llm • AI Security Wire

img of Prompt Injection via Third-Party Plugins: A Growing LLM Supply Chain Risk

Threat actors are embedding prompt injection payloads in third-party LLM plugins and data sources to hijack AI agent actions, exfiltrate data, and pivot within enterprise environments.

img of Many-Shot Jailbreaking: Long-Context Windows as an Attack Surface

May 22, 2026 4 min read

Research

Many-Shot Jailbreaking: Long-Context Windows as an Attack Surface

Research demonstrates that LLMs with large context windows can be reliably jailbroken by embedding hundreds of fictitious dialogues before the target request — a technique that scales with context length and bypasses standard safety training.

img of Incident Report: LLM-Assisted Intrusion at a Mid-Market Financial Firm

May 20, 2026 6 min read

Incident Reports

Incident Report: LLM-Assisted Intrusion at a Mid-Market Financial Firm

A detailed post-mortem of a multi-stage intrusion in which threat actors used LLM-generated spear phishing, AI-assisted credential stuffing, and automated reconnaissance to compromise a wealth management firm — from initial access to detection.

img of Model Inversion Attacks: Extracting Training Data PII from Production LLMs

May 17, 2026 2 min read

Vulnerabilities

Model Inversion Attacks: Extracting Training Data PII from Production LLMs

Model inversion and training data extraction attacks allow adversaries to recover PII, proprietary data, and trade secrets from fine-tuned LLMs exposed via API — a significant compliance and IP risk for enterprises.