AI Coding Agents Exposed 13,000 Internal Images, Including Billing Records, on GitHub

The Hacker News

Overview

A recent report from security firm Glow revealed that AI coding agents inadvertently exposed over 13,000 internal images on public GitHub repositories. These images, belonging to developers from more than 300 organizations, included sensitive information such as customer billing records and unreleased feature screenshots. The issue primarily arose from developers using their personal GitHub accounts to share code review screenshots. This incident raises serious concerns about data privacy and security, as sensitive company information is now publicly accessible. Organizations need to reassess their use of AI tools and ensure that proper security measures are in place to protect internal data.

Key Takeaways

  • Affected Systems: Internal company images, customer billing records, unreleased feature screenshots
  • Action Required: Organizations should review the use of AI coding agents, enforce secure account practices, and implement stricter access controls on sensitive information shared in public repositories.
  • Timeline: Newly disclosed

Original Article Summary

AI coding agents asked to share screenshots of code changes for review have put internal company images in public GitHub repositories, security company Glow said. Its researchers found more than 13,000 internal images from developers at over 300 organizations, including customer billing records and screens of features not yet released. In most cases, they sat under developers' personal accounts

Impact

Internal company images, customer billing records, unreleased feature screenshots

Exploitation Status

No active exploitation has been reported at this time. However, organizations should still apply patches promptly as proof-of-concept code may exist.

Timeline

Newly disclosed

Remediation

Organizations should review the use of AI coding agents, enforce secure account practices, and implement stricter access controls on sensitive information shared in public repositories.

Additional Information

This threat intelligence is aggregated from trusted cybersecurity sources. For the most up-to-date information, technical details, and official vendor guidance, please refer to the original article linked below.

Related Coverage

Attackers Exploit Zimbra Flaw to Deploy Web Shells and Harvest Authentication Secrets

The Hacker News

Recent findings from Microsoft's Security Research team reveal that attackers have exploited a serious vulnerability in the Zimbra Collaboration Suite (ZCS). The flaw, identified as CVE-2026-73570, allows for unauthenticated command injection, leading to remote code execution. This means that attackers can deploy web shells to access sensitive mailbox data. Organizations using ZCS should be particularly vigilant, as this vulnerability carries a high severity score of 8.9. The fact that attackers are taking advantage of this flaw emphasizes the importance of keeping software updated and applying patches promptly to safeguard against potential data breaches.

Sep 30, 2026

Cisco Warns of Attackers Exploiting Critical Authentication Bypass in SD-WAN Manager

The Hacker News

Cisco has issued a warning about a critical zero-day vulnerability affecting its Catalyst SD-WAN Manager, which is used by companies to oversee their SD-WAN networks. The flaw, identified as CVE-2026-76504, allows remote attackers to exploit the system's API without needing any login credentials, essentially granting them admin-level access. This poses a significant risk to organizations using this management software, as it could lead to unauthorized control and potential data breaches. Cisco has released patches to fix the issue, but there are no workarounds available for users who need immediate protection. Companies are urged to apply the fixes as soon as possible to mitigate the risk of exploitation.

Sep 30, 2026

Attackers Abuse ChatGPT Custom GPTs to Deliver RAT via ClickFix Lures

The Hacker News

Cybercriminals are exploiting ChatGPT's Custom GPTs feature to trick users into visiting harmful websites. These sites use ClickFix lures to deliver Remote Access Trojans (RATs), which can take control of victims' devices. This tactic was observed by Huntress in late September 2026 and represents a concerning trend where trusted AI platforms are manipulated to spread malware. Users who interact with these malicious GPTs may unknowingly expose their systems to significant risks. This incident serves as a reminder for individuals and organizations to be cautious when engaging with AI-driven tools and to verify the legitimacy of any offerings before clicking links.

Sep 30, 2026

AI's Third Wave: Coworkers Break the Security Model That Worked for Agents

BleepingComputer

The article discusses emerging security risks associated with persistent AI coworkers, which operate continuously and often have standing access to systems. This creates identity risks that current security models are not equipped to handle. Experts from Token Security emphasize the need for these AI agents to have their own distinct identities, designated owners, and carefully scoped permissions to mitigate potential security issues. Additionally, implementing lifecycle controls for these agents is crucial to ensure that they do not pose an ongoing risk. As AI increasingly integrates into workplaces, addressing these identity and access challenges is vital to protect sensitive information and maintain organizational security.

Sep 30, 2026

TeamViewer urges users to patch severe flaws “as soon as possible”

BleepingComputer

TeamViewer has issued an urgent warning for users to update their remote access software due to the discovery of several high-severity vulnerabilities in both its client and host applications. These flaws could potentially allow unauthorized access or control over user systems, putting sensitive data at risk. Users of TeamViewer are strongly advised to apply the patches as soon as possible to protect against potential exploitation. The vulnerabilities affect multiple versions of the software, which means that a wide range of users could be impacted. Failing to address these issues promptly could lead to serious security breaches for individuals and businesses alike.

Sep 30, 2026

Bitget hacked via zero-day in third-party security products

BleepingComputer

Bitget, a cryptocurrency exchange, reported a significant security breach where attackers stole $387.5 million by exploiting a zero-day vulnerability in third-party security products. This incident raises serious concerns about the reliability of third-party security solutions, especially in the cryptocurrency sector where security is paramount. The breach underscores the importance of continuous monitoring and updating of security systems to prevent similar attacks. Users and investors need to be aware of the potential risks associated with using exchanges that may be vulnerable to such exploits. This incident serves as a stark reminder of the evolving tactics employed by cybercriminals and the need for robust security measures in digital finance.

Sep 30, 2026