I loaded a large JSON file in Python and watched memory jump from 43 MB to 808 MB. Here’s what happened and how streaming changed everything.

LiteLLM version 1.107.0, distributed via PyPI, was found to contain code designed to steal credentials and a backdoor.

Changes That Actually Matter for Developers written by Faisal Haque, I have organized the points that I personally find ...

Compromised LiteLLM 1.107.0 packages run a credential stealer at Python startup and can fetch and execute additional code.

Brotli decoder flaw CVE-2026-111497 affects Node.js, Bun, Python packages and Envoy; code execution was demonstrated on one ...

LiteLLM supply chain attack: 1.107.0 shipped a credential stealer to PyPI and Docker on Oct. 21, two days after OVHcloud CLI ...

Prompt injection test lab: build a 12-step Python harness that scores Claude, GPT, Gemini, Mistral and DeepSeek on 14 OWASP ...

LiteLLM’s second 2026 supply chain attack exposed cloud credentials, AI API keys and SSH keys through malicious v1.107.0 ...

An agent harness, sometimes called an AI harness, is the runtime scaffolding that turns a language model into an agent that can perform work. It drives model and tool calls, manages conversation state ...

Attackers used stolen GitHub credentials to push a backdoored version of the popular AI gateway, exposing cloud, Kubernetes, and CI/CD secrets.

Security researchers have uncovered another LiteLLM package hijack, warning that malicious code in version 1.107.0 can steal ...