NodeBB Patches Eight AI-Found Flaws Exposing Admin Access and Private Chats

Summary

NodeBB has released version 4.14.2 to address eight high-severity vulnerabilities found by Aikido Security's AI pentest agents. These flaws, affecting all versions prior to 4.14.0, could allow unauthorized users to gain admin access, read private messages, and execute code on the forum software. While most were patched quietly in earlier releases, the most significant fix, a rebuild of the page text handling, arrived in version 4.14.0.

Full text

NodeBB Patches Eight AI-Found Flaws Exposing Admin Access and Private Chats Swati KhandelwalJul 24, 2026Web Security / Vulnerability Eight security flaws in NodeBB went public on Wednesday, along with the code to exploit them. Aikido Security rates all eight as high severity and says its AI pentest agents found them in a six-hour review of the forum software's source code. Every version before 4.14.0 is affected. NodeBB has fixed them all, and administrators should be on 4.14.2. The simplest one takes a settings change. A regular forum member could point their homepage setting at the admin address, reload the page, and the admin dashboard opened for them. No password, no exploit code. The forum's own interface blocks that setting, but the block only ran in the browser and could be sidestepped. Most of what a member could then reach was read-only, including the error log and any user list an admin had exported, though they could also swap the site logo. Two more gave an attacker with no account at all access to things meant to be private. One let anyone claim to be any user and read private messages one at a time. The other handed over the contents of private categories to anyone who asked for them the right way. The widest flaw was in how NodeBB builds its pages. The software fills a page in, then makes a second pass to swap in translated text. User input was already sitting in the page by then, and it could smuggle in the codes that second pass looks for. That let an attacker plant a link almost anywhere on the site, including inside ordinary forum posts, that runs their code when a visitor clicks it. The rest let an attacker take over an existing post, inflate a post's vote count, and run two attacks that plant malicious code by way of a fake server on the fediverse, the network of connected social sites a NodeBB forum can join. Who Was Actually Exposed The eight are not equal. Three need no account on the target forum. Two need an ordinary member account. The last three need someone to click a link or open a page. Five of the eight, by The Hacker News' count, sit in NodeBB's federation code, the part that connects a forum to Mastodon and other social sites. That decides who was at risk. Forums installed fresh on version 4 federate by default, so they had all eight. Forums that upgraded from version 3 had federation switched off automatically, and unless an administrator turned it back on, only three of the flaws applied. Aikido published no severity scores for the individual flaws, and NodeBB's release notes do not rate them. NodeBB's own bug bounty scale rates cross-site scripting and account takeover as high, and getting admin access as critical. Patched in Pieces Since May NodeBB fixed most of them quietly, without saying what they were. The Hacker News checked each fix against NodeBB's release history: four shipped in May, two in June, and the biggest, a rebuild of how the software handles page text, arrived in 4.14.0 on July 9. That rebuild touched 325 files. Aikido's writeup says the issues were fixed in early July, which does not match that record. Its link for the admin-panel fix points to a change made in January 2024, two years before the review, while NodeBB's own release notes name a different change from May. Neither side explains the gap. Administrators should upgrade to 4.14.2, released July 23. Expect some work, because 4.14.0 changed how page templates handle text and custom themes and plugins may need updating. Switching federation off is not a full answer either, since three of the flaws have nothing to do with it. None of the eight has a CVE tracking number, and nobody has reported attacks using them. A separate NodeBB federation flaw does have one, CVE-2026-58593, filed on July 1. It is not one of Aikido's eight, but it sits in the same code and lets an outside server post and send messages in the name of any local account, the administrator's included. It needs federation switched on, and the record names no fixed version. NodeBB's bug bounty page says it rejects AI-generated reports and pays only for work the submitter did themselves. That governs payouts rather than fixes, and these eight were reported to the maintainers directly and patched. NodeBB is not the only project fielding them: the automation platform n8n patched a login flaw in June that a different AI pentest agent found. Co-founder Julian Lam's note in the release announcement says valid security reports arrived steadily through the month, "though almost all AI discovered and generated." The pattern behind all eight is the same. NodeBB checked who you were on the main way into a feature, and skipped the check on the side route that reached the same place. Found this article interesting? Follow us on Google News, Twitter and LinkedIn to read more exclusive content we post. SHARE     Tweet Share Share Share SHARE  AI Security, Application Security, Cross-site Scripting, Open Source, Software Security, Vulnerability, Web Security ⚡ Top Stories This Week URGENT - Progress Tells ShareFile Customers to Shut Down Storage Zone Controllers Over Security Threat Misconfigured Server Reveals Three Evilginx Phishing Operations Targeting Microsoft 365 Meta Files Patent for AI That Can Listen All Day and Track How You're Feeling New MemGhost Attack Plants Persistent False Memories in AI Agents Through One Email Microsoft Maps Three Salesforce Attack Paths Tied to a Year of ShinyHunters Activity OAuth Client ID Spoofing Lets Attackers Validate Stolen Microsoft Entra Credentials 11 Old Microsoft-Signed Linux UEFI Shims Could Let Attackers Bypass Secure Boot Researchers Say Claude for Chrome Flaw Lets Rogue Extensions Trigger Gmail Reads Microsoft Patches Record 622 Flaws, Including Two Zero-Days Under Active Attack Cursor Flaw Lets Malicious Cloned Repositories Trigger Windows Code Execution Researcher Drops New Windows Zero-Day PoC Hours After Microsoft Patch Tuesday TuxBot v3 Evolution Shows Signs of LLM-Assisted IoT Botnet Development Unpatched Shark Vacuum Flaw Could Let Attackers Control Other Vacuums Region-Wide New Agent Data Injection Attack Can Make AI Agents Misclick or Run Attacker Commands New ClickLock macOS Stealer Kills Apps Every 210ms Until Victims Type Their Password ThreatsDay: Game Cheat Spyware, 24-Hour Ransomware, Chrome Sync Stalking + 12 More Stories E.U. Orders Google to Open Android Mic, Camera and Screen to Rival AI Assistants OpenSSL HollowByte Flaw Could Freeze Server Memory with 11-Byte TLS Requests New wp2shell WordPress Core Flaw Lets Unauthenticated Attackers Run Code ⭐ Featured Resources What Security Teams Must Defend in the New AI Software Supply Chain Identity Fraud Is Changing Fast. See the Attacks Businesses Face in 2026 What 25 Million Alerts Reveal About the Threats SOCs Ignore How to Find and Control Every Script Running Through Your Marketing Stack Modern SASE Guide: Close the Gaps Traditional Network Security Cannot See

Indicators of Compromise

• cve — CVE-2026-58593

Entities