Attackers Are Targeting the Tools Developers Trust: Google Urges Stronger Defences for Software Build Pipelines

Google Threat Intelligence says software supply chain security has shifted significantly. Recent campaigns show skilled threat actors going after the engineering lifecycle itself by compromising tools that developers and security teams already trust. Because security scanners, utility libraries and AI developer tools are often given elevated privileges inside build pipelines, they make attractive targets.

Google describes three key tactics. First, attackers abuse trusted scanners, libraries and AI tools to exploit their high level of access. Second, they target developer workstations and development environments using tailored social engineering, malicious extensions or typosquatted local dependencies (fake packages with near-identical names), aiming to steal private cryptographic keys, API tokens and active session credentials. Third, attackers have moved beyond stolen static credentials to pipeline manipulation, including GitHub Actions cache poisoning, OpenID Connect (OIDC) token extraction, and subverting mutable action tags. This lets them publish compromised packages that still carry legitimate cryptographic provenance.

The article builds on earlier Google guidance and sets out a blueprint for software and platform architects. Its central point is that treating each stage of a pipeline as a separate security domain is no longer enough, because these attacks span the entire build process. Google recommends a defence-in-depth approach built around five pillars of the software development lifecycle. This summary covers only the opening of the article, so the detailed recommendations for each pillar are not included here.