Key Takeaways
- Risk to your company: a zero-day lets an attacker into a browser, mailbox, firewall or VPN that looks fully up to date, and the first sign is often a fake payment request or a data leak, not a warning pop-up.
- A zero-day vulnerability is a software flaw that attackers exploit before the vendor has released a fix, so the defender has had zero days to patch.
- Zero-days are not rare: Google patched its eighth actively exploited Chrome zero-day of 2026 in November, and HKCERT logged 469 vulnerable-system incidents in Hong Kong in Q3 2026.
- You cannot patch a hole nobody knows about, but you can patch within hours once a fix appears, cut what is exposed to the internet, and add a second check before any money moves.
Your laptop says it is up to date. Your firewall shows a green light. And yet an attacker is reading your finance manager's inbox. That is the uncomfortable reality of a zero-day: a hole in software that nobody, including the vendor, has fixed yet.
What is a zero-day vulnerability?
A zero-day vulnerability is a security flaw in software or hardware that attackers discover and exploit before the vendor knows about it or has released a fix. The name means the defender has had zero days to prepare. A zero-day exploit is the attack code, and a zero-day attack is that code used against real victims.
Think of a building where the security guard checks every door on the list, but nobody knows about a broken window at the back. Until someone notices the window, the guard cannot protect it. A zero-day is that unknown window.
How does a zero-day attack work?
A zero-day attack usually follows five steps: an attacker finds an unknown flaw, builds code to exploit it, delivers it through a website, email or internet-facing device, gains access, then moves on to data or money. Because no patch exists, ordinary updates and antivirus signatures often do not stop the first step.
A typical Hong Kong scenario: a staff member at a trading company clicks a link in what looks like a supplier's shipping notice. The page quietly exploits an unpatched browser flaw, installs a small program, and steals the saved login for the company mailbox. Nothing crashes. Two weeks later a client receives an email, from the real company address, asking them to pay into a new bank account.
- Discovery: a criminal group, a researcher or, increasingly, an AI tool finds the flaw.
- Weaponising: the attacker writes an exploit, or buys one on the underground market.
- Delivery: a malicious web page, an email attachment, or a direct attack on a VPN, firewall or web server facing the internet.
- Foothold: the attacker runs code, steals passwords or session cookies, or creates a hidden admin account.
- Payoff: data theft, ransomware, cryptocurrency mining, or payment fraud using the stolen mailbox.
Hong Kong latest (updated November 2026)
Google patched an actively exploited Chrome flaw, CVE-2026-125350, in November 2026, its eighth Chrome zero-day of the year (The Hacker News, 16 Nov 2026). Microsoft's November update fixed an exploited Outlook flaw, CVE-2026-124194. Locally, HKCERT recorded 469 vulnerable-system incidents in Q3 2026, about 11% of 4,404 incidents (HKCERT, 5 Nov 2026). The window is also shrinking: Google Project Zero says it will publish flaw details 90 days after reporting them whether or not they are fixed, falling to 60 days from February 2027 (Project Zero, Nov 2026).
What types of zero-day vulnerabilities are there?
Zero-days are grouped by where the flaw sits. The ones that matter most to small and mid-sized companies are in web browsers, email and office software, operating systems, internet-facing devices such as VPNs and firewalls, and the web applications or plugins that run your website. Each needs a slightly different defence.
| Where the flaw sits | Typical example | Why it matters to an SME |
|---|---|---|
| Web browser | Chrome, Edge or Safari exploited by visiting a page | Every staff member uses one all day |
| Email and office apps | A crafted message or document leaks passwords | Mailboxes are the key to payment fraud |
| Operating system | A flaw that turns a normal user into an administrator | Lets one infected PC take over others |
| Network edge devices | VPN, firewall or router attacked directly from the internet | No click needed; often unmonitored |
| Website and plugins | A CMS plugin or PHP component on your web server | Customer data and your brand at risk |
Old equipment deserves special mention. VulnCheck found that 42.5% of exploited network-edge flaws it studied affected products at or near end of life, which will never receive a patch (VulnCheck, cited by HKCERT).
How can you tell if you have been hit by a zero-day?
You usually cannot see the flaw itself, but you can see what the attacker does next. Watch for unexpected logins, new mailbox forwarding rules, unknown admin accounts, devices suddenly running slow or hot, and colleagues or clients receiving strange payment requests from your real email address.
- Sign-in alerts from countries or devices nobody recognises.
- Emails disappearing because a hidden rule forwards or deletes them.
- A new administrator account on the firewall, VPN or Microsoft 365 / Google Workspace console.
- Antivirus or endpoint protection switched off without explanation.
How do you protect your business against zero-day attacks?
You cannot patch a flaw before the fix exists, so protection means shrinking the time between patch and install, reducing what is exposed to the internet, limiting what one stolen login can do, and adding a human check before money moves. Most of these steps cost little and can start this week.
For individuals and every staff member
- iPhone: Settings > General > Software Update > Automatic Updates > turn on both download and install options.
- Android (Pixel): Settings > System > Software updates > System update. Samsung: Settings > Software update > Download and install, and turn on auto download over Wi-Fi. Menu names vary by brand and version.
- Chrome on a computer: Menu (three dots) > Help > About Google Chrome, then click Relaunch. Do this whenever a Chrome zero-day is in the news.
- Windows: Settings > Windows Update > Advanced options > turn on "Get me up to date". Mac: System Settings > General > Software Update > the (i) next to Automatic Updates > turn all on.
- Paths last checked on 20 November 2026.
For SMEs and IT teams
- Keep a one-page list of everything facing the internet: VPN, firewall, router, website, remote desktop. Know who owns each one.
- Subscribe to HKCERT security bulletins and your vendors' advisories. Set a rule: "actively exploited" means patch within 24 to 48 hours.
- If no patch exists yet, apply the vendor's workaround, disable the affected feature, or take the device off the internet.
- Replace end-of-life routers, firewalls and VPNs. They will never get a fix.
- Use multi-factor authentication on email, VPN and admin consoles, and give staff only the access they need.
- Keep logs for at least 90 days and back up critical data offline, so you can investigate and recover.
When a stolen mailbox becomes a payment scam
Picture a 20-person accounting firm in Wan Chai. An exploited browser flaw gives an attacker the finance manager's mailbox. The attacker reads past invoices, then emails a regular client from the real address: "Our bank account has changed, please settle this month's fee here." Because the email is genuine, spam filters let it through.
Company rule: any change to bank details, or any urgent transfer request from a boss or supplier, must be confirmed by phone on a number already on file, never the number in the email.
Staff notice you can copy to your WhatsApp group
Security reminder: a new browser and email flaw is being exploited.
1. Please restart Chrome/Edge today and install any pending phone or PC updates.
2. Do not open unexpected attachments or links, even from known contacts.
3. Any request to change bank details or make an urgent payment: call the person on a known number first.
4. Seen a strange login alert? Tell IT now, do not ignore it.
60-second self-check
- Are automatic updates on for every phone, PC and browser?
- Do we know every device that faces the internet, and is any of it end of life?
- Is multi-factor authentication on for email, VPN and admin accounts?
- Is there a phone-verification rule for bank changes and urgent transfers?
- Did anyone independently test our website and VPN in the last 12 months?
What should you do if a zero-day is exploited on your systems?
Contain first, then investigate. Disconnect the affected device or account from the network, but do not wipe it, because the evidence is needed. Reset passwords and sign out all sessions, apply the patch or workaround, report to HKCERT, and check whether personal data was exposed and whether customers must be told.
- Isolate: unplug the PC, or block the VPN account. Keep the device powered on if possible.
- Preserve: save firewall, VPN, mailbox and server logs before they roll over.
- Reset: change passwords, revoke sessions and API keys, remove unknown forwarding rules and admin accounts.
- Report: HKCERT hotline 8105 6060 or its online incident form. If money was lost or a scam is involved, call the police Anti-Scam Helpline 18222 (24 hours) and check suspicious accounts with Scameter on CyberDefender.
- Notify: if personal data may have leaked, consider notifying the Privacy Commissioner (PCPD) and affected people. Operators designated under Cap. 653 have their own reporting deadlines.
- Learn: after recovery, have the exposed systems independently tested so the same door is not left open.
Zero-day vs known vulnerability: what is the difference?
A zero-day has no patch yet, so defence relies on reducing exposure and detecting misuse. A known vulnerability already has a patch, so the risk comes from not installing it. In practice most Hong Kong break-ins use known, unpatched flaws, which is why fast patching protects you from both.
| Zero-day | Known (n-day) vulnerability | |
|---|---|---|
| Patch available? | No | Yes |
| Who knows about it | Attacker, maybe a few researchers | Public, often with ready-made attack tools |
| Main defence | Limit exposure, MFA, monitoring | Install the update quickly |
| Typical targets | High-value or widely used software | Anyone who is slow to update |
The bottom line on zero-days
A zero-day is not a reason to panic, and it is not only a problem for big banks. It is a reminder that "fully updated" is a moment, not a state. Companies that know what they expose to the internet, patch exploited flaws within days, and verify every payment change by phone turn most zero-days into a near miss rather than a loss.
Find the open windows before attackers do
HKCERT recommends regular vulnerability assessments and penetration tests for internet-facing systems. A penetration test has certified testers attack your website, VPN and network the way a real intruder would, so you can fix weak points before a zero-day or a known flaw is used against you.
Learn about penetration testing
Frequently Asked Questions
Can antivirus software stop a zero-day attack?
Not reliably. Traditional antivirus looks for known patterns, and a zero-day is new by definition. Behaviour-based endpoint protection, multi-factor authentication and fast patching give much better odds.
Are small businesses really targeted with zero-days?
Rarely on purpose, but often by accident. Once a zero-day becomes public, attackers scan the whole internet for every unpatched device, regardless of company size. Small firms with old routers or VPNs are easy wins.
Does a zero-day attack have to be reported in Hong Kong?
There is no general duty to report every attack, but operators designated under Cap. 653 must report serious incidents, and data breaches should be considered for notification to the PCPD. HKCERT accepts voluntary reports from any business.
Why are they called zero-day?
Because the vendor has had zero days to fix the flaw when attackers start using it. The count starts on the day the vendor learns about the problem.