Enterprises Face Top Vibe-Coding Risk as 62% Use It for Speed, Exposing Secrets on GitHub
Updated
Updated · Computerworld · Aug 18
Enterprises Face Top Vibe-Coding Risk as 62% Use It for Speed, Exposing Secrets on GitHub
2 articles · Updated · Computerworld · Aug 18
Summary
Hard-coded secrets uploaded through AI-generated apps to GitHub are now the biggest enterprise risk from vibe coding, Gartner analysts said, warning they can open routes into corporate systems and leak sensitive data publicly.
AI coding tools can produce working software in hours instead of weeks, but that speed often bypasses code reviews and security checks, leaving hidden vulnerabilities, logic errors and hallucinated code undetected.
Non-technical employees are increasingly being encouraged to build apps this way, expanding the attack surface as hundreds of ad hoc applications can be created without consistent scanning or patching, Forrester and Gartner said.
Researchers said vibe coding is useful for prototypes, not a replacement for enterprise software development, and should be governed with heavy auditing, validation, skilled prompting and continuous technical-debt monitoring.