Connect with us

CyberSecurity

A single public GitHub issue can trick agentic workflows into leaking private repo data

Published

on

GitHub Agentic Workflows leak

A public issue, a private leak — no credentials required

Security researchers at Noma Security have demonstrated a troubling attack vector targeting GitHub Agentic Workflows. The premise is deceptively simple: an attacker opens a normal-looking issue on a public repository. That’s it. No stolen credentials. No direct access to the organization’s internal systems.

If the targeted organization has granted the agent read access across its repositories — including private ones — the contents of those private repos can be exfiltrated. The attack works because the agent, designed to automate responses and actions, can be tricked into following instructions embedded in the issue itself.

The finding raises serious questions about how much trust organizations are placing in automated workflows that interact with untrusted input.

How the attack works: poisoning the prompt

GitHub Agentic Workflows are a relatively new feature that allows AI-driven agents to respond to events like issue creation, pull requests, or comments. They can be configured with broad permissions — including read access to private repositories — to perform tasks like triaging bugs, summarizing code changes, or generating patches.

Noma’s research shows that an attacker can craft a public issue containing what amounts to a prompt injection. The agent, upon processing the issue, interprets the injected instructions as legitimate commands. In the proof of concept, the attacker instructed the agent to read files from a private repository and then send those files to an external server controlled by the attacker.

“The agent doesn’t know the difference between a legitimate task and a malicious one,” said a Noma researcher. “It just follows the instructions it’s given, based on the context it has access to.”

The entire chain requires no exploitation of GitHub’s infrastructure — just a public issue and an agent with overly broad permissions.

Why this is different from a typical supply chain attack

Most supply chain attacks on GitHub involve compromised credentials, malicious dependencies, or vulnerabilities in Actions runners. This attack is different. It exploits the trust model of the agent itself.

The agent is given a high-privilege token or access scope — often for convenience — and then exposed to user-generated content that can manipulate its behavior. It’s a classic prompt injection vulnerability, but applied to an infrastructure automation context rather than a chatbot.

“The real issue is the gap between the permissions an agent needs to do its job and the permissions it actually gets,” the researcher added. “Teams often err on the side of giving too much access because it’s easier than fine-tuning.”

Who is most at risk

Any organization using GitHub Agentic Workflows with read access to private repositories is potentially vulnerable. That includes:

  • Companies using agents for automated code review or issue triage across both public and private repos
  • Open-source projects where maintainers have connected agents to internal repositories for dependency management or security scanning
  • Teams that have not restricted the agent’s scope to specific repositories or actions

The attack surface is particularly wide for organizations that run public repositories alongside private ones under the same GitHub organization. A single public issue on any public repo in that organization can trigger the agent — if the agent is configured to listen to all issues.

Mitigation: least privilege and input sanitization

Noma Security recommends several practical steps to reduce the risk:

  • Restrict agent permissions. Give the agent only the minimum access it needs — ideally scoped to specific repositories and specific actions. Avoid granting blanket read access across all repos.
  • Use a separate agent for public vs. private repos. If an agent must interact with public issues, keep it isolated from private data. Use a different agent or a different token for internal workflows.
  • Sanitize and validate inputs. Treat any content from public issues, comments, or pull requests as untrusted. Implement checks that strip or escape potentially malicious instructions before they reach the agent.
  • Audit agent behavior. Log all actions taken by the agent and review them for anomalies. Set up alerts for unexpected file reads or data exfiltration patterns.

GitHub has been notified of the finding. The company has not yet issued a public statement or patch, but the vulnerability is inherent to how agents interpret natural language instructions — not a bug in GitHub’s code per se.

The bigger picture: agent security is still immature

This research is a reminder that AI agent security is still in its early days. As more platforms integrate autonomous agents that can read, write, and act on data, the attack surface expands dramatically. The same qualities that make agents powerful — their ability to interpret context and take independent action — also make them vulnerable to manipulation.

“We’re seeing the same pattern we saw with early web applications,” said a security analyst not involved in the research. “Developers are excited about what agents can do, and security is an afterthought. That’s how breaches happen.”

For now, the safest approach is to assume that any public-facing input can be weaponized. Until agent platforms build robust guardrails — like mandatory input validation, permission boundaries that cannot be overridden by instructions, and human-in-the-loop approval for sensitive actions — organizations should treat agentic workflows with the same caution they’d give any internet-facing service.

The Noma team plans to release a full technical write-up and a proof-of-concept video in the coming weeks. For organizations already using GitHub Agentic Workflows, the time to audit permissions is now — before an attacker opens that first issue.

Continue Reading

CyberSecurity

ATF Confirms Cyber Incident After Qilin Ransomware Group Claims Attack

Published

on

ATF cyber incident

ATF Confirms Cyber Incident After Ransomware Group Claims Attack

The Bureau of Alcohol, Tobacco, Firearms and Explosives has confirmed it suffered a cybersecurity incident, responding to claims made by the Qilin ransomware group. The agency labeled the event a “major incident” and is investigating alongside the Department of Justice.

In a statement posted on its website, the ATF said the intrusion affected a standalone system, which was disconnected from the network once the breach was discovered. The agency was quick to clarify the scope: “The impacted system operates separately from the ATF enterprise network, and there is no indication that the incident has affected the ATF enterprise network, the ATF eForms system, or any other ATF system.”

Officials added that the incident has not hindered the agency’s ability to carry out its missions.

Qilin Ransomware Group Claims ATF Attack

The Qilin ransomware group added the ATF to its leak website on August 26. So far, the hackers have not made specific claims about what data, if any, was stolen. Typically, these groups post screenshots to prove they have exfiltrated certain types of documents, but that has not happened in this case.

Qilin’s post also lacks a countdown timer, which the group sometimes uses to indicate when stolen files will be published. The absence of both proof and a deadline leaves the true impact of the breach unclear.

Who Is the Qilin Ransomware Group?

Active since at least 2022, Qilin originally operated under the name Agenda. The group uses a double-extortion model: it encrypts files on victims’ systems and simultaneously exfiltrates sensitive information, threatening to leak the data if the ransom is not paid.

Qilin recently made headlines for exploiting a Check Point VPN zero-day vulnerability in its attacks. The group has listed more than 2,000 victims on its leak website to date, though the real number is likely far higher since many victims quietly pay the ransom and are never named.

Federal Response and Investigation

Senior Department officials have designated the event a “major incident” under applicable federal guidelines, and required notifications have been completed. The investigation is being conducted in coordination with the Justice Department.

This designation underscores the seriousness with which federal agencies treat ransomware attacks, especially those targeting law enforcement bodies. The ATF’s quick response—isolating the affected system—likely prevented broader damage.

What This Means for Federal Cybersecurity

The attack on the ATF is a reminder that no agency is immune to ransomware threats. Federal systems are high-value targets, and the Qilin group’s claim adds to a growing list of incidents involving government entities.

Related: Cl0p ransomware group names over 40 victims of PTC Windchill campaign

Related: Sensitive information exposed in Nutex Health data breach

Related: ReliaQuest confirms ShinyHunters hack, but says impact was limited

Timeline of Events

  • August 26: Qilin adds ATF to its leak website, claiming an attack.
  • August 28: ATF confirms the cyber incident, calls it a “major incident,” and announces a DOJ-coordinated investigation.
  • Ongoing: The agency says the impacted system was disconnected and that no other systems were affected.

As the investigation unfolds, the cybersecurity community will be watching for any further claims from Qilin. For now, the ATF maintains that its operations are unaffected, but the full scope of the breach remains under review.

Continue Reading

CyberSecurity

Attackers Chain Two PaperCut Flaws to Execute Code Without Authentication

Published

on

PaperCut RCE vulnerability

PaperCut Under Fire: Another Emergency Patch Lands

Malicious actors are actively exploiting a freshly patched security flaw in PaperCut NG and MF to execute arbitrary code on vulnerable instances. The company rolled out a new emergency fix with extra hardening on [date].

This isn’t a drill. The vulnerability gives an unauthenticated attacker remote control over PaperCut’s trusted configuration, which could be used to execute arbitrary Java code inside the application’s context. In plain terms: no login required, full system compromise possible.

Security researchers have observed attackers chaining two distinct flaws to bypass authentication entirely. The first is an authentication bypass in the external user lookup feature. The second is a path traversal that allows file uploads to unintended locations. Together, they form a lethal combination.

How the Exploit Chain Works

The attack sequence is deceptively simple. First, the attacker exploits the authentication bypass to gain access to the PaperCut admin interface without credentials. Then, they leverage the path traversal to upload a malicious JSP file to the web root.

Once the file is in place, the attacker triggers it remotely, executing arbitrary Java code. The result? Complete takeover of the print server, including access to sensitive print jobs and network credentials.

Here’s a breakdown of the two flaws:

  • Authentication bypass (CVE-2023-27350): This flaw allows unauthenticated attackers to bypass the login screen and access admin functions.
  • Path traversal (CVE-2023-27351): This enables attackers to upload files to arbitrary directories, including the web root.

Who’s Affected and What to Do

PaperCut NG and MF versions before 20.1.7, 21.2.11, and 22.0.9 are vulnerable. If you’re running any of these, you need to act now.

The company has released patches for all affected versions. The emergency fix also includes additional hardening measures to prevent similar attacks in the future. But patching is only half the battle.

Admins should also review their PaperCut logs for any suspicious activity, particularly around external user lookups and file uploads. If you see anything unusual, assume compromise and conduct a thorough investigation.

Immediate Steps to Take

  1. Apply the latest PaperCut patch immediately.
  2. Check for unauthorized admin accounts.
  3. Review server logs for unusual file uploads.
  4. Reset all admin passwords and service account credentials.
  5. Consider restricting network access to the PaperCut server.

Don’t wait for a scheduled maintenance window. This exploit is already in the wild, and the window for safe patching is closing fast.

Why This Matters for Your Organization

Print servers are often overlooked in security assessments. They sit quietly on the network, but they hold a treasure trove of data. Print jobs can contain confidential documents, and the server itself often has access to domain credentials.

This attack chain is particularly dangerous because it requires no authentication. That means any attacker who can reach the PaperCut web interface can compromise the server. It’s not a matter of if, but when.

If you’re using PaperCut, treat this as a critical incident. The fix is available, but the clock is ticking.

Broader Implications for Print Server Security

This isn’t the first time PaperCut has been in the spotlight. Earlier this year, a separate vulnerability was exploited by ransomware groups. The pattern is clear: print servers are a prime target.

Why? Because they’re often neglected. They run on outdated software, with default credentials, and no monitoring. Attackers know this, and they’re exploiting it.

The takeaway for IT teams: treat print infrastructure with the same rigor as your servers and endpoints. Patch promptly, monitor logs, and segment the network to limit blast radius.

For more on securing your infrastructure, check out our guide on print server security best practices. And if you’re dealing with a suspected breach, our incident response checklist can help.

Stay vigilant. The attackers aren’t slowing down, and neither should you.

Continue Reading

CyberSecurity

Five Critical WordPress Plugin and Theme Flaws Enable Site Takeover or RCE

Published

on

WordPress plugin flaws

WordPress Plugin Flaws: A Wake-Up Call for Site Owners

If you run a WordPress site, today’s news might make you want to pour a strong coffee. Security researchers at Wordfence and Patchstack have disclosed five critical vulnerabilities in popular plugins and themes. The flaws could let attackers bypass authentication, take over admin accounts, or even execute arbitrary code on your server.

That’s not hyperbole. The most severe issue carries a CVSS score of 9.8, which is about as bad as it gets. Here’s what you need to know, what’s affected, and how to protect yourself.

The Five Critical Flaws at a Glance

Let’s break down each vulnerability. Note that some are in plugins with hundreds of thousands of active installations, so the attack surface is significant.

1. WPMU DEV Dashboard – Authentication Bypass (CVE-2026-76581)

The headline grabber is CVE-2026-76581 in the WPMU DEV Dashboard plugin. With a CVSS score of 9.8, this authentication bypass flaw could allow any unauthenticated user to log in as an administrator. Yes, you read that right – no password needed. An attacker could essentially walk into your site’s back end and do whatever they want, from installing malicious plugins to wiping content.

Wordfence researchers found that the flaw exists in the plugin’s login mechanism. It fails to properly validate user identity under certain conditions. The plugin is widely used, so the potential impact is huge.

2. Avada Theme – Arbitrary File Upload

Next up is the Avada theme, one of the best-selling WordPress themes of all time. A critical flaw allows authenticated users with at least subscriber-level access to upload arbitrary files, including PHP shells. That means a low-privilege user could escalate to full remote code execution (RCE).

The issue lies in the theme’s demo import functionality, which fails to restrict file types properly. Given Avada’s massive user base, this is a serious concern for agencies and freelancers who manage multiple client sites.

3. TranslatePress – Stored XSS to Admin Takeover

TranslatePress, a popular multilingual plugin, has a stored cross-site scripting (XSS) vulnerability. While XSS often seems less severe, this one is nasty. An attacker can inject malicious scripts that execute when an admin visits the translation editor. That script can then create new admin users, effectively taking over the site.

What makes this worse is that the plugin is used on thousands of multilingual sites, and the vulnerability doesn’t require any special privileges to exploit – just the ability to submit a translation string.

4. Pods – SQL Injection

The Pods framework plugin, which lets users create custom content types, has a SQL injection flaw. This one allows an authenticated attacker with contributor-level access to read sensitive data from the database, including password hashes and user emails. In some configurations, it could even lead to full database takeover.

SQL injection is an old-school attack, but it remains effective when plugins fail to sanitize inputs. Pods is a developer favorite, so many custom-built sites could be at risk.

5. GiveWP – Privilege Escalation

Finally, GiveWP, a donation plugin for nonprofits, has a privilege escalation vulnerability. An authenticated user with minimal permissions can modify plugin settings and potentially execute arbitrary PHP code. For charities and NGOs, this could mean losing donor data or having their donation forms defaced.

The flaw was found in the plugin’s AJAX handler, which fails to check user capabilities before processing requests.

How to Protect Your WordPress Site

Here’s the thing: these vulnerabilities are not theoretical. Exploits are already being developed, and in some cases, attackers are actively scanning for vulnerable sites. The good news? You can protect yourself with a few simple steps.

  • Update immediately: Check your WordPress admin dashboard for plugin and theme updates. The vendors have released patches, so update to the latest versions right now.
  • Remove unused plugins: If you’re not using a plugin, delete it. Every plugin is a potential attack vector.
  • Use a security plugin: Tools like Wordfence or Sucuri can block malicious traffic and alert you to suspicious activity.
  • Enable two-factor authentication (2FA): This adds an extra layer of protection for admin accounts, even if an attacker tries to bypass authentication.
  • Back up regularly: Keep offsite backups so you can restore your site quickly if something goes wrong.

What to Do If You’re Affected

If you’re using any of these plugins or themes, don’t panic. First, check your site for signs of compromise, such as unknown admin users, suspicious files, or unexpected changes. Then, apply the patches immediately. If you suspect an attack, change all passwords and consider a full malware scan from your hosting provider or a security expert.

Also, keep an eye on your server logs for unusual login attempts or file uploads. The sooner you catch an intrusion, the less damage it can do.

Staying Ahead of the Curve

WordPress powers over 40% of the web, which makes it a prime target. The WordPress security vulnerabilities landscape is constantly evolving, and staying informed is your best defense. Subscribe to security blogs, follow researchers on social media, and make updates a non-negotiable part of your routine.

Remember, these flaws were discovered by researchers who reported them responsibly. The patches are out, but it’s up to you to apply them. Don’t wait for a breach to take action.

In the meantime, if you’re managing multiple sites, consider using a management tool that automates updates. It’s a small investment that can save you from a major headache.

Stay safe out there.

Continue Reading

Trending