Cell Phone Spyware Vulnerability

Cell phone spyware is commercially developed surveillance software, sold to government customers, that turns a target’s phone into a listening device. Once installed it can read messages inside encrypted apps, capture photos and files, track location, and activate the microphone and camera. The delivery method is what makes it a distinct security problem: the most capable versions require no user interaction at all.

FORCEDENTRY, disclosed by Citizen Lab in September 2021, is the reference case. It is not a historical curiosity. In August 2026 Apple sent mercenary spyware threat notifications to users in 110 countries in a single wave, the largest in the program’s history, and the company reports having notified users in more than 150 countries since the program began in 2021.

This information is provided for general awareness and defense purposes only.  This information also is not intended to be a complete description of the functionality or risks of the identified tools.

Quick Facts

• What it is: Commercial surveillance software sold to government customers, capable of reading encrypted messages, exfiltrating files, tracking location, and activating the microphone and camera.
• The reference case: FORCEDENTRY, tracked as CVE-2021-30860, an integer overflow in Apple’s CoreGraphics image rendering that Apple patched in iOS 14.8 on September 13, 2021.
• Zero interaction required: No link to click and no file to open. Security awareness training provides no protection against this class.
• Still active: Apple notified users in 110 countries on August 13, 2026, its largest single notification wave, and in more than 150 countries since 2021.
• The one control with a track record: Apple’s Lockdown Mode, which Citizen Lab confirmed blocked the 2023 BLASTPASS chain and which Apple says it has never seen defeated.
• The physical consequence: A compromised phone is an attacker-controlled microphone, camera, and location beacon in whatever room it is carried into.

What Was FORCEDENTRY?

Citizen Lab at the University of Toronto identified the exploit in March 2021 on the phone of a Saudi activist and provided the artifacts to Apple in September 2021. Apple confirmed a zero-day affecting iOS and macOS and shipped fixes on September 13, 2021.

• The vulnerability: CVE-2021-30860, an integer overflow in image rendering. Apple described it as processing a maliciously crafted PDF leading to arbitrary code execution.
• The delivery: A malicious file arriving over iMessage, processed automatically by the receiving device.
• Why it mattered technically: It defeated BlastDoor, the sandbox Apple had introduced specifically to contain malicious message content. Google’s Project Zero, which published a detailed analysis, described the chain as among the most technically sophisticated exploits it had examined.
• Scope: iPhones running iOS before 14.8, Macs before macOS Big Sur 11.6 and the corresponding Catalina security update, and Apple Watches before watchOS 7.6.2. Evidence of use dates to at least February 2021.
• The payload: NSO Group’s Pegasus.

What Zero-Click Actually Means

Most phone compromise requires a mistake: tapping a link, installing an app, entering credentials. Zero-click exploitation removes that step. The target’s device processes attacker-supplied content automatically, as part of normally receiving a message, and the exploit runs before anything appears on screen.

Three consequences follow, and they are the ones security teams most often miss:

• User training is irrelevant here. There is no decision the target makes, and no amount of caution changes the outcome. Awareness programs that treat phishing discipline as the control do not touch this class.
• There may be no user-visible artifact. Some chains delete the delivery message, leaving nothing for the target to notice or report.
• Attack surface is whatever parses untrusted input automatically. Messaging apps, image and font libraries, and codecs are the recurring targets precisely because they run without being asked.

The Threat Class Today

FORCEDENTRY was neither the first nor the last of its kind. The pattern since has been consistent: a new zero-click chain is found in the wild on the device of a journalist, activist, or official, it is patched, and another appears.

Case	Year	What it shows
FORCEDENTRY, CVE-2021-30860	2021	Zero-click via iMessage defeating a purpose-built sandbox; Pegasus payload
BLASTPASS, CVE-2023-41064 and CVE-2023-41061	2023	Zero-click chain using PassKit image attachments against a fully updated iPhone; Citizen Lab confirmed Lockdown Mode blocked it
Graphite, CVE-2025-43200	2025	Zero-click flaw in Messages used against European journalists in January and February 2025, delivering Paragon’s spyware
Predator resurgence	2024 onward	Documented return of activity despite public exposure, sanctions, and policy intervention
Apple notification wave	2026	110 countries notified in a single wave on August 13, 2026; more than 150 countries since 2021

The market matured rather than shrank. Where a capability of this kind was once the preserve of a handful of intelligence services, several commercial vendors now sell comparable products, and the customer list extends well beyond the countries most people would assume. Public exposure has raised costs for these vendors without ending the business.

Am I Actually a Target?

Probably not, and that framing is more useful than either dismissal or alarm. This is targeted tooling, not mass malware. Operators pay per target and spend valuable exploit chains deliberately, which means selection is narrow.

• Consistently targeted: journalists, human rights defenders, lawyers, opposition politicians, diplomats, and government officials.
• Also in scope, and frequently underestimated: executives during mergers and acquisitions, negotiators, defense and aerospace personnel, researchers with export-controlled knowledge, and anyone with routine physical proximity to sensitive discussions.
• Proximity matters as much as position. An executive assistant, a driver, or a spouse can be a more practical target than the principal, and the microphone works the same either way.

For an organization the useful question is not whether every employee is at risk. It is whether the handful of people who sit in the room where decisions get made have a device posture that reflects being worth targeting.

Can Cell Phone Spyware Be Detected?

Sometimes, with effort, and never with certainty. Detection here is probabilistic, not binary, and content that promises otherwise is selling something.

Method	What it gives you	Limits
Vendor threat notification	The strongest single signal available; Apple and WhatsApp both notify targeted users	Only covers what the vendor detects and attributes; absence of a notification proves nothing
Mobile Verification Toolkit (MVT)	Forensic analysis of backups and system logs against published indicators of compromise	Requires expertise to run and interpret; relies on known indicators; a clean result is not an all-clear
Behavioral scanners such as iVerify	Anomaly and integrity checks accessible to non-specialists	Heuristic, and a determined operator is specifically trying to avoid this
Backup-based tooling such as iMazing	Automates parts of the MVT workflow for less technical users	Same underlying dependence on known indicators
Device management profile review	Catches the cruder cases that install configuration profiles	Not how the high-end chains operate
Battery, heat, and data anomalies	Occasionally the first thing a target notices	Weak evidence; a hundred benign explanations exist

Two further complications are worth naming. Some implants are non-persistent by design, meaning a reboot removes the running code while leaving the target just as exposed to reinfection, and it also erases much of the evidence. And forensic traces live in logs that rotate, so the window for confirming an infection can close quickly. If you suspect compromise, preserve a backup before rebooting or resetting anything.

Does Lockdown Mode Work?

It has the best track record of any user-accessible control against this threat class. Lockdown Mode reduces attack surface deliberately, blocking most message attachment types, disabling link previews, restricting certain web technologies, and limiting connections from unknown accounts and accessories.

• Documented success: Citizen Lab reported that Apple confirmed Lockdown Mode blocks the BLASTPASS chain, and Apple states it has not seen a device successfully compromised while in Lockdown Mode.
• Apple’s own guidance: anyone receiving a threat notification should enable it, and contact the Access Now Digital Security Helpline for rapid-response assistance.
• The honest caveat: “never seen” is a statement about observed cases, not a proof. Lockdown Mode raises cost substantially; it does not make a device unexploitable.
• The cost: real usability friction, which is why it works best applied to a defined high-risk group rather than an entire workforce.

Android has no single equivalent switch. The nearest practical posture combines a device with a long security-update commitment, prompt patching, Google’s Advanced Protection Program, and minimizing installed messaging surfaces.

What Does Not Help

• A VPN. It moves where your traffic egresses. The compromise happens on the device, above the tunnel, and a VPN sees none of it.
• Mobile antivirus. On iOS, third-party apps have no visibility into the system internals this would require. On Android, detection of custom, targeted implants is unreliable.
• Encrypted messaging alone. End-to-end encryption protects messages in transit. An implant reads them on the endpoint after decryption, which is the entire point of endpoint compromise.
• Security awareness training. Effective against phishing and social engineering, irrelevant against zero-click delivery.
• A factory reset, on its own. It may remove an implant while leaving the account, the phone number, and the targeting entirely intact. Reinfection is a routine outcome.
• Turning the phone off during a meeting. Better than nothing, unverifiable in practice, and it depends on trusting that the software controlling the power state has not been compromised.

Why This Is a Physical Security Problem

The reason this research sits alongside our wireless threat work is that a compromised phone is not primarily a data breach. It is a microphone, a camera, and a location beacon that a person carries into rooms where the organization assumes it is speaking privately.

That reframes the mitigation. You cannot reliably tell whether a given phone is compromised, but you can control whether phones are present where it would matter most:

• Board rooms and executive suites during merger, earnings, litigation, and product discussions.
• Research and engineering spaces holding trade secrets or export-controlled work.
• SCIFs and SAPFs, where personal electronic devices are already prohibited and continuous wireless detection is mandated rather than optional.
• Any space where a policy exists on paper and nothing verifies it.

The gap in most programs is not the policy. It is that the policy runs on the honor system, and an unenforced device ban provides the feeling of security without the fact of it.

Policy, Sanctions, and Litigation

Governments and platforms have pushed back, with mixed results that are worth understanding if you are assessing how durable this threat is.

• Export controls. The US Commerce Department added NSO Group and Candiru to the Entity List in November 2021, followed by Cytrox and Intellexa in July 2023.
• Executive Order 14093, signed in March 2023, prohibits US government operational use of commercial spyware that poses national security or counterintelligence risks.
• Sanctions. The Treasury Department designated individuals and entities in the Intellexa consortium in 2024 for developing and distributing spyware used against Americans.
• Litigation. A jury found for WhatsApp against NSO Group in May 2025, awarding roughly $444,000 in compensatory damages and $167 million in punitive damages. A court later reduced the punitive award to about $4 million as constitutionally excessive while ordering NSO to stop targeting WhatsApp. NSO appealed, and in 2026 Meta returned to court seeking contempt findings over continued activity.

The practical read: enforcement raises the cost of doing this business and has not stopped it. Predator activity resumed after exposure and sanctions, and new vendors continue to enter. Plan for the capability to remain available.

If You Receive a Threat Notification

1. Treat it as credible. These notifications are deliberately conservative and are not sent casually.
2. Preserve evidence before changing anything. Take an encrypted backup first. Rebooting, resetting, or updating can destroy the artifacts a forensic analyst would need.
3. Enable Lockdown Mode and update the operating system.
4. Get expert help. Apple points recipients to the Access Now Digital Security Helpline, which provides rapid-response assistance at no cost to civil society targets.
5. Assume account compromise, not just device compromise. Rotate credentials and review session and recovery settings from a different, trusted device.
6. Think about what the device heard. Reconstruct which meetings and conversations happened in its presence, and treat that as the actual scope of the incident.

How Bastille Fits, and Where It Does Not

Being direct about scope, because it matters here more than usual: Bastille does not detect spyware running on a phone. No RF monitoring system does, and any vendor claiming otherwise should be asked how.

What Bastille addresses is the physical consequence of that threat, which is the presence of a potentially compromised device in a place where sensitive conversations happen. Monitoring is 100% passive across 100 MHz to 7.125 GHz using software-defined radios.

The pairing to aim for is endpoint hardening for the people most likely to be targeted, and airspace monitoring for the rooms where a compromised device would do the most damage. See wireless intrusion detection for secure facilities and related coverage of eavesdropping and surveillance devices.

Frequently Asked Questions

What was the FORCEDENTRY vulnerability?

FORCEDENTRY, CVE-2021-30860, was an integer overflow in Apple’s image rendering code that allowed arbitrary code execution when a device processed a maliciously crafted file received over iMessage. Citizen Lab captured it in the wild in 2021 and Apple patched it in iOS 14.8 on September 13, 2021. It delivered NSO Group’s Pegasus spyware.

What is a zero-click attack?

A zero-click attack compromises a device without any action by the user. The target does not click a link or open a file; the device processes attacker-controlled content automatically as part of receiving a message. This is why security awareness training offers no protection against it.

Can cell phone spyware turn on my microphone and camera?

Yes. Commercial spyware of this class can access messages, photos, and files, track location, and activate the microphone and camera. That capability is why a compromised phone should be treated as a surveillance device rather than only as a compromised endpoint.

How do I know if my phone has spyware on it?

Usually you do not, absent a notification from Apple or WhatsApp or a forensic examination. The Mobile Verification Toolkit from Amnesty International’s Security Lab analyzes backups and logs against known indicators, and tools such as iVerify apply behavioral checks. All of these are probabilistic, and a clean result is not proof of a clean device.

Does Lockdown Mode actually protect against Pegasus?

It has the strongest track record of any user-accessible control. Citizen Lab reported that Lockdown Mode blocked the BLASTPASS chain, and Apple says it has never observed a successful compromise of a device in Lockdown Mode. That is an observation rather than a guarantee, and the mode carries genuine usability costs.

Am I likely to be targeted by mercenary spyware?

Most people are not. Targeting is deliberate and expensive, concentrating on journalists, activists, lawyers, diplomats, and officials, plus executives and negotiators in sensitive commercial situations. The relevant organizational question is whether the small number of people near your most sensitive decisions are hardened accordingly.

Will a factory reset remove spyware?

It may remove the implant, and it does not remove the targeting. Many implants are non-persistent anyway, so a reboot clears the running code while reinfection remains available to the operator. A reset also destroys forensic evidence, so preserve a backup first if you intend to investigate.

Does a VPN or encrypted messaging protect me?

No. A VPN protects traffic in transit and the compromise happens on the device itself. End-to-end encryption protects messages between endpoints, and an implant reads them on the endpoint after decryption.

What should I do if Apple sends me a threat notification?

Treat it as credible, take an encrypted backup before changing anything, enable Lockdown Mode, update the operating system, and contact the Access Now Digital Security Helpline, which Apple recommends for emergency assistance. Then rotate credentials from a separate trusted device and assess what the phone was present for.

Is this still happening in 2026?

Yes. Apple sent threat notifications to users in 110 countries on August 13, 2026, the largest single wave since the program started, and has notified users in more than 150 countries since 2021. Zero-click chains delivering Paragon’s Graphite were documented against journalists in 2025, and Predator activity resumed after sanctions.

Can wireless monitoring detect spyware on a phone?

No, and Bastille does not claim to. RF monitoring detects that a device is present and where it is, which is what lets an organization enforce device policy in rooms where a compromised phone would be a listening device. Detecting the implant itself is an endpoint forensics problem.

Is commercial spyware legal?

It is sold to government agencies under export licensing regimes, and its use against a given individual may or may not be lawful in that jurisdiction. Several vendors face US export controls and sanctions, and Executive Order 14093 restricts US government operational use of commercial spyware that poses national security risks.

What Can I Do to Defend Against This Threat?

• Identify your actual high-risk group, which is usually a few dozen people, and harden those devices specifically: Lockdown Mode, aggressive patching, minimal messaging surfaces.
• Patch on a short clock. These chains get burned when disclosed, and the window between a patch and a compromised device being updated is the exposure that remains.
• Establish a phone-free rule for sensitive discussions, covering mergers, earnings, litigation, product plans, and trade secrets, and apply it to visitors and guests as well.
• Do not rely on the honor system. Deploy detection that verifies which devices are actually present in those rooms.
• Write the response plan first: who is called when someone receives a threat notification, who takes the forensic backup, and how you scope what the device overheard.
• Keep expectations honest with leadership. Against a well-funded operator with a fresh zero-click chain, endpoint hardening reduces risk and does not eliminate it, which is exactly why controlling device presence matters.

Assume the phone in the room may not be trustworthy, and design the room accordingly. See a demo to understand how Bastille detects and locates devices in your most sensitive spaces, or browse more Bastille wireless security research.