Bleeping Computer
PulseAugur coverage of Bleeping Computer — every cluster mentioning Bleeping Computer across labs, papers, and developer communities, ranked by signal.
8 day(s) with sentiment data
-
AI tool Artex exploited in separate cyberattacks
An AI tool named Artex has been exploited by malicious actors in two separate security incidents. In one case, attackers used Artex to breach WorkOS's internal sales application, leading to a data breach. In another ins…
-
Coding agents vulnerable to malicious instruction files in Downloads folder
Security researchers have identified a vulnerability where malicious Markdown files, such as AGENTS.md or CLAUDE.md, can be silently downloaded into a user's Downloads folder. When coding agents like Codex or Claude Cod…
-
Northflank breach exposes customer AI API keys
A security breach at Northflank has resulted in the compromise of customer AI API keys. The incident, reported by Bleeping Computer, highlights the risks associated with storing sensitive credentials. The extent of the …
-
Malicious markdown files hijack AI coding agents like Codex and Claude
A security vulnerability has been discovered where malicious markdown files, such as AGENTS.md or CLAUDE.md, can be placed in an AI coding agent's working directory to hijack its operations. These files trick agents lik…
-
AI agents vulnerable to email-based prompt injection attacks
A recent experiment demonstrated how AI agents can be manipulated through carefully crafted emails, leading to unauthorized command execution and data exfiltration without exploiting software vulnerabilities. This indir…
-
Coding agents vulnerable to malicious Markdown files, research shows
A security research firm, Forever Security, has uncovered a vulnerability in coding agents where malicious instructions embedded in Markdown files can be executed. These files, often saved in user download folders, tric…
-
AI coding agents hijacked via malicious AGENTS(dot)md file
A security vulnerability has been discovered that affects AI coding agents, allowing them to be hijacked. The attack exploits a file named AGENTS(dot)md, which, when placed in a user's Downloads folder, can trick these …
-
LiteLLM release pipeline compromised to distribute info-stealer malware
The release pipeline for LiteLLM, an open-source tool for simplifying LLM API calls, was compromised. Attackers injected a malicious info-stealer into the release process, which was then distributed through Docker and P…
-
AI agents vulnerable to malicious Markdown files; System 1 models show mixed results in medicine
A new study published on Hugging Face evaluates the performance of System 1 models in medical contexts, finding they excel at knowledge recall and text understanding but struggle with image interpretation and clinical j…
-
Microsoft Teams to integrate third-party deepfake detection tools
Microsoft Teams is integrating support for third-party deepfake detection tools. This feature aims to enhance security and privacy by helping users identify and protect against impersonation and malicious deepfakes with…
-
Fake AI Sites Steal Ad Accounts and MFA Codes
Cybercriminals are using fake websites mimicking popular AI tools like ChatGPT and Gemini to steal advertising accounts and multi-factor authentication (MFA) codes. These phishing campaigns specifically target agencies …
-
DIVD: AI-driven attacks exploit Zammad zero-days
The cybersecurity group DIVD has reported that zero-day vulnerabilities in the Zammad customer support platform were exploited to conduct an AI-driven network breach. Attackers leveraged these vulnerabilities to gain un…
-
Anthropic's Claude Opus 5.5 adopts simpler writing style, Arena.ai finds
Anthropic's Claude Opus 5.5 has been analyzed by Arena.ai, revealing significant changes in its writing style compared to its predecessor, Opus 5. The analysis found that Opus 5.5 uses em dashes 95% less frequently and …
-
Man sentenced to 15 years for AI deepfake extortion
A man has been sentenced to 15 years in prison for using AI-generated deepfake pornography to extort women. The perpetrator created non-consensual explicit images of his victims and then used these images to demand mone…
-
AI-powered attacks escalate credential theft, targeting 395 organizations
Hackers are leveraging AI to develop sophisticated frameworks for widespread credential theft. One such AI-powered attack exploited vulnerabilities in PaperCut software, compromising 395 organizations. These advancement…
-
Critical LangFlow flaw exploited to steal OpenAI and AWS keys
A critical vulnerability in the LangFlow platform has been actively exploited, allowing attackers to steal sensitive API keys for OpenAI and AWS. The flaw enables unauthorized access to credentials, potentially leading …
-
Anthropic cuts Claude Code weekly limits by 17% despite 'permanent increase' framing · 4 sources tracked
Anthropic is reducing the weekly usage limits for its Claude Code service by 17%. While the company frames this as a permanent 25% increase over the original baseline for its Pro, Max, Team, and Enterprise plans, users …
-
Microsoft Blames RGB Software for Windows 11 Game Crashes
Microsoft has identified RGB lighting software as the cause of recent Windows 11 instability, including game crashes and blue screens. The issue stems from RGB software drivers interfering with core system files, partic…
-
Pokemon Center and Steam customers impacted by CEVA Logistics data breach
A third-party data breach at logistics provider CEVA Logistics has exposed customer information from the Pokémon Center and Valve Corporation. The breach, which occurred between July 29 and August 1, 2026, involved cust…
-
Apple warns 110 countries of mercenary spyware attacks on iPhones
Apple has issued a spyware warning to iPhone users across 110 countries, alerting them to potential mercenary spyware attacks. These targeted attacks, often costing millions and utilizing zero-click exploits, are design…