Return to Threats

Gheist Uses Stolen GitHub Tokens to Steal Private Repository Secrets via Actions Logs

thehackernews.com 2026-10-26 data leakage Critical
Run AI Security Assessment Talk to AI CISO

Executive Summary

Attackers used stolen GitHub tokens to exfiltrate private repository secrets through malicious GitHub Actions workflows and then delete the runs. Teams operating AI agents or LLM applications through GitHub should treat repository tokens, CI/CD workflows, and stored secrets as high-priority security assets.

What Happened

An automated campaign is using stolen GitHub tokens to plant workflows that print a repository's secrets into its GitHub Actions logs, which the attackers then download before deleting the runs. The attacks have succeeded in private repositories and GitHub Enterprise environments, according to an analysis by cloud security company Wiz. A GitHub safeguard that pauses

Affected Products

  • GitHub
  • GitHub Actions
  • GitHub Enterprise

Vulnerability / Attack Technique

Stolen-token abuse through malicious CI/CD workflows

Technical Mechanism

Attackers used stolen GitHub tokens to plant workflows that printed repository secrets into GitHub Actions logs. They then downloaded the logs containing the secrets and deleted the workflow runs; the reported attacks affected private repositories and GitHub Enterprise environments.

AI Security Category

AI-assisted attack

Part of the AI-Assisted Attacks topic.

Relevant Frameworks

MITRE ATLAS: Credential AccessMITRE ATLAS: ExecutionMITRE ATLAS: CollectionMITRE ATLAS: ExfiltrationMITRE ATLAS: Defense Evasion

RealGround Analysis

Although the reported target is GitHub rather than an AI system, the technique can expose source code, deployment credentials, prompts, tools, or configuration used by AI agents and LLM applications. Teams should include CI/CD token abuse and workflow tampering in AI security assessments, especially where repositories contain agent logic, model integrations, or secrets.

Recommended Controls

  • Use short-lived, narrowly scoped GitHub tokens and rotate or revoke exposed credentials immediately.
  • Restrict GitHub Actions permissions using least privilege and require approval for workflow changes in protected repositories.
  • Prevent secrets from being printed to CI/CD logs and apply secret scanning to repositories, workflows, and build outputs.
  • Monitor workflow creation, modification, execution, log access, token use, and deletion of workflow runs.
  • Protect repositories containing AI agent code, prompts, tool definitions, and deployment configuration with branch protection and mandatory reviews.

Related RealGround Service

AI Security Readiness Assessment · or start with the free AI Agent Security Scanner.

Sources

thehackernews.com: https://thehackernews.com/2026/10/gheist-uses-stolen-github-tokens-to.html

Summary and analysis written by RealGround from the source above; last updated 2026-10-26.

Talk to AI CISO