Visa launches AI tool to automatically patch code vulnerabilities
Visa has released an open-source AI tool that identifies vulnerabilities, writes fixes, and tests them before human review. The tool aims to accelerate security patching processes.
Financial services company Visa has launched a new open-source artificial intelligence tool designed to automatically identify security vulnerabilities in software production code, write fixes for them, and test the patched code before human review.
The tool, named Visa Vulnerability Agentic Harness (VVAH), operates by default, running through all stages to detect issues, write code corrections, and apply them directly to the target repository unless manually capped. Visa states this automated process is necessary as AI can now find vulnerabilities faster than humans can fix and verify them.
VVAH originated from Visa's participation in Anthropic's Project Glasswing. During this project, Visa utilized AI models to analyze the network behind billions of daily transactions, observing how the models could chain together minor weaknesses into functional exploits.
The tool has been made available on GitHub, where it has rapidly gained traction within the developer community. Visa has also expanded its advisory services to support clients with cybersecurity concerns. The company claims the tool's automated remediation process can reduce the time to resolve vulnerabilities from weeks to hours.