Marcelo Morales

Summary: Trail of Bits published a post by Anish Naik, Dan Guido, Benjamin Samuels, and Marcelo Morales arguing that 1Password's AI patching benchmark is misleading. The post says the firm shares real-world data on human and agent patch quality from its consulting work and Patch the Planet, and releases two agent skills for testing and reviewing security fixes.
Key Takeaway: A benchmark is what vendors and buyers use to argue that AI-written patches are safe; Trail of Bits is asserting that this one misrepresents the picture and is countering with its own consulting data on how good those patches actually are.

Read the original at Trail of Bits Blog ↗