Defending LLMs Against LoRA-Enabled Supply Chain Attacks

- Posted in Hardenings by
Background: As an LLM trainer, it is a best practice to use 2–3 isolated environments for LLM training. One of the major risks in this process is supply chain attacks, and LoRA (Low-Rank [...] Read more

Securing LLM Outputs: Preventing Insecure Output Handling and Injection Attacks

- Posted in Hardenings by
Background: The threat landscape in the LLM era has become hybrid compared with the classic OWASP Top Ten. Interconnected LLM services have brought new types of attacks, and the boundaries between [...] Read more

Sensitive Data Disclosure over Leveraging LLM

- Posted in Threat Analyze by
Background: When leveraging agentic and non-agentic AI capabilities, especially when your solution is connected to your database, your organization can face another threat besides prompt injection: [...] Read more

Prompt Injection: Attack against LLM

- Posted in Threat Analyze by
Background: Prompt injection is one of the most well-known attacks against LLMs. The primary goal of a threat actor in such attacks is to extract secrets and other sensitive data from the environment [...] Read more

Data Poisoning Attacks on LLM leveraged Product Support Bots: Attack, Risk, Prevention

- Posted in Threat Analyze by
Background: Because industry started leveraging AI capabilities for routine cases like general customer support, a new threat has emerged: data poisoning. Core attack category: Based on MITRE ATLAS, [...] Read more

Malicious Use of Ansible by Threat Actors

- Posted in Hardenings by
Background: Ansible is a legitimate tool used by specialists to manage environments at scale. However, threat actors can also abuse this tool to deploy malicious payloads or establish persistence. [...] Read more

Morse Code and the Failure of Text-Based Detection

- Posted in Leak by
Background: Sometimes, in alert detection engineering, we can rely on plain-text detection, and even IDS rules can fail during body analysis. Attackers may achieve this by using more exotic methods [...] Read more

Kubernetes Forensics: Secret Rotation from Ruin to Recovery

- Posted in Incident Response by
Background: In this section, we will discuss specific actions during evidence collection, log analysis, and recovery for situations where it is necessary to determine if unencrypted secrets or [...] Read more

Shadow Workloads: How to Find and Monitor Unchecked Static Pods on Your Nodes

- Posted in Incident Response by
Background: During incident response activities at the cluster level, we should not focus solely on pods created via kubectl that are managed by the Kubernetes control plane (kubelet and etcd). It is [...] Read more

Kubernetes Incident Response Hunting: Identifying Anomalies in etcd During a Rapid Response Situation

- Posted in Incident Response by
Background: During incident response activities in Kubernetes, we need to ensure that artifacts have not been altered. One of the best places to hunt for artifacts is etcd, where Kubernetes stores [...] Read more
Page 1 of 11