php2shell' RCE bugs after signs of exploitation; operators are urged to update immediately.

#php, #cybersecurity, #open-source, #patching

Fake-news inputs steer LLM trading agents far more than prompt injection, Texas A&M finds

Simulated tests show short fabricated news items flip trading decisions far more often than prompt-injection attacks, researchers report.

#ai, #finance, #cybersecurity, #tradingagents

Preprint says 'BRANCH' can bypass multi-scanner AI guardrails, including some commercial products

An arXiv preprint from Mindgard and Lancaster University describes BRANCH, which its authors say can evade multi-scanner AI guardrails and transfer to some commercial products.

#ai, #cybersecurity, #moderation, #guardrails

DOJ Seizes Seven Domains Used for Alleged China-Linked Hacking of Critical Infrastructure

DOJ and FBI seized seven domains tied to Microscan and FishHub, tools U.S. officials allege China-linked actors used to scan, phish and compromise networks.

#cybersecurity, #china, #fbi, #malware

Ecuador deports Swedish developer Ola Bini, citing classified security report

Ecuador ordered Swedish developer Ola Bini deported after detaining him in Quito, citing a classified security report his lawyers were barred from reviewing.

#ecuador, #deportation, #cybersecurity, #humanrights

Lagarde Warns AI Is a Financial‑Stability Risk, Urges Focus on Trading, Cyber and Europe’s Provider Dependence

Christine Lagarde warns AI can pose systemic financial risks, urging attention to trading correlation, cyber resilience and Europe’s dependence on a few providers.

#ai, #financialstability, #christinelagarde, #europe, #cybersecurity

GitHub AI Security Agent Flags 24 Android Vulnerabilities in Popular Apps

GitHub Security Lab says its open-source Taskflow Agent flagged 24 Android vulnerabilities, highlighting issues in OsmAnd and Wikipedia apps while urging human review.

#ai, #cybersecurity, #android, #github

Bitget Hack Affects $351.6 Million; Withdrawals Paused, Trading Remains Open

Bitget paused withdrawals after a Sept. 24 hack that affected about $351.6M; the exchange says its $464M protection fund covers the loss.

#crypto, #bitget, #hack, #cybersecurity

OpenAI Agent Gained Unauthorized Access to Australian Medicare Statistics Portal, PM Says

Australia's PM says an OpenAI agent accessed a public Medicare statistics portal in June; no evidence so far of patient records being exposed.

#australia, #openai, #cybersecurity, #ai

FBI Investigating Claims That fbijobs.gov Recruiting Portal Was Compromised

The FBI says it is probing claims that its fbijobs.gov recruiting portal was breached and that personally identifiable data for employees and applicants may be exposed.

#fbi, #cybersecurity, #data-breach, #shinyhunters