Skip to content

Threat Research

When the Test Escaped the Lab: OpenAI’s “Rogue” Models and the Hugging Face Breach

The short version During an internal cyber-capability test in July 2026, two OpenAI models broke out of an isolated testing sandbox, reached the open internet, and compromised the AI hub Hugging Face to obtain the answers to the benchmark they were being scored on. The models were run with their safety refusals switched off, deliberately,… Read More »When the Test Escaped the Lab: OpenAI’s “Rogue” Models and the Hugging Face Breach

Before Google Cloud Backed gRPC for MCP, FireCompass Had Already Built It for Switchblade

Before Google Cloud Backed gRPC for MCP, FireCompass Built It for Our AI Agents

Early January, I published a deep dive into why we ripped out JSON-RPC and rewrote our Model Context Protocol (MCP) server using gRPC for our internal AI initiatives. The idea was simple: if you are building enterprise-grade agents, you cannot rely on the “guesswork” of dynamic JSON payloads. You need the strict guarantees of Protobufs.… Read More »Before Google Cloud Backed gRPC for MCP, FireCompass Built It for Our AI Agents

Verizon DBIR 2026: The Year Vulnerability Exploitation Beat Credentials. What It Means for Your Pen Testing Program

The Verizon 2026 Data Breach Investigations Report dropped this week, and for the first time in the report’s 19-year history, vulnerability exploitation has overtaken credential abuse as the top initial access vector. Exploited vulnerabilities now account for 31% of breaches, up from 20% the year before. Credentials dropped from 22% to 13%. Read past the… Read More »Verizon DBIR 2026: The Year Vulnerability Exploitation Beat Credentials. What It Means for Your Pen Testing Program

Use Cases
Technology
About
Partner
Resources

FireCompass – An EC Council Ecosystem Company.
©2025 . All Rights Reserved.

Firecompass ranked #1 AI on HackerOne. Read more →