Hackers launched a coordinated command injection campaign targeting 61 public GitHub repositories, aiming to steal data from automated development systems This article explores malicious issues pull. . Radware researchers discovered 83 malicious issues, pull requests, and comments posted on September 1, 2026, with evidence that injected commands executed inside two GitHub Actions workflows.
The findings highlight how routine project discussions can serve as entry points into software build systems.
According to Radware’s investigation, 77 malicious items appeared within about 43 minutes, with six more following in a second wave lasting roughly two minutes. GitHub Actions automatically respond to issues, pull requests, or comments, performing tasks such as checking titles, assigning labels, or sending notifications. The payload ran env to gather environment variables, encoded the output with Base64, stripped newline characters, and used curl to attempt an HTTP POST request.
PX4/PX4-Autopilot — Pull Request #28488 (source: Radware) Every verified payload pointed to 54.205.16[. ]142 on port 8443, with a separate collection identifier for each attempt. Researchers discovered signs that some participating accounts may have been hijacked, including one owner publicly denying they opened a malicious issue.
Maintainers should pass untrusted titles through environment variables instead of inserting them into shell scripts, avoid using eval, and limit workflow permissions.
Teams should review available secrets, rotate potentially exposed credentials, inspect audit logs, and investigate unexpected outbound traffic.
