Trusted AI Blog

• Agentic AI Security

Anthropic's Mythos completed a 32-step network attack autonomously in hours.

• LLM Security

IICL is a structural jailbreak bypassing GPT-5.4 safety with just two words.

• Agentic AI Security

A practical framework for comparing manual, in-house, and continuous red teaming of AI agents across coverage, cost, staffing, and compliance needs.

Adversa AI won in the Safety and Alignment category, recognized for advancing real-world AI safety through continuous adversarial testing of AI systems.

• Agentic AI Security

Full technical guide to Agent Goal Hijack, the #1 risk in the OWASP Agentic Top 10.

• GenAI Security

April 2026 in GenAI security: 19 resources on the LiteLLM supply chain compromise, 128K context window poisoning, and compound RAG database exploits.

• MCP Security

Our April 2026 MCP resources digest highlights the latest vulnerability research and practical defenses.

• Agentic AI Security

Adversa AI red team found Claude Code's deny rules silently stop working after 50 subcommands. The fix exists in Anthropic's codebase. They never shipped it

• Agentic AI Security

April 2026 in agentic AI security: privilege escalation in OpenClaw, the hijacking of Chrome's Gemini Live assistant, and the RSA conversation around both.

• Agentic AI Security

Recognized Among Hundreds of Vendors for Advancing Continuous AI Red Teaming and Agentic AI Security.

• Agentic AI Security

The six threat actors your red team should be simulating, the five expertise domains needed to find them, and the uncomfortable math between the two.

• Agentic AI Security

Our agent made it to the top 3 in Gandalf CTF for agents. It predicts vulnerabilities before sending a single attack.