Originally reported by Wiz Blog
TL;DR
Security researchers at Wiz have identified prt-scan, a new AI-powered supply chain attack campaign that exploited GitHub's pull_request_target workflow feature across multiple accounts. The campaign operated for three weeks before detection, highlighting the growing sophistication of automated attacks targeting CI/CD pipelines.
While this represents a sophisticated supply chain attack campaign using AI automation, it appears to be contained to specific GitHub repositories without evidence of widespread compromise or critical infrastructure impact.
Wiz security researchers have uncovered prt-scan, a sophisticated supply chain attack campaign that leveraged AI automation to exploit GitHub's pull_request_target workflow feature. The campaign represents the latest evolution in automated attacks targeting software development infrastructure.
The researchers traced the attack activity across six different GitHub accounts, with operations beginning three weeks before the campaign was first detected. This extended dwell time demonstrates the attackers' ability to maintain persistence while avoiding detection in CI/CD environments.
The prt-scan campaign specifically targeted GitHub repositories utilizing the pull_request_target workflow trigger, a feature that allows workflows to run with elevated privileges when processing pull requests from forks. This mechanism, designed to enable certain automation scenarios, creates a potential attack vector when improperly configured.
According to Wiz's analysis, the attackers employed AI-powered automation to:
pull_request_targetThe research team identified coordinated activity across six distinct GitHub accounts, suggesting a single threat actor operating multiple personas. The campaign timeline reveals:
This campaign follows the earlier hackerbot-claw attacks, indicating that AI-powered supply chain threats are becoming an established attack pattern rather than isolated incidents. The use of automation allows threat actors to scale their operations significantly while maintaining consistent operational security.
The targeting of pull_request_target workflows specifically demonstrates attackers' deep understanding of GitHub Actions security models and their willingness to exploit complex CI/CD features for malicious purposes.
Organizations using GitHub Actions should review their workflow configurations, particularly those utilizing pull_request_target triggers. Key security measures include:
pull_request_targetOriginally reported by Wiz Blog