Threat Research

No tap required: how zero-click and one-click attacks reach a phone

A zero-click attack needs nothing from the phone’s owner; a one-click attack needs a single tap, open or scan. Cases documented from 2021 to 2026 show how each works, why so little remains afterward, and how rare these attacks are.

Contents

The word "click" is doing a lot of work in mobile security. It is the difference between a phone that must be talked into an attack and one that is attacked while it sits on a desk.

Readers meet the terms in headlines, and the Lab’s assessment is that they are used loosely. This article sets out what they mean, walks through the documented cases, and says plainly what remains on a phone afterward.

A vocabulary that matters

A zero-click attack runs without any action by the owner: a message arrives and is processed in the background. A one-click attack needs one action, such as opening a link, a file or a scanned code. A page visit, sometimes called a watering hole, sits between them: no tap on a suspicious thing, but a trip to a web page. A zero-day is a flaw exploited before its maker can fix it. The two ideas are independent; an old, patched flaw can be used with zero clicks against a phone that was never updated.

The categories are also contested at the edges. Apple’s September 2026 fix for a CoreGraphics flaw says only that processing a crafted file may lead to code execution, and does not say how the real attackers delivered it[S12]. Without the delivery route, "zero-click" or "one-click" would be a guess.

The anatomy of a chain

Entry is only the first step. Documented chains have the same later stages: code execution inside the component that handled the content, an escape from that component’s sandbox, a climb to higher privilege, and an implant (Figure 1). Google noted that multiple chains found in 2025 included three or more flaws[S7].

Figure 1Three ways into the same chainThree entry routes that end at the same later stages. Which routes a given attack uses depends on the case; the later stages are common to documented chains. General schematic compiled from Google Project Zero (2021), Citizen Lab (2023, 2025) and Google Threat Intelligence Group (2026). Diagram by SectraTEK Research Lab.Sources:[S1][S2][S4][S7]
Describe this figure

Three entry boxes on the left: a message attachment processed in the background (zero-click), a link opened from a message or browser (one-click), and a page visit. All three arrows converge on the same four stages: code execution, sandbox escape, privilege escalation, implant.

Four cases

A file named like an image (2021). Google Project Zero analyzed what it titled an NSO zero-click iMessage exploit. A file named as an image was in fact a PDF; the system identified it by content, and a flaw in the PDF parser’s image decoder gave code execution with no user interaction[S1].

A forensic mystery (2019 to 2023). Kaspersky found Operation Triangulation when suspicious network traffic from iPhones on its own corporate Wi-Fi was noticed. An iMessage attachment of about 242 KB triggered code execution without interaction, and the exploit and the message were deleted afterward. The oldest traces date to 2019, with activity still under way in June 2023. The strongest indicator was traces of a deprecated backup-related process, not the attachment[S3]. In September 2023 Citizen Lab described BLASTPASS, a zero-click chain on fully patched iOS 16.6 that used image attachments inside Apple’s wallet component, and attributed it to NSO Group; Citizen Lab and Apple believed Lockdown Mode blocked it[S2].

Logs, not a file (2025). Citizen Lab confirmed Paragon’s Graphite spyware on the iPhone of a European journalist, running iOS 18.2.1 and compromised between January and early February 2025. Apple said the zero-click flaw (CVE-2025-43200) was mitigated in iOS 18.3.1. The case was confirmed by device logs showing requests to a server matching a known fingerprint, until at least 12 April 2025, and by an attacker’s iMessage account, rather than by a malicious file[S4]. In August 2025 WhatsApp said a linked-device flaw chained with an Apple image flaw was used against a small number of people, reportedly dozens[S11].

Serbia (2025 to 2026). Citizen Lab, working with the SHARE Foundation, confirmed Pegasus on the iPhone of a member of a student protest movement, delivered by a zero-click iMessage exploit with indicators from December 2025 to January 2026; Apple fixed the flaw in iOS 18.4.1[S5]. A second report adds an infected phone belonging to a university official, and Apple threat notifications received by a former prosecutorial associate, which Citizen Lab treats as a targeting indicator and not proof of infection[S6]. SHARE has documented at least 14 notification cases. Citizen Lab does not name a government customer in the pages the Lab reviewed.

Figure 2Documented cases, 2019 to 2026Bars show a documented activity period; diamonds show a year or month when only that is known. Triangulation: oldest traces 2019, active June 2023 (Kaspersky). Graphite: compromise January to early February 2025, requests logged to at least 12 April 2025 (Citizen Lab). Serbia: indicators December 2025 to January 2026 (Citizen Lab). Sources: Kaspersky (2023); Citizen Lab (2023, 2025, 2026); Google Project Zero (2021); Malwarebytes (2025). Chart by SectraTEK Research Lab.Sources:[S1][S2][S3][S4][S5][S11]
Describe this figure

A timeline from 2019 to 2027. A long bar for Operation Triangulation from 2019 to June 2023. Diamonds for FORCEDENTRY in 2021, BLASTPASS in September 2023, a WhatsApp and Apple image chain in August 2025. Short bars for Graphite in early 2025 and the Serbian case from December 2025 to January 2026.

What hardening changed

Lockdown Mode, Apple’s optional strict setting, limits message attachment types, link previews and complex web features. Apple has said it is not aware of any successful mercenary spyware attack against a Lockdown Mode device[S10]. That is Apple’s statement, not an independent measurement.

Hardening raises the price. Google found chains of three or more flaws common in 2025 and counted 17, 9 and 15 mobile operating-system zero-days in 2023, 2024 and 2025; these counts are floors that are revised upward later [S7][S8].

How rare is it

Rarity is real. Apple says it has sent threat notifications to users in more than 150 countries since 2021, and that "the vast majority of users will never be targeted"[S9]. Its latest wave, on 13 August 2026, reached users in 110 countries, according to Apple’s statement to BleepingComputer[S10]. Neither figure is a prevalence rate: confirmed victims in public cases run from tens to low hundreds per campaign.

Where the label gets contested

Two 2026 episodes show why precision matters. In September an independent researcher reported a memory flaw in how an iPhone decodes one image format, reachable by an image sent over iMessage and processed in the background during indexing and thumbnailing. He says it reached memory corruption but not demonstrated code execution, that it was reported to Apple in May and fixed in iOS 27, and that there is no evidence of use in the wild. A public vulnerability entry scores it as needing user interaction, which conflicts with his zero-click description[S13]. It is one researcher’s account, and the Lab treats it as an example of the class, not as a confirmed case.

In October, Calif listed a "Signal 1-click" demonstration among its Hexacon announcements, with no technical detail, no software version and no advisory from Signal[S14]. A one-click label on an unreleased demonstration tells a reader nothing yet about what the click is or what it triggers.

Our assessment

First, the attachment is rarely the evidence. In Triangulation the file was deleted[S3]; in Graphite the confirming evidence was a log of requests[S4]. Examiners who look for the document will often find nothing.

Second, the same component, the one that previews content, keeps appearing. Chains keep entering through image, document and media handlers because those run automatically. Hardening of those handlers is the most meaningful trend to watch.

Third, no public source offers a median for how long a mobile compromise goes unnoticed. The honest answer is a range: from months in the Graphite case to years in Triangulation[S3][S4], in individual cases and not as an average.

What the Lab is watching

  • Whether Apple or researchers disclose how the 2026 CoreGraphics flaw was delivered.
  • New researcher or vendor reports that identify the entry route of a confirmed infection.
  • Changes to what previews and attachment handlers do automatically.
  • Updated counts in Google’s zero-day reviews.

What this means for examiners

Preserve the whole device state, not only the suspected message. Note the iOS version and the dates a log covers. Treat an Apple threat notification as a reason to examine, not a verdict.

Our approach

How SectraTEK approaches this

The routes in this article leave different traces. A message-borne attack may leave an attachment record, or a gap where one was deleted; a link-borne one leaves a message and a browser record; all of them may leave a crash entry or a network connection. In iOS backups, iOS full file system extractions and Android full file system extractions, SectraTEK lists the documents and message attachments recorded in the evidence, with their types and the conversation they belong to. It also examines browser records, background network activity, installed apps, and the crash and system logs a device retains, including searchable decoded log events where a system diagnostic archive is examined, and it states the dates each covers.

Zero-click chains often leave little. SectraTEK does not claim to detect any named exploit, and a result with no findings means no known indicators were found in the evidence provided; it is not proof that a device was never attacked. A result is reported with the coverage behind it, so a reader knows what was and was not examined.

Facts about third parties in this article are attributed to the publisher named and were last checked on 21 October 2026. Corrections carry their date.

Sources

Facts about third parties are attributed to the publisher named in the sentence. Each marker such as [S1] opens its row here.

Sources cited in this article
MarkerPublisher and titlePublishedAccessedAddress
[S1]Google Project Zero
A deep dive into an NSO zero-click iMessage exploit: Remote Code Execution
projectzero.google
[S2]Citizen Lab
BLASTPASS: NSO Group iPhone zero-click, zero-day exploit captured in the wild
https://citizenlab.ca/2023/09/blastpass-nso-group-iphone-zero-click-zero-day-exploit-captured-in-the-wild/
[S3]Kaspersky (Securelist)
Operation Triangulation
securelist.com
[S4]Citizen Lab
First forensic confirmation of Paragon's iOS mercenary spyware finds journalists targeted
Address not listed
[S5]Citizen Lab
Pegasus spyware infection of a Serbian activist
Address not listed
[S6]Citizen Lab
Pegasus spyware and Serbian pro-democracy figures
Address not listed
[S7]Google Threat Intelligence Group
Look What You Made Us Patch: 2025 Zero-Days in Review
cloud.google.com
[S8]Google Threat Intelligence Group
Hello 0-Days, My Old Friend: A 2024 Zero-Day Exploitation Analysis
cloud.google.com
[S9]Apple
About Apple threat notifications and protecting against mercenary spyware
support.apple.com
[S10]BleepingComputer
Apple sends new threat notification alerts over mercenary spyware attacks
www.bleepingcomputer.com
[S11]Malwarebytes
WhatsApp fixes vulnerability used in zero-click attacks
www.malwarebytes.com
[S12]Apple
About the security content of iOS 26.7.1
support.apple.com
[S13]ironPeak (independent researcher)
Ex-arrr: sailing the 0-click seas
ironpeak.be
[S14]Calif
iPhone (newsletter post)
blog.calif.io

The sources were last read on . If a fact here changes, the correction carries its date.

All research