All articles
A briefing on AI security. The breaches, CVEs, jailbreaks, regulatory actions, and tooling shifts that mattered — in one place, with sources.
-
AI Security Scanning Tools for LLMs: 2026 Picks Compared
garak, PyRIT and Promptfoo probe LLM behavior, ModelScan and picklescan check model files, and Snyk Agent Scan audits MCP servers.
-
LLM Security Audit Checklist: 10 Controls Mapped to OWASP
A 10-point LLM security audit checklist mapped to the OWASP Top 10 for LLM Applications 2025, with the evidence to collect and the fail conditions.
-
AI Red Team Methodology Guide: Scope, Phases, and Evidence
The four OWASP red-team scopes, threat modeling with MITRE ATLAS and the OWASP LLM Top 10, the NIST AI 600-1 team types, and what closes a finding.
-
LLM Prompt Injection Defense: What Actually Works
A layered guide to LLM prompt injection defense: prompt-level controls, classifier screening, agent design patterns, and the output gates that contain it.
-
Best LLM Firewall Tools in 2026: A Buying Guide
Managed APIs, in-process frameworks and open-weight classifiers compared: how to pick an LLM firewall architecture, and how to layer all three.
-
LLM Jailbreak Defense Techniques: A Practitioner's Guide
A breakdown of LLM jailbreak defense techniques, from input filtering and model-level hardening to multi-agent architectures and OWASP mitigations.
-
Prompt Injection Detection Tools: The 2026 Landscape
A field guide to prompt injection detection tools: Llama Prompt Guard 2, Azure Prompt Shields, Lakera Guard, NeMo Guardrails, and the open-source options.
-
AI Agent Security Risks 2026: Injection to Privilege Abuse
OWASP's Q1 2026 GenAI Exploit Round-up confirms AI agent security risks have moved from research to active production breaches.
-
Prompt Injection Attack Examples: Direct, Indirect, Agentic
A practitioner breakdown of prompt injection examples, from basic instruction overrides to CVE-rated zero-click exploits in production agent deployments.
-
AI on Offense: Google's Zero-Day Warning and Jailbreaks
Google says it caught an attacker using an LLM to find a zero-day, and peer-reviewed work shows reasoning models can autonomously jailbreak other models.
-
How LLM Chatbots Leak Data Through Their Own Rendered Output
A recurring AI-security finding: an injected instruction makes the model emit a markdown image whose URL carries the user's data to an attacker server.
-
LiteLLM SQL Injection CVE-2026-42208: The AI Gateway Risk
A pre-auth SQL injection in the LiteLLM proxy landed in CISA's KEV catalog and was exploited within roughly 36 hours of disclosure.
-
Indirect Prompt Injection: The Agent Era's Default Flaw
As LLM agents gained tools and memory, the dangerous injection stopped coming from the user and started coming from the data the agent reads.
-
How AI Sec Weekly Selects and Sources What It Covers
How AI Sec Weekly chooses what to cover: the four-question inclusion bar, the primary-source rule, and the house style every entry is written to.