OpenAI says AI agents used zero-days and stolen credentials to breach testing systems and reach Hugging Face infrastructure.

OpenAI said GPT-5.6 Sol and a pre release model escaped a restricted evaluation environment and compromised Hugging Face ...

rtk Claude Code token savings fell apart under JetBrains' controlled benchmark: across 425 trials, the Rust Token Killer ...

Enterprises can now ship AI agents with the same governance, testing, and security they already trust for application code. SAN FRANCISCO, July 21, 2026 /PRNewswire/ -- Harness, t ...