Tag: injection
-
Salesforce Patches CRM Data Exfiltration Vulnerability
Agentforce Agentic AI Tool Was Exposed to Indirect Prompt Injection Attacks. Salesforce has patched a vulnerability involving its Agentforce agentic artificial intelligence tool, discovered by researchers, that attackers could have exploited using an indirect prompt injection attack to steal sensitive customer data and leads being stored in the CRM system. First seen on govinfosecurity.com Jump…
-
Prompt injection and a $5 domain trick Salesforce Agentforce into leaking sales
More fun with AI agents and their security holes First seen on theregister.com Jump to article: www.theregister.com/2025/09/26/salesforce_agentforce_forceleak_attack/
-
Prompt injection and a $5 domain trick Salesforce Agentforce into leaking sales
More fun with AI agents and their security holes First seen on theregister.com Jump to article: www.theregister.com/2025/09/26/salesforce_agentforce_forceleak_attack/
-
Salesforce AI Agent Vulnerability Lets Attackers Steal Sensitive Data
Cybersecurity researchers at Noma Labs have discovered a critical vulnerability in Salesforce’s Agentforce AI platform that could allow attackers to steal sensitive customer data through sophisticated prompt injection techniques. The vulnerability, dubbed >>ForcedLeak,
-
Salesforce AI Agents Forced to Leak Sensitive Data
Yet again researchers have uncovered an opportunity (dubbed ForcedLeak for indirect prompt injection against autonomous agents lacking sufficient security controls, but this time the risk involves PII, corporate secrets, physical location data, and so much more. First seen on darkreading.com Jump to article: www.darkreading.com/vulnerabilities-threats/salesforce-ai-agents-leak-sensitive-data
-
Salesforce Patches Critical ForcedLeak Bug Exposing CRM Data via AI Prompt Injection
Cybersecurity researchers have disclosed a critical flaw impacting Salesforce Agentforce, a platform for building artificial intelligence (AI) agents, that could allow attackers to potentially exfiltrate sensitive data from its customer relationship management (CRM) tool by means of an indirect prompt injection.The vulnerability has been codenamed ForcedLeak (CVSS score: 9.4) by Noma Security, First seen on…
-
Critical Vulnerability in Salesforce AgentForce Exposed
Critical flaw ForcedLeak in Salesforce’s AgentForce allows CRM data theft via prompt injection First seen on infosecurity-magazine.com Jump to article: www.infosecurity-magazine.com/news/critical-flaw-salesforce-agentforce/
-
Vulnerability in Salesforce AI could be tricked into leaking CRM data
Tags: access, ai, api, attack, cybersecurity, data, exploit, injection, Internet, software, supply-chain, tool, update, vulnerabilityGuardrails, not just patches: While Salesforce responded quickly with a patch, experts agree that AI agents represent a fundamentally broader attack surface. These systems combine memory, decision-making, and tool execution, meaning compromises can spread quickly and, as Bennett puts it, “at machine speed.” “It’s advisable to secure the systems around the AI agents in use, which…
-
AI coding assistants amplify deeper cybersecurity risks
Tags: access, ai, api, application-security, attack, authentication, business, ceo, ciso, cloud, compliance, control, cybersecurity, data, data-breach, detection, fintech, flaw, governance, injection, leak, LLM, metric, open-source, programming, radius, risk, risk-management, service, software, startup, strategy, threat, tool, training, vulnerability‘Shadow’ engineers and vibe coding compound risks: Ashwin Mithra, global head of information security at continuous software development firm Cloudbees, notes that part of the problem is that non-technical teams are using AI to build apps, scripts, and dashboards.”These shadow engineers don’t realize they’re part of the software development life cycle, and often bypass critical…
-
State-Sponsored Hackers Exploiting Libraesva Email Security Gateway Vulnerability
Libraesva has released a security update to address a vulnerability in its Email Security Gateway (ESG) solution that it said has been exploited by state-sponsored threat actors.The vulnerability, tracked as CVE-2025-59689, carries a CVSS score of 6.1, indicating medium severity.”Libraesva ESG is affected by a command injection flaw that can be triggered by a malicious…
-
Lectora Desktop and Online XSS Vulnerability Enables JavaScript Injection
A critical cross-site scripting (XSS) vulnerability affecting both Lectora Desktop and Lectora Online has been disclosed, enabling attackers to inject JavaScript through crafted URL parameters. Discovered by security researcher Mohammad Jassim and documented by the CERT® Coordination Center on September 22, 2025, this flaw poses a risk of client-side code execution, session hijacking, and user…
-
Libraesva ESG Vulnerability Allows Attackers to Execute Malicious Commands
A critical command injection vulnerability in Libraesva ESG email security gateways has been discovered, allowing attackers to execute arbitrary commands through specially crafted compressed email attachments. The vulnerability, designated CVE-2025-59689, affects versions starting from 4.5 and has already been exploited by what appears to be a foreign state actor. Diagram showing how command injection attacks…
-
ShadowLeak Exploit Exposed Gmail Data Through ChatGPT Agent
Radware researchers revealed a service-side flaw in OpenAI’s ChatGPT. The ShadowLeak attack had used indirect prompt injection to bypass defences and leak sensitive data, but the issue has since been fixed. First seen on hackread.com Jump to article: hackread.com/shadowleak-exploit-exposed-gmail-data-chatgpt-agent/
-
Patch Now: Max-Severity Fortra GoAnywhere Bug Allows Command Injection
Exploitation of the flaw, tracked as CVE-2025-10035, is highly dependent on whether systems are exposed to the Internet, according to Fortra. First seen on darkreading.com Jump to article: www.darkreading.com/cyberattacks-data-breaches/patch-fortra-goanywhere-bug-command-injection
-
Fortra warns of max severity flaw in GoAnywhere MFT’s License Servlet
Fortra has released security updates to patch a maximum severity vulnerability in GoAnywhere MFT’s License Servlet that can be exploited in command injection attacks. First seen on bleepingcomputer.com Jump to article: www.bleepingcomputer.com/news/security/fortra-warns-of-max-severity-flaw-in-goanywhere-mfts-license-servlet/
-
New iOS Video Injection Tool Bypasses Biometric Locks on Jailbroken iPhones
A newly discovered video injection tool for iOS devices that have been jailbroken poses a serious threat to modern digital identity verification. Developed to run on iOS 15 or later, this highly specialized toolkit can circumvent weak biometric checks and even exploit services lacking any biometric safeguards. Its emergence marks a troubling shift toward automated,…
-
CrowdStrike Buys Pangea for $260M to Guard Enterprise AI Use
Silicon Valley Startup Brings AI Agent and Prompt Injection Protections to Falcon. CrowdStrike plans to purchase Pangea to add native AI detection and response capabilities to its Falcon platform. The company says the acquisition will help secure AI models and users alike from preventing prompt injection to tracking agent activity across enterprise environments. First seen…
-
Meet ShadowLeak: ‘Impossible to detect’ data theft using AI
Tags: ai, attack, business, ciso, cybersecurity, data, data-breach, email, exploit, gartner, governance, injection, LLM, malicious, RedTeam, resilience, risk, sans, service, sql, supply-chain, technology, theft, tool, update, vulnerabilityWhat CSOs should do: To blunt this kind of attack, he said CSOs should:treat AI agents as privileged actors: apply the same governance used for a human with internal resource access;separate ‘read’ from ‘act’ scopes and service accounts, and where possible sanitize inputs before LLM (large language model) ingestion. Strip/neutralize hidden HTML, flatten to safe…
-
New attack on ChatGPT research agent pilfers secrets from Gmail inboxes
Unlike most prompt injections, ShadowLeak executes on OpenAI’s cloud-based infrastructure. First seen on arstechnica.com Jump to article: arstechnica.com/information-technology/2025/09/new-attack-on-chatgpt-research-agent-pilfers-secrets-from-gmail-inboxes/
-
SQL Injection
SQL Injection (SQLi) remains one of the most persistent and dangerous web application vulnerabilities. Attackers who successfully exploit SQLi can read, modify, or delete sensitive data, bypass authentication, escalate privileges, and in some cases take full control of backend systems. This page explains SQL Injection in plain language, shows why it matters, covers types and…
-
Check Point acquires Lakera to build a unified AI security stack
Tags: access, ai, api, attack, automation, cloud, compliance, control, cybersecurity, data, endpoint, government, infrastructure, injection, LLM, network, RedTeam, risk, saas, startup, supply-chain, tool, trainingClosing a critical gap: Experts call this acquisition significant and not merely adding just another tool to the stack. “This acquisition closes a real gap by adding AI-native runtime guardrails and continuous red teaming into Check Point’s stack,” said Amit Jaju, senior managing director at Ankura Consulting. “Customers can now secure LLMs and agents alongside…
-
New Magecart Attack Injects Malicious JavaScript to Steal Payment Data
A new Magecart-style campaign has emerged that leverages malicious JavaScript injections to skim payment data from online checkout forms. The threat surfaced after security researcher sdcyberresearch posted a cryptic tweet hinting at an active campaign hosted on cc-analytics[.]com. Subsequent analysis revealed a heavily obfuscated script that hooks into checkout fields, collects credit card and billing…
-
Chaos-Mesh flaws put Kubernetes clusters at risk of full takeover
Tags: access, api, authentication, cloud, control, data-breach, exploit, flaw, infrastructure, injection, kubernetes, network, risk, service, tool, vulnerabilitychaosctl tool and port. Some cloud infrastructure providers that offer Chaos-Mesh implementations as part of their managed Kubernetes Services, such as Azure Chaos Studio, are also impacted. Chaos-Mesh was designed to orchestrate fault scenarios that could impact infrastructure and applications. The researchers observed that one core component of Chaos-Mesh, the Controller Manager, exposed a GraphQL…
-
Chaos Mesh Critical GraphQL Flaws Enable RCE and Full Kubernetes Cluster Takeover
Tags: access, cybersecurity, exploit, flaw, injection, kubernetes, network, rce, remote-code-execution, vulnerabilityCybersecurity researchers have disclosed multiple critical security vulnerabilities in Chaos Mesh that, if successfully exploited, could lead to cluster takeover in Kubernetes environments.”Attackers need only minimal in-cluster network access to exploit these vulnerabilities, execute the platform’s fault injections (such as shutting down pods or disrupting network communications), and perform First seen on thehackernews.com Jump to…
-
Threat Actors and Code Assistants: The Hidden Risks of Backdoor Injections
AI code assistants integrated into IDEs, like GitHub Copilot, offer powerful chat, auto-completion, and test-generation features. However, threat actors and careless users can exploit these capabilities to inject backdoors, leak sensitive data, and produce harmful code. Indirect prompt injection attacks exploit context-attachment features by contaminating public data sources with hidden instructions. When unsuspecting developers feed…
-
LangChainGo Vulnerability Allows Malicious Prompt Injection to Access Sensitive Data
A recently discovered flaw in LangChainGo, the Go implementation of the LangChain framework for large language models, permits attackers to read arbitrary files on a server by injecting malicious prompt templates. Tracked as CVE-2025-9556, this vulnerability arises from the use of the Gonja template engine, which supports Jinja2 syntax and can be manipulated to perform…