Most Wi-Fi networks are not hacked by breaking encryption. They are hacked through a short or guessable password, a router setting left on from the factory, firmware that has not been updated in years, or a fake access point that a phone joins on its own. Each of those has a specific fix, and most of the fixes take minutes.
This guide explains every major Wi-Fi attack the way an attacker sees it, so you can see exactly where your own network is exposed. For each one you get what it is, why it works, what the attacker needs, a real documented example, how to spot it, and how to shut it down. It ends with hardening checklists for a home, a small office and an enterprise, and with the part most guides skip: the router itself is a small Linux computer built from open source components, and its firmware is a software supply chain problem.
It is written for anyone who runs a network: a careful home user, the person who ended up looking after the office Wi-Fi, an IT manager, or a security engineer who wants one well-sourced reference. Every claim links to a primary source or a reputable report.
A note on the law. This is a defender's guide. It explains how attacks work conceptually and deliberately leaves out step-by-step procedures, commands and attack configurations. Testing a network you do not own, or do not have written permission to test, is illegal in most countries. If you want your own network tested, hire an authorized tester and put the scope in writing.
Contents
- How does Wi-Fi get hacked? The short answer
- How does Wi-Fi security work? WEP, WPA, WPA2 and WPA3
- What happens in the WPA2 4-way handshake?
- Where do Wi-Fi attacks actually land?
- How do attackers crack Wi-Fi passwords?
- Why is WPS a security risk?
- What is a deauthentication attack, and how does PMF stop it?
- What is an evil twin attack?
- What was the KRACK attack on WPA2?
- Is WPA3 secure? What Dragonblood showed
- What are FragAttacks?
- What is the SSID Confusion attack?
- How do attackers get into Wi-Fi routers?
- Is public Wi-Fi safe?
- How do enterprise Wi-Fi networks get compromised?
- Why are IoT devices on your Wi-Fi a risk?
- Which Wi-Fi risks should you fix first?
- How do you secure home Wi-Fi?
- How do you secure small office Wi-Fi?
- How do you secure enterprise Wi-Fi?
- How can you tell if your Wi-Fi has been hacked?
- Why is router firmware a software supply chain problem?
- How do you scan router and access point firmware for vulnerabilities?
- Fix it without waiting for upstream
- A fair counter-argument
- How to test your own Wi-Fi legally
- Frequently asked questions
How does Wi-Fi get hacked? The short answer
Wi-Fi usually gets hacked in one of four ways: guessing a weak password offline after recording a login exchange, abusing a router feature such as WPS or a default admin password, tricking a device into joining a fake access point, or exploiting old router firmware. Strong encryption is rarely broken directly. Configuration and patching decide most outcomes.
That distinction matters because it tells you where to spend effort. The cryptography in WPA2 and WPA3 is sound when it is implemented correctly and paired with a good passphrase. The well-known protocol flaws covered below, such as KRACK and FragAttacks, are real, but they were disclosed responsibly and fixed with updates. The attacks that succeed day to day are cheaper: a ten-digit phone number used as the password, WPS left on, a router admin page reachable from the internet with its factory password, or a laptop that happily rejoins any network called "Airport Free WiFi".
Think of a Wi-Fi network as five layers an attacker can target: the client device, the radio link in the air, the access point, the router behind it, and the internal network the router connects you to. Every attack in this guide hits at least one of them, and the attack map later in the guide shows which.
How does Wi-Fi security work? WEP, WPA, WPA2 and WPA3
Wi-Fi security has had four generations. WEP (1997) is broken and should never be used. WPA (2003) was a stopgap. WPA2 (2004) added AES encryption and is still widely deployed. WPA3 (2018) replaced the password exchange with SAE, which resists offline guessing, and requires Protected Management Frames. Use WPA3 where you can.
Wi-Fi security generations, and the published attacks that pushed each one forward.
WEP (Wired Equivalent Privacy). WEP arrived with the original 802.11 standard. It encrypted traffic with the RC4 stream cipher, using a key built by joining the shared secret to a 24-bit initialization vector. Twenty-four bits is far too small: the vector repeats quickly on a busy network, and repeated keystreams leak the key. In August 2001 Scott Fluhrer, Itsik Mantin and Adi Shamir published a cryptanalysis showing the key could be recovered from captured traffic. By 2007, researchers showed a 104-bit WEP key could be recovered with a 50 percent probability using only 40,000 captured packets. The IEEE formally deprecated both WEP key sizes in 2004.
WEP's real-world cost is well documented. The TJX intrusion, which by the end of March 2007 had affected 45.7 million customers' cards, was possible because of a non-secure wireless network in one of the company's stores, and the use of WEP contributed to the break-in. If any device on your network still offers WEP, treat it as an open network.
WPA (Wi-Fi Protected Access). WPA became available in 2003 as an interim fix that could run on WEP-era hardware. It used TKIP, which wrapped RC4 with per-packet keys and a message integrity check. TKIP is now deprecated and is not allowed in WPA3. Turn it off.
WPA2. WPA2 added CCMP, an AES-based encryption mode, and the 4-way handshake that most networks still use today. It comes in two flavors. WPA2-Personal, also called WPA2-PSK, uses one shared passphrase for everyone. WPA2-Enterprise uses 802.1X, where each user or device logs in with its own credentials through a RADIUS server. WPA2 with AES and a strong passphrase is still a reasonable baseline for older devices, but it has a structural weakness covered in the next section: anyone who records a login exchange can try passphrase guesses offline, as fast as their hardware allows.
WPA3. The Wi-Fi Alliance announced WPA3 in January 2018 and certification began that June. It has been required for Wi-Fi CERTIFIED devices since 1 July 2020. WPA3-Personal replaces the pre-shared key exchange with Simultaneous Authentication of Equals (SAE), a password-authenticated key exchange based on the Dragonfly handshake. SAE gives each session forward secrecy and is designed so that recording a login does not hand an attacker something to guess against offline; the Wi-Fi Alliance describes WPA3-Personal as giving "increased protections from password guessing attempts". WPA3-Enterprise offers an optional mode with equivalent 192-bit cryptographic strength. WPA3 also enforces Protected Management Frames (IEEE 802.11w), and the Wi-Fi Alliance states that PMF is required for all new certified devices.
Two more pieces round out the picture. Wi-Fi Enhanced Open, built on Opportunistic Wireless Encryption (RFC 8110), encrypts traffic on open networks such as cafes without a password. It does not authenticate the access point, so it stops passive eavesdropping but not a fake access point. And on the 6 GHz band used by Wi-Fi 6E and Wi-Fi 7, WPA3 is mandatory, and Cisco Meraki's configuration guide notes that WPA2 is not permitted there, Protected Management Frames are mandatory, and the hash-to-element (H2E) method for SAE is required because the older hunting-and-pecking method is prone to brute force dictionary attacks.
What happens in the WPA2 4-way handshake?
In the WPA2 4-way handshake, the access point and the client prove they share the same pairwise master key (PMK) without sending it. They swap random numbers, each derives a fresh session key called the PTK, and each proves it with a message integrity code. The weakness: a recorded handshake lets anyone test passphrase guesses offline.
The WPA2 4-way handshake. The key never crosses the air, but the messages are enough to check a password guess.
The handshake is easier to defend once you see its moving parts. The IEEE 802.11i amendment defines them.
The PMK. After authentication, both sides hold a shared secret called the Pairwise Master Key. In WPA2-Personal, the PMK is derived directly from the passphrase: the passphrase and the network name (SSID) go through PBKDF2 with 4,096 iterations of HMAC-SHA1 to produce a 256-bit key. The passphrase can be 8 to 63 printable characters. In WPA2-Enterprise, the PMK comes out of each user's 802.1X login instead, so there is no single shared passphrase to guess.
The four messages. The access point sends a random number called the ANonce. The client picks its own random SNonce, derives the session key, and replies with the SNonce plus a message integrity code (MIC) that proves it knows the PMK. The access point derives the same session key, checks the MIC, and sends the group key used for broadcast traffic (the GTK), encrypted, with its own MIC. The client confirms, and both sides install their keys.
The PTK. The Pairwise Transient Key is computed from the PMK, both random numbers and both devices' MAC addresses, run through a pseudo-random function, and split into separate keys for integrity and encryption. A new PTK is created for every session.
Why this matters to a defender. Look at what crosses the air in the first two messages: two random numbers, two MAC addresses, the SSID and a MIC computed from a key that depends only on the passphrase. Someone who records those messages can take a candidate passphrase, run it through the same derivation, and see whether the MIC matches. Nothing has to touch your network after the recording. That is the offline guessing problem, and it is the reason the single most important WPA2 setting is the passphrase itself. WPA3's SAE was designed to remove exactly this property.
Where do Wi-Fi attacks actually land?
Wi-Fi attacks land in five places: the client device, the radio link, the access point, the router, and the internal network behind it. Password guessing targets the radio link. Evil twins target the client. WPS and firmware bugs target the access point and router. IoT pivots target the internal network. Each place needs its own control.
The attack map. Most real compromises chain two zones, for example an evil twin on the client followed by credential reuse on the network.
A strong passphrase does nothing against a router with a remotely exploitable firmware bug, and a patched router does nothing for a laptop that joins a fake hotspot at an airport. Good Wi-Fi security means a control in every zone, and the checklists later in this guide are organized that way.
How do attackers crack Wi-Fi passwords?
Attackers crack Wi-Fi passwords offline. They record a WPA2 handshake, or a single PMKID value that some access points send without any client present, then test guesses on their own hardware until one matches. Short passwords, phone numbers and dictionary words fall quickly. A long random passphrase, or WPA3-SAE, makes this attack impractical.
What it is. Offline password guessing against WPA2-Personal. The attacker never needs to be connected to your network while guessing. They only need a recording that lets them check guesses.
Why it works. As the handshake section showed, the first two handshake messages contain everything needed to check whether a candidate passphrase is correct. In 2018 the hashcat developer known as atom described a second route on the hashcat forum. Many access points that support roaming features include a value called the PMKID in the first message they send to a device trying to associate. The PMKID is an HMAC-SHA1 value, truncated to 128 bits, computed from the PMK, the fixed label "PMK Name" and the two MAC addresses. Because it depends on the same PMK as the handshake, it is just as useful for checking guesses, and the attacker can request it from the access point directly. No legitimate user has to be present. The method applies to pre-shared key networks, not to 802.1X Enterprise logins.
What the attacker needs. Physical proximity for a few seconds or minutes, an inexpensive wireless adapter, and computing power for the guessing. Tools such as Aircrack-ng and hashcat are widely available and are used legitimately in authorized wireless assessments and password audits. GPUs make guessing fast, which is why passphrase length matters more than anything else on WPA2.
Real example. In 2021 a CyberArk researcher collected about 5,000 Wi-Fi network hashes while walking around Tel Aviv, using equipment that cost about $50, and recovered the passwords for more than 70 percent of them. About 2,200 fell because they were mobile phone numbers, a common local habit, and roughly 900 more fell to a list of previously leaked passwords. The researcher found that the success rate dropped as password length increased, and his advice was a password of at least 10 characters mixing lower case, upper case, digits and symbols. Security Affairs reported the same 70 percent figure.
Impact. Full access to the network, as if the attacker were a member of the household or office. From there they can reach every device that trusts the local network: file shares, printers, cameras, the router's admin page.
How to detect it. You usually cannot. Recording a handshake is passive, and requesting a PMKID looks like an ordinary association attempt. A burst of deauthentication frames, covered below, is a sign someone may be trying to force a fresh handshake. On business networks a wireless intrusion detection system can flag that pattern.
How to protect against it.
- Use WPA3-Personal (SAE) wherever every device supports it. SAE forces each guess to be made live against your access point.
- On WPA2, use a passphrase that is long and random. Four or more random words, or 16 or more random characters, puts it far beyond practical guessing. Do not use phone numbers, addresses, names or anything printed on the router label.
- Never reuse the Wi-Fi passphrase anywhere else.
- If you do not need roaming features on a home network, leave them off. The Tel Aviv researcher pointed out that only routers supporting roaming were exposed to the PMKID method, and that a personal network usually has no need for roaming.
- For offices, move to WPA3-Enterprise or per-user credentials, so there is no single shared passphrase to recover.
Why is WPS a security risk?
WPS is a risk because its eight-digit PIN can be guessed in about 11,000 attempts instead of 100 million. The router confirms each half of the PIN separately, and the last digit is a checksum. Some routers also generate weak random numbers, which lets the PIN be recovered offline. Turn WPS off entirely.
What it is. Wi-Fi Protected Setup was designed to let people connect devices without typing a passphrase, either by pressing a button or entering a PIN printed on the router. Once a device completes WPS, the router hands it the real network passphrase.
Why it works. In December 2011 Stefan Viehböck reported a design flaw, published by CERT/CC as VU#723755 and CVE-2011-5053. The protocol tells the attacker when the first half of the eight-digit PIN is correct, so each half can be guessed separately, and the last digit is a checksum. In CERT/CC's words, "The number of attempts goes from 10^8 to 10^4 + 10^3 which is 11,000 attempts in total." Many routers had no lockout after failed attempts. In 2014 Dominique Bongard described a second weakness, now called the Pixie Dust attack: some implementations use low-entropy pseudo-random number generators, which lets the PIN be worked out offline from a single exchange on known vulnerable devices.
What the attacker needs. Radio range, time, and a router with WPS PIN mode enabled. Public tools automate both attacks, which is exactly why the setting matters.
Impact. The attacker recovers the network passphrase no matter how long or random it is. CERT/CC also notes the attacker may be able to change the access point's configuration or cause a denial of service.
How to detect it. Some routers log repeated failed WPS attempts or temporarily lock WPS. Check your router's log for WPS events you did not initiate.
How to protect against it. CERT/CC's recommended workaround is simply to disable WPS. Turn off the PIN method and, ideally, the push-button method too, then check the router's status page to confirm WPS shows as disabled. If your router cannot disable WPS, replace it. If a device can only join through WPS, that device is the problem.
What is a deauthentication attack, and how does PMF stop it?
A deauthentication attack sends forged "disconnect" frames that appear to come from the access point. Without Protected Management Frames, devices obey them and drop off the network, which lets an attacker force a fresh handshake to record, or push users toward a fake access point. PMF (802.11w) authenticates those frames so forged ones are ignored.
Protected Management Frames turn a one-frame disconnect into a frame the client discards.
What it is. The 802.11 standard includes management frames that tell a device it has been disconnected. On older networks the protocol does not require any encryption for this frame, even when the session itself uses WEP, WPA or WPA2, and the attacker only needs the target's MAC address, which is visible over the air.
Why it works. Without integrity protection, a device cannot tell a genuine disconnect from a forged one.
What the attacker needs. Radio range and a wireless adapter that can transmit raw frames.
Impact. By itself, a denial of service: devices keep dropping off. As part of a chain, it is more serious. Each forced reconnect produces a new handshake to record for offline guessing, and a client that loses its real network may jump to an evil twin with a stronger signal.
Real example. Deauthentication is not only an attacker technique. In 2014 Marriott agreed to pay a $600,000 fine to the US communications regulator, the FCC, after an investigation found that staff at a Nashville resort and convention center had used "containment features of a Wi-Fi monitoring system" to stop guests using their own personal hotspots, while charging exhibitors up to $1,000 per device for the hotel's Wi-Fi. The lesson for defenders runs both ways: containment features are powerful, and pointing them at networks you do not own can be unlawful.
How to detect it. Wireless intrusion detection systems flag bursts of deauthentication or disassociation frames. At home, the symptom is devices repeatedly disconnecting and reconnecting with a strong signal.
How to protect against it. Turn on Protected Management Frames. IEEE 802.11w adds integrity, data origin authenticity and replay protection to management frames. WPA3 requires PMF, and many WPA2 routers offer it as an "optional" or "capable" setting that protects devices which support it without locking out those that do not. PMF does not stop radio jamming, which is a physical layer problem, but it removes the cheap one-frame disconnect.
What is an evil twin attack?
An evil twin is a fake access point that copies a trusted network's name, often with a stronger signal, so phones and laptops join it automatically. The attacker then sits in the middle of the victim's traffic and often shows a fake sign-in page asking for email or social media logins. A rogue access point is the inside version.
An evil twin needs no password cracking at all. It relies on devices trusting a network name.
What it is. An evil twin is a fraudulent access point that appears legitimate and is set up to eavesdrop. A common variant pairs it with a captive portal: the fake network shows a sign-in page that looks like the venue's, and whatever the victim types goes to the attacker. A related technique, the KARMA attack published in 2004 by Dino dai Zovi and Shane Macaulay, answers the network names devices broadcast while looking for networks they have joined before, so the device connects to a network it thinks it knows.
A rogue access point is the inside version: an unauthorized access point plugged into your wired network, whether by an attacker or by an employee who wanted better coverage. Either way it creates a wireless door that bypasses your controls.
Why it works. Open networks and many captive portal networks do not authenticate the access point at all, so a device has nothing to check except the name. Phones and laptops are built to rejoin known names automatically.
What the attacker needs. A laptop or a small portable access point, a network name worth copying, and a place where people expect free Wi-Fi.
Real example. In 2024 Australian police charged a Western Australian man with setting up "evil twin" free Wi-Fi networks at airports in Perth, Melbourne and Adelaide and on domestic flights. People who connected were sent to a fake page asking them to sign in with email or social media accounts, and the details were saved to his devices. The investigation began when an airline reported a suspicious network during a flight. In late 2025 he was sentenced to prison, with at least five years to serve. The equipment was a Wi-Fi Pineapple, a device built for authorized wireless auditing.
Impact. Stolen credentials, exposure of any traffic that is not encrypted end to end, and malicious redirects. For a business, one stolen email password from an employee's phone can be the start of a much larger incident.
How to detect it. Two networks with the same name where you expect one, a network that suddenly asks for an email or social login to get online, or a certificate warning in your browser are all red flags. Business wireless intrusion detection systems compare broadcast networks against an approved list and alert on lookalike names and unknown access points.
How to protect against it.
- On phones and laptops, forget open networks you no longer use and turn off auto-join for public networks.
- Never enter account passwords into a Wi-Fi sign-in page. Real venue portals ask you to accept terms, not to log in with your email.
- Use a VPN on any network you do not control, and keep HTTPS-only browsing on.
- At work, use 802.1X with server certificate validation (see the Enterprise section) so devices can tell the real network from a copy.
- Run rogue access point detection, and use 802.1X on wired switch ports so an unauthorized access point cannot get a network connection in the first place.
What was the KRACK attack on WPA2?
KRACK (Key Reinstallation Attacks) was a set of 2017 flaws in the WPA2 standard itself. By replaying handshake messages, an attacker in range could make a device reinstall a key already in use, resetting counters so traffic could be decrypted or, in some cases, forged. It was fixed by software patches. WPA2 did not need to be replaced.
What it is. In October 2017 Mathy Vanhoef of KU Leuven published KRACK, a family of attacks against the 4-way handshake and several related handshakes. They are tracked as ten CVEs, CVE-2017-13077 through CVE-2017-13088, covering pairwise, group and integrity key reinstallation in the 4-way handshake, the group key handshake, Fast BSS Transition (802.11r), PeerKey, TDLS and WNM sleep mode.
Why it works. The weakness was in the standard's description of when to install keys, so any correct implementation was likely affected. When a key is reinstalled, the counters that ensure each encryption nonce is used once are reset, and nonce reuse breaks the encryption's guarantees.
What the attacker needs. Radio range and the ability to sit between the client and access point at the radio level. No password is needed.
Impact. Decryption of traffic, and on some setups injection or modification of data. The worst case affected wpa_supplicant 2.4 and later, which includes Android 6.0 and later at the time: the KRACK site explains that these clients could be tricked into installing an all-zero encryption key, making interception trivial.
How to detect it. Practically, you check versions rather than traffic. Vendors published advisories listing fixed firmware and operating system releases.
How to protect against it. Patch both sides. The researchers stated plainly that "both the client and AP must be patched to defend against all attacks." Home routers were mainly exposed if they used client or repeater mode or 802.11r. Changing the Wi-Fi password does not help, and falling back to WEP makes things worse. Any device that has not received updates since 2017, especially IoT devices and old phones, should be assumed vulnerable and kept off networks that carry anything sensitive.
Is WPA3 secure? What Dragonblood showed
WPA3 is the most secure option for most networks, but its first implementations had flaws. In 2019 the Dragonblood research showed downgrade attacks against WPA3 transition mode, side channels that leaked password information, and denial-of-service weaknesses. Vendors patched them, and newer devices support a stronger hash-to-element method. Keep WPA3 devices updated.
What it is. In April 2019 Mathy Vanhoef and Eyal Ronen published Dragonblood, an analysis of WPA3's Dragonfly handshake. The attacks targeted WPA3-Personal.
Why it works. Three classes of weakness were found:
- Downgrade. In WPA3 transition mode, a network accepts both WPA2 and WPA3 so older devices can still join. An attacker can push a WPA3-capable client onto WPA2, where offline guessing works again, or onto a weaker cryptographic group.
- Side channels. The researchers showed that the time an access point takes to respond, and memory access patterns during the handshake, could leak information about the password. These are tracked as CVE-2019-9494 for SAE and CVE-2019-9495 for the related EAP-pwd method in hostapd and wpa_supplicant, with further CVEs (CVE-2019-9497 through CVE-2019-9499, plus CVE-2019-11234 and CVE-2019-11235 for FreeRADIUS) covering reflection and invalid curve flaws in EAP-pwd implementations. The researchers estimated that the leaked information could support a brute force attack against the largest known dictionaries for less than $1 of cloud computing.
- Denial of service. An access point could be overloaded by as few as 16 forged commit frames per second.
What the attacker needs. Radio range, and for the side channel attacks either precise timing or the ability to observe the target's memory behavior. They are harder to pull off than WPA2 offline guessing.
Impact. Potential password recovery on unpatched implementations, or denial of service.
How to detect it. Again, by version. Check that access points and clients run firmware released after the Dragonblood fixes.
How to protect against it. Update access points and clients. Use WPA3-only mode once all your devices support it, and treat transition mode as temporary. On 6 GHz networks, WPA3 with hash-to-element is required and WPA2 is not permitted, which removes the downgrade path on that band. The practical takeaway: WPA3 is a clear improvement, and its early bugs are a reason to keep firmware current, not a reason to stay on WPA2.
What are FragAttacks?
FragAttacks are a set of Wi-Fi vulnerabilities disclosed in 2021 that affect how devices split and reassemble frames. Three are design flaws dating back to Wi-Fi's 1997 release, and more are common implementation bugs. Testing found every Wi-Fi product was affected by at least one. Updates fix them, and HTTPS limits the damage.
What it is. FragAttacks (fragmentation and aggregation attacks), published by Mathy Vanhoef on 11 May 2021, affect all modern Wi-Fi security protocols including WPA3, and even WEP.
Why it works. Three design flaws sit in the standard: the aggregation attack (CVE-2020-24588), because the flag that marks a frame as aggregated is not authenticated; the mixed key attack (CVE-2020-24587); and the fragment cache attack (CVE-2020-24586). Several of these design flaws have been part of Wi-Fi since its release in 1997. Alongside them are implementation flaws, such as devices accepting plaintext frames on an encrypted network (for example CVE-2020-26140 and CVE-2020-26145), tracked as CVE-2020-26139 through CVE-2020-26147.
What the attacker needs. Radio range and, for the injection attacks, a way to get the victim to contact an attacker-controlled server, which the researcher notes can be as simple as making the victim load an image from it.
Impact. Injection of frames into a protected network and, in some cases, extraction of data. A documented example is tricking a client into using a malicious DNS server. The researcher noted that the design flaws are hard to abuse and that "the biggest concern are the programming mistakes in Wi-Fi products." He also wrote that "every Wi-Fi product is affected by at least one vulnerability." He was not aware of any exploitation in the wild.
How to detect it. By version, through vendor advisories.
How to protect against it. Install updates on routers, access points, laptops and phones. Prefer HTTPS everywhere. For devices that will never be patched, which is common for IoT gear, isolate them on their own network segment so a compromise stays contained.
What is the SSID Confusion attack?
SSID Confusion is a 2024 design flaw, CVE-2023-52424, that can trick a Wi-Fi client into connecting to a different network than the one it shows. It works when two networks share credentials, such as separate 2.4 GHz and 5 GHz names with one password. It can also switch off VPN apps that trust specific network names.
What it is. Héloïse Gollier and Mathy Vanhoef presented the SSID Confusion attack at ACM WiSec in May 2024. In their words, "a client can be tricked into connecting to a different protected Wi-Fi network than the one it intended to connect to," and "the root cause is a design flaw in the IEEE 802.11 standard, causing the SSID to not always be authenticated." The vulnerability was assigned CVE-2023-52424, and all tested devices were vulnerable.
Why it works. Passwords or other credentials are verified in both directions when you connect, but the network name is not guaranteed to be part of that check. If two networks accept the same credentials, an attacker in the middle can make the client think it joined one while it actually joined the other.
What the attacker needs. Two networks in range that use the same credentials, and a position between the client and access point at the radio level.
Impact. The paper describes downgrading clients from a better-secured network to a weaker one, for example from a 5 GHz network to a 2.4 GHz network with a different name but the same password. It also shows that some VPN apps, which can be set to turn off automatically on "trusted" networks recognized by name, may switch themselves off, exposing traffic.
How to detect it. Hard to see directly. The practical signal is configuration: any place where two network names share one password or one RADIUS server identity.
How to protect against it. Use a different password for each network name at home, and on enterprise networks give each network a distinct RADIUS server certificate name. The researchers also propose beacon protection and adding the SSID to the handshake; from Wi-Fi 7 onward, access points must support beacon protection. Do not rely on a VPN's "trusted network" auto-disable as a security control.
How do attackers get into Wi-Fi routers?
Attackers get into routers through default or weak admin passwords, admin panels exposed to the internet, UPnP features that let anything on the network open ports, and known vulnerabilities in old firmware. Botnets such as Mirai scan the internet for these automatically. Keeping firmware updated and turning off remote admin and UPnP closes most of the door.
The router is the most valuable target on a home or small office network. It sees every packet, controls DNS for every device, and often runs for years without anyone logging in. Four problems account for most router compromises.
1. Default and hard-coded credentials. The Mirai botnet identified vulnerable devices using a table of more than 60 common factory default usernames and passwords and logged in over Telnet. A USENIX Security study traced its growth to a peak of 600,000 infections, made up mainly of embedded and IoT devices including home routers. In October 2016 Mirai-infected devices took part in DDoS attacks on the DNS provider Dyn that made sites including GitHub, Twitter, Reddit and Netflix unreachable for many users. Hard-coded credentials are worse than defaults, because the owner cannot change them. The Fraunhofer FKIE router study found hard-coded login credentials in the firmware of 50 of 127 home routers, and well-known or easily cracked ones in 16.
2. Exposed admin panels. A router's web interface should only be reachable from inside the network. When remote administration is turned on, any vulnerability in that interface becomes reachable from the whole internet. CVE-2023-1389 is a clear example. It is an unauthenticated command injection in the web management interface of the TP-Link Archer AX21 (AX1800), in firmware before 1.1.4 Build 20230219, that lets an attacker gain root access with a single crafted request. It has a CVSS base score of 8.8, and FortiGuard reported that it was exploited in the wild to deploy Mirai.
3. UPnP. Universal Plug and Play lets devices on your network ask the router to open ports automatically, so game consoles and video apps work without manual setup. It also lets malware do the same. In 2018 Akamai found that 277,000 of 3.5 million UPnP devices it surveyed ran vulnerable implementations, and confirmed more than 45,000 had been compromised in a campaign that injected port forwarding rules to expose Windows file sharing ports (TCP 139 and 445) on machines behind the router. Those 45,113 routers exposed 1.7 million machines. Separately, CallStranger (CVE-2020-12695), disclosed in 2020, showed that the UPnP SUBSCRIBE feature could be abused for data exfiltration and reflected, amplified TCP DDoS across a protocol used by billions of devices.
4. Old firmware. In May 2018 Cisco Talos reported VPNFilter, modular malware that had infected at least 500,000 devices in at least 54 countries, including Linksys, MikroTik, NETGEAR and TP-Link routers and QNAP storage devices. It could steal website credentials, monitor industrial control traffic and render a device unusable, and its first stage survived a reboot. Talos noted that all of the affected models it had found had well-known, public vulnerabilities, and many were reachable through known exploits or default credentials.
What the attacker needs. For most of these, nothing but an internet connection and a scanner. Router compromise is automated and indiscriminate. Botnets do not care whose router it is.
Impact. A compromised router can redirect your DNS to phishing sites, intercept or modify unencrypted traffic, expose devices behind it, join DDoS attacks, and serve as a persistent foothold that survives password changes on every other device.
How to detect it. Warning signs include DNS settings you did not set, unfamiliar port forwarding or UPnP rules, admin accounts you did not create, remote management switched on, unexplained reboots or sluggishness, and outbound traffic at odd hours. Check the router's admin page every few months, and look at the list of connected devices for anything you cannot identify.
How to protect against it.
- Change the admin password to a long unique one, and change the admin username if the router allows it.
- Turn off remote administration, Telnet and SSH on the WAN side. If you must manage remotely, use a VPN into the network instead.
- Turn off UPnP unless something specific needs it, then reboot to clear any rules that were already injected. Akamai notes that turning UPnP off may not remove existing injected rules.
- Turn on automatic firmware updates, or check monthly.
- Replace routers that no longer receive updates. A router past its end of support is a permanently unpatched Linux server on your network edge.
- After a suspected compromise, factory reset, update firmware before reconnecting to the internet, and set a new admin password. Talos gave the same advice for VPNFilter.
Is public Wi-Fi safe?
Public Wi-Fi is reasonably safe for browsing HTTPS sites on an up-to-date device, but it carries real risks: fake hotspots, fake sign-in pages, interception of unencrypted traffic, and other users on the same network. Use a VPN, keep HTTPS-only mode on, turn off auto-join, and never enter passwords into a Wi-Fi sign-in page.
Public Wi-Fi is safer than it was ten years ago, mainly because most websites now encrypt traffic end to end with HTTPS, a shift Google tracks in its HTTPS transparency report. Someone sniffing a cafe network today sees which sites you visit, not the content of your banking session. That is real progress, and it is why panic about public Wi-Fi is often overstated.
The remaining risks are specific:
- Fake hotspots and portals. The evil twin section above describes the main one. The Australian airport case shows it is not theoretical.
- Unencrypted traffic. Older apps, some IoT companion apps and plain HTTP pages still send data in the clear.
- Other users on the same network. Many public networks rely on client isolation to stop customers reaching each other's devices. In 2023 Vanhoef and colleagues published MacStealer (CVE-2022-47522), a technique that can bypass client isolation and intercept traffic meant for other clients at the MAC layer. It needs a malicious insider with credentials on the same network, which on a hotel or conference network can be anyone.
- Downgrade and protocol flaws. KRACK, FragAttacks and SSID Confusion all matter more on networks you do not control and cannot patch.
How to protect yourself on public Wi-Fi. Keep your device updated. Use a reputable VPN on any network you do not run. Keep your browser in HTTPS-only mode. Turn off auto-join for public networks and forget them after use. Turn off file sharing and AirDrop-style discovery for everyone. Use your phone's hotspot for anything sensitive. If you operate a public network, prefer Wi-Fi Enhanced Open so traffic is encrypted even without a password, and enable client isolation.
How do enterprise Wi-Fi networks get compromised?
Enterprise Wi-Fi is most often compromised through clients that do not validate the RADIUS server's certificate. A fake access point then collects login exchanges, and the widely used MSCHAPv2 method can be cracked. Fix it with EAP-TLS certificates where possible, strict server certificate validation pushed by device management, rogue access point detection and PMF.
What it is. WPA2-Enterprise and WPA3-Enterprise use 802.1X, where a RADIUS server checks each user's credentials. The most common method, PEAP with MSCHAPv2, builds a TLS tunnel to the server and sends the user's login inside it. The security of the whole design depends on one check: the client must verify that the server certificate is the real one.
Why it works. If the client does not validate the CA certificate, in general it is trivial to introduce a fake wireless access point that then gathers MS-CHAPv2 handshakes. MSCHAPv2 itself is weak. As of 2012, MS-CHAP had been completely broken: its security reduces to DES operations keyed by the NTLM password hash, and cracking a single DES key is not difficult with modern hardware. A captured exchange can therefore lead to the user's domain password, or to a hash that can be used directly.
Misconfiguration is common for ordinary reasons. Users get a "do you trust this certificate?" prompt and tap yes. Bring-your-own devices are set up by hand without a pinned server name. A wildcard or public certificate is used on the RADIUS server, so any certificate from the same public authority looks valid.
What the attacker needs. A fake access point broadcasting the corporate network name, a fake RADIUS server, and proximity to employees, for example in a car park or a nearby cafe.
Impact. Corporate credentials that often unlock email, VPN and cloud applications as well as Wi-Fi.
How to detect it. Wireless intrusion detection that alerts on your corporate network name broadcast by unknown access points. RADIUS and identity logs that show logins from unusual locations shortly after a user was near an unknown network.
How to protect against it.
- Prefer EAP-TLS, where both sides authenticate with certificates and there is no password to steal. Microsoft recommends moving away from MSCHAPv2-based connections to certificate-based authentication such as PEAP-TLS or EAP-TLS for Wi-Fi and VPN, and notes that Credential Guard on Windows does not work with MSCHAPv2 configurations.
- Push Wi-Fi profiles through device management with the RADIUS server name and the issuing CA pinned, and with user prompts to trust new certificates disabled.
- Use a private CA for the RADIUS server certificate rather than a public one.
- Give each network name its own RADIUS server identity, which also blocks the SSID Confusion credential reuse case.
- Require PMF, and use WPA3-Enterprise where your clients support it.
- Run wireless intrusion detection with rogue and lookalike access point alerts.
Why are IoT devices on your Wi-Fi a risk?
IoT devices such as cameras, smart TVs and plugs are risky because they often run old software, rarely get updates, and sit on the same network as your laptops. Once one is compromised, an attacker can use it to reach everything else. Put IoT devices on a separate network or VLAN that cannot reach your main devices.
Why it works. Many connected devices ship with old Linux kernels and libraries, get firmware updates for only a short time, and have no screen to warn anyone when something is wrong. They are also always on. Mirai's victims were mostly cameras and routers. FragAttacks' author singled out smart and IoT devices as the ones that are hard to keep patched.
Real example. In 2018 Darktrace's chief executive described an investigation into an unnamed casino that was breached through an internet-connected thermometer in a lobby fish tank. The attackers reportedly used it as a foothold to find the casino's high-roller database and pull data out. The thermometer was never the target. It was a way in.
Impact. Lateral movement to laptops, file shares and the router, and use of the device itself for spying (cameras, microphones) or for botnet traffic.
How to detect it. Watch for IoT devices talking to unusual destinations or to other devices on the local network. Many business firewalls and some consumer routers show per-device traffic.
How to protect against it. Segment the network. Put IoT devices on a separate SSID or VLAN that can reach the internet and, where needed, its own controller, but nothing else. Change default passwords on every device, update firmware where updates exist, and retire devices that no longer receive updates.
Segmentation keeps one weak device from becoming a path to everything else.
Which Wi-Fi risks should you fix first?
Fix the cheap, high-impact problems first: default router passwords, outdated router firmware, weak Wi-Fi passphrases, WPS, and exposed admin panels or UPnP. Then plan the larger work, such as network segmentation and enterprise certificate validation. Protocol-level flaws like FragAttacks and Dragonblood are handled mainly by keeping devices updated.
A qualitative view for a typical home or small office. Your own ordering depends on your devices and how the network is used.
The matrix is a judgment, not a measurement. Default credentials and old firmware sit at the top because they are exploited automatically at internet scale, as Mirai, VPNFilter and CVE-2023-1389 show, and fixing them costs little. Weak passphrases and WPS are high risk because the attacks are cheap and well tooled, and the fixes take minutes. Segmentation and enterprise certificate validation are high value but take planning, new hardware or device management work. The named protocol flaws are serious in principle, yet each was disclosed responsibly with patches available, so the main job is keeping firmware current.
How do you secure home Wi-Fi?
To secure home Wi-Fi, use WPA3 (or WPA2 with AES) with a long random passphrase, turn off WPS, enable PMF, change the router admin password, turn off remote admin and UPnP, keep firmware updated automatically, and put guests and smart devices on separate networks. Replace routers that no longer get updates.
Each column builds on the one before it.
Here is the home checklist in full.
- Encryption: WPA3-Personal, or WPA2/WPA3 mixed if an older device needs it, otherwise WPA2 with AES only. Never WEP, never TKIP, never open.
- Passphrase: four or more random words, or 16 or more random characters. This one setting decides whether offline guessing ever succeeds. A password manager can generate and store it.
- WPS: off. Both PIN and push-button if your router allows it.
- PMF: on. Choose "required" if every device supports it, otherwise "optional" or "capable".
- Admin password: changed to something long and unique. If the router has a companion app with a cloud account, protect that account with a strong password and two-step verification.
- Remote administration: off. Telnet and SSH off on the WAN side.
- UPnP: off unless a specific device needs it.
- Firmware: automatic updates on. Check the vendor's support page for the end-of-support date of your model.
- Guest network: on for visitors, with access to the internet only.
- IoT network: separate from your laptops and phones, using a second SSID or a guest network if the router cannot do VLANs.
- Network names: unique per band, each with its own password, or a single name for both bands. This avoids the SSID Confusion credential reuse case.
- Inventory: know what is connected. Review the connected device list every few months and remove anything you cannot identify.
How do you secure small office Wi-Fi?
Secure small office Wi-Fi with business-grade access points that get regular updates, WPA3-Enterprise or per-user credentials instead of one shared password, separate VLANs for staff, guests, IoT and management, an admin interface reachable only from the management network, and a written owner and patch schedule.
A small office adds these steps to the home list:
- Business-grade hardware with a published support lifecycle. Consumer routers often have short update windows.
- No single shared passphrase for staff. Use WPA3-Enterprise with a hosted RADIUS service, or an access point feature that gives each user or device its own key, so one departing employee does not mean changing the password for everyone.
- VLANs: staff, guest, IoT (printers, cameras, door controllers) and management, with firewall rules between them.
- Management interfaces only on the management VLAN. The access points' and switches' admin pages should not be reachable from the guest or staff networks.
- Firmware inventory. Keep a list of every access point, router, switch and firewall with its model and firmware version, and check for updates monthly.
- Legacy features off: WEP, TKIP and old data rates you do not need.
- Alert on new networks nearby that copy your network name. Many business access point platforms include basic rogue detection.
- Laptops: VPN on untrusted networks and auto-join off for public networks.
- An owner. Someone is responsible for the network, by name, with a calendar reminder to review it.
How do you secure enterprise Wi-Fi?
Secure enterprise Wi-Fi with WPA3-Enterprise and EAP-TLS certificates, clients that strictly validate the RADIUS server certificate through device management, PMF on every network, wireless intrusion detection with rogue access point detection, 802.1X on wired ports, and firmware for every access point and controller tracked like any other software with an SBOM.
The enterprise list builds on the small office list:
- WPA3-Enterprise with EAP-TLS wherever clients support it, with the 192-bit mode for high-security networks.
- Strict server certificate validation pushed through device management: pinned server name, pinned private CA, no user prompts.
- Retire PEAP-MSCHAPv2 where possible, following Microsoft's guidance above.
- PMF required on every network name.
- A wireless intrusion prevention system. A WIPS monitors the radio spectrum for unauthorized access points and can take countermeasures. Use containment only against devices on your own premises and network, and take legal advice before turning on automatic containment; the Marriott case shows why.
- 802.1X on wired switch ports so a rogue access point plugged into a wall socket gets no network.
- Firmware as software. Every access point, controller, router and firewall runs a firmware image built from open source components. Keep an SBOM for each, scan it for vulnerabilities, and set patch deadlines by exploitability rather than by CVSS score alone. The next sections cover how.
- Authorized testing at least once a year, with a written scope, covering rogue access points, evil twin resistance of managed devices, and segmentation.
How can you tell if your Wi-Fi has been hacked?
Signs your Wi-Fi or router has been hacked include unknown devices in the connected list, DNS or port forwarding settings you did not make, a changed admin password, remote management switched on, frequent disconnects, and browsers landing on unexpected pages. If you see these, factory reset the router, update its firmware and set new passwords.
At home, check four things on the router's admin page: the list of connected devices, the DNS servers, the port forwarding and UPnP rules, and whether remote management is on. Anything you did not set deserves attention. Repeated disconnections with a strong signal can point to deauthentication. A browser that shows certificate warnings for well-known sites, or search pages that look slightly wrong, can point to DNS tampering.
In a business, detection is a combination of wireless intrusion detection (deauthentication floods, unknown access points, lookalike network names), RADIUS logs (failed and unusual logins), DHCP logs (unknown devices), and firewall logs (IoT devices talking to unexpected destinations). The most useful habit is a baseline: know which access points, network names and device types are normal, so anomalies stand out.
If you find signs of compromise, disconnect the router from the internet, factory reset it, install the latest firmware before reconnecting, set a new admin password and a new Wi-Fi passphrase, and then change passwords for important accounts used on the network. If the router no longer receives updates, replace it.
Why is router firmware a software supply chain problem?
Router firmware is a software supply chain problem because a router is a small Linux system assembled from open source parts: the Linux kernel, BusyBox, a TLS library such as OpenSSL, dnsmasq, hostapd and a UPnP daemon. Each has its own vulnerabilities, and fixes must pass through chip vendors, device makers and sometimes internet providers before reaching you.
A router firmware image is a stack of components, most of them open source, each with its own advisories.
Every attack in the router section traced back to a software component. CVE-2023-1389 lived in a vendor web interface. CallStranger and UPnProxy lived in UPnP implementations. The Dragonblood EAP-pwd vulnerabilities lived in hostapd and wpa_supplicant, the open source projects that implement Wi-Fi authentication on most Linux devices. KRACK fixes shipped as updates to those same projects.
The supporting components carry their own history. In January 2021 researchers at JSOF disclosed DNSpooq, seven vulnerabilities in dnsmasq (CVE-2020-25681 through CVE-2020-25687), the DNS and DHCP server used in many routers. Three allowed DNS cache poisoning and four were buffer overflows. Versions 2.78 to 2.82 were affected, the fix was version 2.83, and OpenWrt firmware, which runs on hundreds of router models, was among those affected. In 2021 Claroty's Team82 and JFrog found 14 vulnerabilities in BusyBox, fixed in version 1.34.0. When they checked more than 10,000 embedded firmware images, 40 percent contained a BusyBox executable linked with one of the affected applets.
The patching gap is measurable. The Fraunhofer FKIE study of 127 home routers from seven large vendors, using firmware available on 27 March 2020, found that 116 ran Linux and that 46 had received no security update in the previous year. More than a third used Linux kernel 2.6.36 or older, whose last security update was on 17 February 2011, nine years earlier. The researchers' summary was blunt: "There is no router without flaws," and "even if the routers got recent updates, many of these known vulnerabilities were not fixed."
What 127 home routers looked like inside in 2020. The pattern, old components inside recently shipped images, is the supply chain problem in one chart.
Why is it so slow? A typical router's software passes through several hands. The Wi-Fi chipset vendor supplies a board support package with a kernel and drivers. A device manufacturer builds the product on that package. A brand may resell it, and an internet provider may customize and distribute it to subscribers. A fix in an upstream open source project has to move through each of those steps, and any of them can stall, especially for older models. The kernel version in particular is often pinned to whatever the chipset vendor's package supports.
This is exactly the problem software composition analysis was built for in application code: know every component and version, match them to known vulnerabilities, decide which ones matter, and get fixes moving. The difference is that firmware arrives as a binary image rather than a repository with a lockfile, so the inventory has to be recovered from the image itself.
For device makers, internet providers and enterprises that buy network equipment at scale, that inventory is now a basic requirement. If you ship routers or access points, your customers increasingly ask for an SBOM. If you buy them, you need to know which of the hundreds of components in your fleet are exposed when the next dnsmasq or BusyBox advisory lands.
How do you scan router and access point firmware for vulnerabilities?
To scan router or access point firmware, extract the filesystem from the image if it is packed, generate an SBOM listing every component and version, match those components to known vulnerabilities, rank them by exploitability and reachability, then fix what you build and pressure vendors on what you buy. Safeguard does this from a binary upload.
Treating Wi-Fi firmware like any other software: inventory, ranked vulnerabilities, then fixes.
Here is how that works in Safeguard, step by step.
1. Get the image. For devices you build, use the output of your firmware build. For devices you buy, download the vendor's published firmware file for the exact model and version you run.
2. Extract it if needed. Many firmware files are packed or compressed. Open source tools such as binwalk can identify and extract the embedded filesystem, so the scanner sees the real files and binaries rather than one opaque blob.
3. Upload it. In the Safeguard console, binaries and images are uploaded under Integrations > Upload App Binary. The same route accepts virtual machine and disk images and container tarballs, which covers virtual appliances such as wireless controllers and the containers that run network management software.
4. Get an SBOM. Safeguard generates an SBOM for the image in SPDX or CycloneDX, listing components such as the kernel, BusyBox, the TLS library, dnsmasq and hostapd with their versions. Keep it next to the firmware release it describes, so every version in your fleet or product line has a matching inventory.
5. See the vulnerabilities. Components are matched to known CVEs and enriched with EPSS exploit-probability scores and known-exploited data, so a dnsmasq version in the DNSpooq range or a router web interface bug that botnets are using rises to the top.
6. Rank by reachability. Reachability analysis asks whether the vulnerable code is actually called in your software, and it typically cuts the noise by roughly 60 to 80 percent. For a firmware team, that is the difference between a list of hundreds of CVEs and a short list of the ones that matter in this build.
7. Act on it. For firmware you build, Safeguard's Griffin models can open a fix pull request that upgrades or patches the vulnerable component, and your own CI and hardware tests decide whether it merges. You choose which remediation categories run automatically and which stay manual. For equipment you buy, the Third-Party Risk Manager lets you request SBOMs from vendors, score and monitor them, and apply the same policies to their components as to yours. Policies and gates can stop a release that contains a reachable critical vulnerability.
The same platform covers the firmware's source repositories, companion apps, secrets in the build system, open source license compliance, and SLSA, Sigstore and in-toto attestations for what you ship.
Safeguard runs as SaaS, in your own VPC, or fully on premises. Its AI features can use Safeguard's own security-tuned models or your own model, or run in Zero AI mode with no LLM in the loop. Teams that will not grant repository access can use the local runner CLI to scan on their own machines and send only encrypted results.
Fix it without waiting for upstream
Spotting a vulnerable dnsmasq or BusyBox in a firmware image is half the job. The harder half is getting a fixed build when the upstream project, the chipset vendor's package, or both have not caught up.
Safeguard Gold closes that gap. Gold rebuilds patched versions of open source packages and images from source, signs each build, attaches an SBOM, SLSA provenance and VEX, and serves them from registry.safeguard.sh. When an upstream project has fixed only its newest major version while your firmware is pinned to an older branch, or when a fix exists upstream but has not reached the package your toolchain uses, Gold gives you a patched build to depend on. When upstream will not move at all, Safeguard can fork and patch the package so you still have a fixed component.
For a router or access point maker, that means a reachable vulnerability in a core component does not have to hold up a firmware release while you wait on someone else's schedule. Pair it with the fix pull requests above and the loop closes: the vulnerability is found, ranked, patched with a signed component, and tested in your own CI.
The free Gold Open Source directory at gold.safeguard.sh lets anyone look up packages, known vulnerabilities and safe versions without logging in, which is a quick way to check whether the dnsmasq or BusyBox version in your router is affected.
A fair counter-argument
A reasonable objection: most Wi-Fi risk is configuration, not code, so why talk about SBOMs at all? For a household, that is mostly right. Change the admin password, turn off WPS and UPnP, pick a long passphrase and turn on automatic updates, and you have handled the large majority of practical risk. You do not need a firmware scanner for one router.
Free tools also exist for one-off analysis. Fraunhofer FKIE built its Firmware Analysis and Comparison Tool (FACT) for exactly the study cited above, and binwalk is the standard way to unpack an image. A curious engineer can learn a lot from them.
The case for a platform starts where scale and accountability start. A device maker shipping dozens of firmware versions, an internet provider distributing routers to subscribers, or an enterprise running hundreds of access points needs inventories that stay current, vulnerabilities ranked by real exploitability, vendor SBOMs requested and tracked, and fixes that turn into releases. That is software supply chain work, and it is where Safeguard fits.
How to test your own Wi-Fi legally
Only test networks you own, or networks whose owner has given you written permission with a clear scope. Testing a network you do not own or have written permission for is illegal in most countries, and that includes capturing handshakes, sending deauthentication frames and running fake access points.
For a home network, you can do a lot without any attack tooling at all:
- Read your router's settings against the home checklist above.
- Look up your router model's end-of-support date and latest firmware on the vendor's site.
- Check the firmware's components in the free Gold directory or with an SBOM.
- Look at the connected device list and identify everything on it.
- Check whether your router's admin page is reachable from outside your network, for example by trying it from a phone on mobile data.
For a business, hire a qualified tester for an authorized wireless assessment. A good scope covers passphrase strength for any pre-shared key networks, WPS, rogue and evil twin resistance of managed devices, enterprise certificate validation, segmentation between guest, IoT and corporate networks, and the router and access point firmware itself.
Frequently asked questions
How do hackers hack Wi-Fi?
Mostly by guessing weak passwords offline after recording a WPA2 login exchange or PMKID, abusing WPS, logging into routers with default passwords, exploiting old router firmware, or running fake access points that devices join automatically. Breaking WPA2 or WPA3 encryption directly is rare. Strong passphrases, WPS off, updated firmware and careful auto-join settings stop most attacks.
Can someone hack my Wi-Fi if I have a strong password?
A strong, random passphrase makes offline password guessing impractical, which removes the most common attack. Someone could still get in through WPS if it is enabled, through a router admin page with a default password or an unpatched vulnerability, or by tricking your devices onto a fake network. Turn off WPS, update firmware and change the admin password too.
Is WPA2 still secure?
WPA2 with AES and a long random passphrase is still reasonably secure for most homes, provided devices are patched against KRACK and FragAttacks. Its weakness is that a recorded handshake allows unlimited offline guessing, so short or predictable passwords fall quickly. Move to WPA3 when all your devices support it, and enable Protected Management Frames now.
Is WPA3 secure?
WPA3 is the most secure Wi-Fi option available. Its SAE handshake resists offline password guessing and it requires Protected Management Frames. Early implementations had flaws, shown by the 2019 Dragonblood research, including downgrade attacks in WPA2 transition mode and side channels. Those were patched. Keep devices updated and switch to WPA3-only mode when you can.
Should I turn off WPS on my router?
Yes. WPS lets an attacker recover your Wi-Fi passphrase by guessing its eight-digit PIN in about 11,000 attempts, and some routers are vulnerable to the faster offline Pixie Dust attack. CERT/CC's recommended workaround is to disable WPS. Turn off the PIN method and, ideally, push-button WPS too, and replace routers that cannot disable it.
Is public Wi-Fi safe to use?
Public Wi-Fi is fairly safe for HTTPS browsing on an updated device, but fake hotspots, fake sign-in pages and other users on the network are real risks. Use a VPN, keep browsers in HTTPS-only mode, turn off auto-join for public networks, and never enter account passwords into a Wi-Fi sign-in page. Use your phone's hotspot for sensitive work.
How can I tell if someone has hacked my Wi-Fi or router?
Look for unknown devices in the router's connected list, DNS or port forwarding settings you did not make, remote management switched on, a changed admin password, frequent disconnects or unexpected web pages. If you see these, disconnect the router, factory reset it, update its firmware, and set new admin and Wi-Fi passwords.
How do I secure my home Wi-Fi?
Use WPA3, or WPA2 with AES, and a passphrase of four or more random words. Turn off WPS, turn on Protected Management Frames, change the router's admin password, turn off remote administration and UPnP, and enable automatic firmware updates. Put guests and smart devices on separate networks, and replace your router when it stops receiving updates.