Tag: injection
-
Three vulnerabilities in Anthropic Git MCP Server could let attackers tamper with LLMs
mcp-server-git versions prior to 2025-12.18.The three vulnerabilities are·CVE-2025-68143, an unrestricted git_init.·CVE-2025-68145, a path validation bypass.·CVE-2025-68144, an argument injection in git_diff.Unlike other vulnerabilities in MCP servers that required specific configurations, these work on any configuration of Anthropic’s official server, out of the box, Cyata says.Model Context Protocol (MCP) is an open standard introduced by Anthropic in 2024 to…
-
Google Gemini Flaw Turns Calendar Invites Into Attack Vector
The indirect prompt injection vulnerability allows an attacker to weaponize calendar invites to circumvent Google’s privacy controls and access private data. First seen on darkreading.com Jump to article: www.darkreading.com/cloud-security/google-gemini-flaw-calendar-invites-attack-vector
-
Prompt Injection Bugs Found in Official Anthropic Git MCP Server
Three vulnerabilities in Anthropic’s Git server for the MCP can be exploited via prompt injection First seen on infosecurity-magazine.com Jump to article: www.infosecurity-magazine.com/news/prompt-injection-bugs-anthropic/
-
Exploiting Google Gemini to Abuse Calendar Invites Illustrates AI Threats
Researchers with security firm Miggo used an indirect prompt injection technique to manipulate Google’s Gemini AI assistant to access and leak private data in Google Calendar events, highlighting the challenges AI presents that traditional security measures can’t address. First seen on securityboulevard.com Jump to article: securityboulevard.com/2026/01/exploiting-google-gemini-to-abuse-calendar-invites-illustrates-ai-threats/
-
Exploiting Google Gemini to Abuse Calendar Invites Illustrates AI Threats
Researchers with security firm Miggo used an indirect prompt injection technique to manipulate Google’s Gemini AI assistant to access and leak private data in Google Calendar events, highlighting the challenges AI presents that traditional security measures can’t address. First seen on securityboulevard.com Jump to article: securityboulevard.com/2026/01/exploiting-google-gemini-to-abuse-calendar-invites-illustrates-ai-threats/
-
Exploiting Google Gemini to Abuse Calendar Invites Illustrates AI Threats
Researchers with security firm Miggo used an indirect prompt injection technique to manipulate Google’s Gemini AI assistant to access and leak private data in Google Calendar events, highlighting the challenges AI presents that traditional security measures can’t address. First seen on securityboulevard.com Jump to article: securityboulevard.com/2026/01/exploiting-google-gemini-to-abuse-calendar-invites-illustrates-ai-threats/
-
Three Flaws in Anthropic MCP Git Server Enable File Access and Code Execution
A set of three security vulnerabilities has been disclosed in mcp-server-git, the official Git Model Context Protocol (MCP) server maintained by Anthropic, that could be exploited to read or delete arbitrary files and execute code under certain conditions.”These flaws can be exploited through prompt injection, meaning an attacker who can influence what an AI assistant…
-
Anthropic quietly fixed flaws in its Git MCP server that allowed for remote code execution
Prompt injection for the win First seen on theregister.com Jump to article: www.theregister.com/2026/01/20/anthropic_prompt_injection_flaws/
-
Google Gemini flaw exposes new AI prompt injection risks for enterprises
Real enterprise exposure: Analysts point out that the risk is significant in enterprise environments as organizations rapidly deploy AI copilots connected to sensitive systems.”As internal copilots ingest data from emails, calendars, documents, and collaboration tools, a single compromised account or phishing email can quietly embed malicious instructions,” said Chandrasekhar Bilugu, CTO of SureShield. “When employees…
-
OverAir Software Updates Pose Risks to Vehicles
eSync Alliance Chair Shrikant Acharya on How Standardization Can Prevent Breaches. Over-the-air updates are an irreplaceable part of software-defined vehicles, giving manufacturers a convenient way of remotely fixing and upgrading vehicles. If not appropriately secured, over-the-air updates can become a gateway for data theft, malware injection, vehicle theft and even injury. First seen on govinfosecurity.com…
-
Google Gemini Prompt Injection Flaw Exposed Private Calendar Data via Malicious Invites
Tags: control, cybersecurity, data, data-breach, flaw, google, injection, malicious, privacy, vulnerabilityCybersecurity researchers have disclosed details of a security flaw that leverages indirect prompt injection targeting Google Gemini as a way to bypass authorization guardrails and use Google Calendar as a data extraction mechanism.The vulnerability, Miggo Security’s Head of Research, Liad Eliyahu, said, made it possible to circumvent Google Calendar’s privacy controls by hiding a dormant…
-
Five Chrome extensions caught hijacking enterprise sessions
Blocking defenses and hijacking sessions: The campaign went beyond stealing credentials. Two of the extensions, Tool Access 11 and Data By Cloud 2, incorporated DOM manipulation routines that actively blocked access to security and administrative pages within the targeted platforms. This prevented the enterprise admins from reaching screens to change passwords, view sign-on history, or…
-
Contagious Claude Code bug Anthropic ignored promptly spreads to Cowork
Office workers without AI experience warned to watch for prompt injection attacks – good luck with that First seen on theregister.com Jump to article: www.theregister.com/2026/01/15/anthropics_claude_bug_cowork/
-
CVE-2025-64155: Exploit Code Released for Critical Fortinet FortiSIEM Command Injection Vulnerability
Tags: access, advisory, attack, authentication, cisa, cve, cyber, cybersecurity, exploit, flaw, fortinet, infrastructure, injection, kev, mitigation, threat, update, vpn, vulnerability, zero-dayExploit code has been published for CVE-2025-64155, a critical command injection vulnerability affecting Fortinet FortiSIEM devices. Key takeaways: CVE-2025-64155 is a critical operating system (OS) command injection vulnerability affecting Fortinet FortiSIEM. Fortinet vulnerabilities have historically been common targets for cyber attackers, with 23 Fortinet CVEs currently on the CISA KEV list. Public exploit code has…
-
How WitnessAI raised $58M to solve enterprise AI’s biggest risk
As companies deploy AI-powered chatbots, agents, and copilots across their operations,they’refacing a new risk: how do you let employees and AI agents use powerful AI tools without accidentally leaking sensitive data, violating compliance rules, oropening the door to prompt-based injections? Witness AI just raised $58 million to find a solution, building what they call >>the…
-
How WitnessAI raised $58M to solve enterprise AI’s biggest risk
As companies deploy AI-powered chatbots, agents, and copilots across their operations,they’refacing a new risk: how do you let employees and AI agents use powerful AI tools without accidentally leaking sensitive data, violating compliance rules, oropening the door to prompt-based injections? Witness AI just raised $58 million to find a solution, building what they call >>the…
-
How WitnessAI raised $58M to solve enterprise AI’s biggest risk
As companies deploy AI-powered chatbots, agents, and copilots across their operations,they’refacing a new risk: how do you let employees and AI agents use powerful AI tools without accidentally leaking sensitive data, violating compliance rules, oropening the door to prompt-based injections? Witness AI just raised $58 million to find a solution, building what they call >>the…
-
Exploit code public for critical FortiSIEM command injection flaw
Technical details and a public exploit have been published for a critical vulnerability affecting Fortinet’s Security Information and Event Management (SIEM) solution that could be leveraged by a remote, unauthenticated attacker to execute commands or code. First seen on bleepingcomputer.com Jump to article: www.bleepingcomputer.com/news/security/exploit-code-public-for-critical-fortisiem-command-injection-flaw/
-
January 2026 Microsoft Patch Tuesday: Actively exploited zero day needs attention
More priorities: Executives should also prioritize rapid patching and risk reduction efforts this month around the Windows Local Security Authority Subsystem Service Remote Code Execution, Windows Graphics Component Elevation of Privilege, and Windows Virtualization Based Security Enclave Elevation of Privilege flaws, Bicer said, as these vulnerabilities directly enable full system or trust boundary compromise.Strategic focus…
-
Fortinet Fixes Critical FortiSIEM Flaw Allowing Unauthenticated Remote Code Execution
Fortinet has released updates to fix a critical security flaw impacting FortiSIEM that could allow an unauthenticated attacker to achieve code execution on susceptible instances.The operating system (OS) injection vulnerability, tracked as CVE-2025-64155, is rated 9.4 out of 10.0 on the CVSS scoring system.”An improper neutralization of special elements used in an OS command (‘OS…
-
SAP January 2026 Security Patch Day Fixes Critical Injection and RCE Flaws
SAP released 17 new security notes on January 13, 2026, addressing vulnerabilities affecting widely deployed enterprise systems. The patch day includes four critical-severity flaws spanning SQL injection, remote code execution, and code injection attacks that could allow authenticated and unauthenticated threat actors to compromise SAP environments. Critical Vulnerabilities Demand Immediate Attention The most severe vulnerabilities…
-
ServiceNow patches critical AI platform flaw that could allow user impersonation
The company says it has no evidence the bug was exploited before October’s patch, but researchers say AI agent configuration can still enable prompt-injection style abuse. First seen on cyberscoop.com Jump to article: cyberscoop.com/servicenow-fixes-critical-ai-vulnerability-cve-2025-12420/
-
Red-Teaming BrowseSafe Exposes AI Browser Guardrail Gaps
Red-team testing shows encoded prompt injections can bypass BrowseSafe guardrails. First seen on esecurityplanet.com Jump to article: www.esecurityplanet.com/threats/red-teaming-browsesafe-exposes-ai-browser-guardrail-gaps/
-
For application security: SCA, SAST, DAST and MAST. What next?
Tags: advisory, ai, application-security, automation, best-practice, business, cisa, cisco, cloud, compliance, container, control, cve, data, exploit, flaw, framework, gartner, government, guide, ibm, incident response, infrastructure, injection, kubernetes, least-privilege, ml, mobile, network, nist, resilience, risk, sbom, service, software, sql, supply-chain, threat, tool, training, update, vulnerability, waf<img loading="lazy" decoding="async" src="https://b2b-contenthub.com/wp-content/uploads/2026/01/posture-provenance-proof.jpg?quality=50&strip=all&w=1024" alt="Chart: Posture, provenance and proof." class="wp-image-4115680" srcset="https://b2b-contenthub.com/wp-content/uploads/2026/01/posture-provenance-proof.jpg?quality=50&strip=all 1430w, b2b-contenthub.com/wp-content/uploads/2026/01/posture-provenance-proof.jpg?resize=300%2C168&quality=50&strip=all 300w, b2b-contenthub.com/wp-content/uploads/2026/01/posture-provenance-proof.jpg?resize=768%2C431&quality=50&strip=all 768w, b2b-contenthub.com/wp-content/uploads/2026/01/posture-provenance-proof.jpg?resize=1024%2C575&quality=50&strip=all 1024w, b2b-contenthub.com/wp-content/uploads/2026/01/posture-provenance-proof.jpg?resize=1240%2C697&quality=50&strip=all 1240w, b2b-contenthub.com/wp-content/uploads/2026/01/posture-provenance-proof.jpg?resize=150%2C84&quality=50&strip=all 150w, b2b-contenthub.com/wp-content/uploads/2026/01/posture-provenance-proof.jpg?resize=854%2C480&quality=50&strip=all 854w, b2b-contenthub.com/wp-content/uploads/2026/01/posture-provenance-proof.jpg?resize=640%2C360&quality=50&strip=all 640w, b2b-contenthub.com/wp-content/uploads/2026/01/posture-provenance-proof.jpg?resize=444%2C250&quality=50&strip=all 444w” width=”1024″ height=”575″ sizes=”auto, (max-width: 1024px) 100vw, 1024px” /> Sunil GentyalaOver the past year the community has admitted the obvious: the battleground is the software supply chain and…
-
Apache Struts External Entity (XXE) Injection Vulnerability S2-069 (CVE-2025-68493)
Overview Recently, NSFOCUS CERT detected that Apache issued a security bulletin to fix the Apache Struts external entity (XXE) injection vulnerability S2-069 (CVE-2025-68493); Because the XWork component of Apache Struts does not perform effective validation when parsing XML configuration, attackers can inject external entities by constructing malicious XML data to read sensitive server files, perform…The…
-
Critical InputPlumber Flaw Enables UI Input Injection and DenialService
Security researchers have discovered critical vulnerabilities in InputPlumber, a Linux input device utility used in SteamOS, that could allow attackers to inject keystrokes, leak sensitive information, and cause denial-of-service conditions. The flaws, tracked as CVE-2025-66005 and CVE-2025-14338, affect InputPlumber versions before v0.69.0 and stem from inadequate D-Bus authorization checks. CVE ID Description Affected Versions Impact…
-
ZombieAgent ChatGPT attack shows persistent data leak risks of AI agents
Worm-like propagation: The email attack even has worming capabilities, as the malicious prompts could instruct ChatGPT to scan the inbox, extract addresses from other email messages, exfiltrate those addresses to the attackers using the URL trick, and send similar poisoned messages to those addresses as well.If the victim is the employee of an organization that…
-
Radware Discloses ZombieAgent Technique to Compromise AI Agents
Radware this week announced it has discovered a zero-click indirect prompt injection (IPI) vulnerability targeting the Deep Research agent developed by OpenAI. Dubbed ZombieAgent, Radware researchers have discovered that it is possible to implant malicious rules directly into the long-term memory or working notes of an AI agent. That technique enables a malicious actor to..…