Tag: LLM
-
AI Has Created a New Attack Surface and Encryption Is Not Enough
Tags: access, ai, attack, ciso, cryptography, cybersecurity, data, data-breach, defense, encryption, endpoint, exploit, framework, injection, intelligence, leak, LLM, microsoft, mitigation, network, risk, service, side-channel, sql, training, vulnerability<div cla Executive Insight For decades, enterprises relied on strong encryption to protect sensitive data in transit, and encryption used to be the end of the conversation. If an organization could say “we use TLS 1.3 and modern cipher suites,” that was enough to reassure boards, regulators, and customers that data in transit was…
-
Try our new dimensional analysis Claude plugin
We’re releasing a new Claude plugin for developing and auditing code that implements dimensional analysis, a technique we explored in our most recent blog post. Most LLM-based security skills ask the model to find bugs. Our new dimensional-analysis plugin for Claude Code takes a different approach: it uses the LLM to annotate your codebase with…
-
Malicious LiteLLM versions linked to TeamPCP supply chain attack
TeamPCP backdoored LiteLLM v1.82.71.82.8, likely via Trivy CI/CD, adding tools to steal credentials, move in Kubernetes, and keep persistent access. Threat actor TeamPCP compromised LiteLLM versions 1.82.7 and 1.82.8, likely through a Trivy CI/CD breach. LiteLLM, with over 95 million monthly downloads, helps developers route LLM requests via a single API. The malicious releases, now…
-
Malicious LiteLLM versions linked to TeamPCP supply chain attack
TeamPCP backdoored LiteLLM v1.82.71.82.8, likely via Trivy CI/CD, adding tools to steal credentials, move in Kubernetes, and keep persistent access. Threat actor TeamPCP compromised LiteLLM versions 1.82.7 and 1.82.8, likely through a Trivy CI/CD breach. LiteLLM, with over 95 million monthly downloads, helps developers route LLM requests via a single API. The malicious releases, now…
-
Training an AI agent to attack LLM applications like a real adversary
Most enterprise software development teams now ship AI-powered applications faster than traditional penetration testing can keep up with. A security team with 500 applications … First seen on helpnetsecurity.com Jump to article: www.helpnetsecurity.com/2026/03/25/novee-ai-pentesting-agent/
-
Compromised LiteLLM Package With 95M Downloads Tied to TeamPCP, After Trivy KICS Hacks
Security researchers discovered that the popular Python library litellm was compromised on PyPI. With over 95 million monthly downloads, this open-source tool helps developers route requests across various LLM providers through a single API. The threat actor, identified as TeamPCP, injected malicious code into versions 1.82.7 and 1.82.8. This devastating supply chain attack directly follows the group’s…
-
Compromised LiteLLM Package With 95M Downloads Tied to TeamPCP, After Trivy KICS Hacks
Security researchers discovered that the popular Python library litellm was compromised on PyPI. With over 95 million monthly downloads, this open-source tool helps developers route requests across various LLM providers through a single API. The threat actor, identified as TeamPCP, injected malicious code into versions 1.82.7 and 1.82.8. This devastating supply chain attack directly follows the group’s…
-
Julius v0.2.0: From 33 to 63 Probes, Now Detecting Cloud AI, Enterprise Inference, and RAG Pipelines
TL;DR: Julius v0.2.0 nearly doubles LLM fingerprinting probe coverage from 33 to 63, adding detection for cloud-managed AI services (AWS Bedrock, Azure OpenAI, Vertex AI), high-performance inference servers (SGLang, TensorRT-LLM, Triton), AI gateways (Portkey, Helicone, Bifrost), and self-hosted RAG platforms (PrivateGPT, RAGFlow, Quivr). This release also hardens the scanner itself with response size limiting and……
-
Training effective models without the annotation budget
Learn how to bypass costly annotation workflows by using LLM-generated labels and lightweight fine-tuning to build high-quality NER models with minimal human input. First seen on securityboulevard.com Jump to article: securityboulevard.com/2026/03/training-effective-models-without-the-annotation-budget/
-
LiteLLM loses game of Trivy pursuit, gets compromised
Python interface for LLMs infected with malware via polluted CI/CD pipeline First seen on theregister.com Jump to article: www.theregister.com/2026/03/24/trivy_compromise_litellm/
-
Why CISOs should embrace AI honeypots
Tags: access, ai, api, attack, breach, business, ciso, credentials, cyberattack, cybercrime, cybersecurity, data, defense, detection, exploit, hacker, LLM, mitigation, open-source, RedTeam, risk, service, threat, tool, vulnerabilityWhy CISOs should consider honeypots: Another player in the AI honeypot space is Deutsche Telekom (DT). The firm is both a user and purveyor of AI-powered honeypots through its free, open-source platform ‘T-Pot.’ The most obvious advantage to their use, explains Marco Ochse, DT’s lead for threat analytics and mitigation, lies in how little these…
-
The OWASP Top 10 for LLM Applications (2025): Explained Simply
6 min readThe OWASP Top 10 for LLM Applications is the most widely referenced framework for understanding these risks. First released in 2023, OWASP updated the list in late 2024 to reflect real-world incidents, emerging attack techniques and the rapid growth of agentic AI. First seen on securityboulevard.com Jump to article: securityboulevard.com/2026/03/the-owasp-top-10-for-llm-applications-2025-explained-simply/
-
Llamafile, Mozilla’s portable LLM runner, gets GPU support and a rebuilt core
Running a large language model on a single machine without cloud access or a container runtime remains a priority for practitioners working in air-gapped or … First seen on helpnetsecurity.com Jump to article: www.helpnetsecurity.com/2026/03/20/llamafile-0-10-0-released/
-
AI Conundrum: Why MCP Security Can’t Be Patched Away
MCP introduces security risks into LLM environments that are architectural and not easily fixable, researcher says at RSAC 2026 Conference. First seen on darkreading.com Jump to article: www.darkreading.com/application-security/mcp-security-patched
-
AI Conundrum: Why MCP Security Can’t Be Patched Away
MCP introduces security risks into LLM environments that are architectural and not easily fixable, researcher says at RSAC 2026 Conference. First seen on darkreading.com Jump to article: www.darkreading.com/application-security/mcp-security-patched
-
AI Conundrum: Why MCP Security Can’t Be Patched Away
MCP introduces security risks into LLM environments that are architectural and not easily fixable, researcher says at RSAC 2026 Conference. First seen on darkreading.com Jump to article: www.darkreading.com/application-security/mcp-security-patched
-
Salt Security Launches Agentic Security Platform for the AI Stack Across LLMs, MCP Servers and APIs
This week, Salt Security has announced the launch of the Salt Agentic Security Platform, a new platform designed to secure the rapidly expanding Agentic Security Graph inside modern enterprises and enable organisations to adopt AI agents safely and at scale. As enterprises deploy AI agents to drive greater efficiency and productivity, their success depends on…
-
5 key priorities for your RSAC 2026 agenda
Tags: ai, api, attack, automation, ciso, compliance, conference, cybersecurity, data, detection, framework, governance, identity, infrastructure, injection, LLM, risk, service, soc, threat, tool, trainingEnable AI adoption fast enough to stay competitive.Secure the enterprise against a threat landscape that AI itself is creating.These are not sequential problems, unfortunately; they are parallel ones. I’d argue that RSAC 2026 is your best opportunity this year as a security leader to close the knowledge gap. AI prioritised Learning Framework: RSAC can be…
-
Everyone Is Deploying AI Agents. Almost Nobody Knows What They’re Doing.
Tags: access, ai, api, attack, ceo, ciso, credentials, data, data-breach, finance, infrastructure, Internet, LLM, risk, service, tool, vulnerability, wafOne constant I hear from CISOs I speak with is that AI agents are not coming. They are already inside organizations, reasoning through goals, selecting tools, and taking action through the same APIs that connect your most sensitive systems. And most security teams have no idea what those agents are doing. The problem Is not…
-
Everyone Is Deploying AI Agents. Almost Nobody Knows What They’re Doing.
Tags: access, ai, api, attack, ceo, ciso, credentials, data, data-breach, finance, infrastructure, Internet, LLM, risk, service, tool, vulnerability, wafOne constant I hear from CISOs I speak with is that AI agents are not coming. They are already inside organizations, reasoning through goals, selecting tools, and taking action through the same APIs that connect your most sensitive systems. And most security teams have no idea what those agents are doing. The problem Is not…
-
CISOs rethink their data protection strategies
Tags: access, ai, attack, automation, breach, business, cisco, ciso, cloud, compliance, computing, control, cyber, data, defense, framework, governance, healthcare, identity, jobs, LLM, privacy, resilience, risk, service, strategy, technology, tool, zero-trustFactors driving strategy evaluations CISOs, security experts, and data practitioners cite the expanding use of AI in the enterprise as the main reason they’re rethinking their data protection strategies.”AI is exposing more sensitive information as [workers] are taking that information and typing it into LLMs,” says Errol Weiss, CSO at Health-ISAC.AI tools make it easy…
-
prompted: Key Insights from the AI Security Practitioners Conference FireTail Blog
Tags: ai, api, application-security, attack, automation, conference, cybersecurity, data, defense, detection, exploit, google, infrastructure, injection, LLM, malicious, malware, monitoring, openai, risk, strategy, theft, threat, tool, training, update, vulnerability, zero-dayMar 17, 2026 – Jeremy Snyder – The State of AI Security: Moving Beyond TheoryThe biggest shift evident at the [un]prompted AI Security Practitioners Conference was the move from purely theoretical discussions about “what could go wrong” to concrete, battle-tested methodologies for “what is going wrong and how we fix it.” It’s clear that AI…
-
prompted: Key Insights from the AI Security Practitioners Conference FireTail Blog
Tags: ai, api, application-security, attack, automation, conference, cybersecurity, data, defense, detection, exploit, google, infrastructure, injection, LLM, malicious, malware, monitoring, openai, risk, strategy, theft, threat, tool, training, update, vulnerability, zero-dayMar 17, 2026 – Jeremy Snyder – The State of AI Security: Moving Beyond TheoryThe biggest shift evident at the [un]prompted AI Security Practitioners Conference was the move from purely theoretical discussions about “what could go wrong” to concrete, battle-tested methodologies for “what is going wrong and how we fix it.” It’s clear that AI…
-
Anton’s Vibe Coding Experience: A Reflection on Risk Decisions
Tags: access, ai, application-security, authentication, business, compliance, corporate, credentials, data, google, linkedin, LLM, risk, toolLook, I’m not a developer, and the last time I truly “wrote code” was probably a good number of years ago (and it was probably Perl so you may hate me). I am also not an appsec expert (as I often remind people). Below I am describing my experience “vibe coding” an application. Before I go…
-
Heading to RSA Conference 2026? Mark your Calendar and Meet Thales!
Tags: access, ai, application-security, attack, communications, compliance, conference, container, control, cybersecurity, data, defense, firewall, framework, GDPR, google, HIPAA, iam, ibm, injection, LLM, malicious, risk, tool, vulnerabilityHeading to RSA Conference 2026? Mark your Calendar and Meet Thales! madhav Tue, 03/17/2026 – 05:14 The countdown is on. From March 2326, the cybersecurity community will gather once again at the Moscone Center in San Francisco, and Thales will be at the heart of it. Cybersecurity Chad Couser – Director Marketing Communications Thales More…
-
Augustus v0.0.9: Multi-Turn Attacks for LLMs That Fight Back
Single-turn jailbreaks are getting caught. Guardrails have matured. The easy wins, “ignore previous instructions,” base64-encoded payloads, DAN prompts, trigger refusals on most production models within milliseconds. But real attackers don’t give up after one message. They have conversations. Augustus v0.0.9 now ships with a unified engine for LLM multi-turn attacks, with four distinct… First seen…
-
Researchers Discover Major Security Gaps in LLM Guardrails
Palo Alto Networks’ Unit 42 has developed a successful attack to bypass safety guardrails in popular generative AI tools First seen on infosecurity-magazine.com Jump to article: www.infosecurity-magazine.com/news/major-security-gaps-llm-guardrails/