# Corrupted Word Attachments Deliver QR-Code Phishing Lures

- **Canonical URL:** https://www.probackend.com/phishing-smishing-campaigns/corrupted-word-attachments-deliver-qr-code-phishing-lures
- **Section:** Threats & Attacks
- **Published:** 2026-11-19
- **Summary:** Any.Run reported payroll and HR phishing emails that used damaged Word attachments, recovery prompts and QR-code sign-in lures to target credentials.

---

## Corrupted Word Files Lead to QR-Code Phishing

A damaged Word attachment can look less like an attack than a routine document problem. That’s the trick in a campaign reported by malware-hunting firm Any.Run and detailed by BleepingComputer: an email poses as payroll or HR, the attachment triggers Word’s recovery prompt, and the repaired document directs its reader to scan a QR code. That code leads to a page impersonating Microsoft sign-in and attempting to steal credentials.

The important point is the chain. The reporting does not say that artificial intelligence created or enabled this campaign. Its relevance to a cybersecurity threat discussion is narrower and more useful: an unusual file-handling technique can make a familiar credential lure harder for some automated checks to inspect, while the document’s instructions move the recipient toward a login page.

## How the payroll and HR lure starts

BleepingComputer reported that the emails impersonated payroll and human resources departments. Their themes centered on employee benefits and bonuses, including annual benefits, Q4 bonuses and payment-related language. Attachment names varied, with examples such as “Annual_Benefits_&_Bonus” and “Due_&_Payment,” often personalized with a recipient name. Some filenames used a `.docx` extension; others appeared as `.docx.bin` files.

Those details are specific to the samples described in the report, not a rule that every payroll email or similarly named attachment is malicious. Still, the subject matter is doing practical work: benefits and payment messages give an employee a plausible reason to open a document. A familiar department name can lower suspicion, especially when the file appears relevant to work.

BleepingComputer also noted a repeated encoded string in the campaign’s document filenames. It decodes to `##TEXTNUMRANDOM45##`. That is an observable artifact in the reported samples; the coverage does not establish that the string itself performs a malicious function. It’s better treated as a hunting clue than as proof of compromise.

## What the Word recovery prompt does

On opening one of the attachments, Word detects unreadable or corrupted content and asks whether the user wants to recover the file. The samples were damaged in a way that allowed recovery to display the document, according to BleepingComputer. In other words, the error message isn’t the end of the interaction. It’s a step in it.

After recovery, the document shows an instruction to scan a QR code to retrieve the document. The displayed page can be branded with the targeted organization’s logo; BleepingComputer’s example showed a campaign aimed at the Daily Mail. The report describes the document as a lure, not as a conventional macro-bearing file that executes code when opened.

That distinction matters. A file can be used to deliver a phishing instruction without running a malicious program. Here, the recipient is asked to take the next action—scan the code—outside the attachment itself. The damaged-file prompt and the apparent recovery process give the recipient a reason to keep engaging rather than simply close the document.

## From QR code to credential-harvesting page

BleepingComputer reported that scanning the code takes the user to a phishing site impersonating a Microsoft login page. The page attempts to collect the user’s credentials. The reported sequence is therefore direct: payroll- or HR-themed email, corrupted Word attachment, recovery prompt, recovered document with a QR code, then a counterfeit sign-in page.

A QR code changes where the next step happens. Instead of clicking a visible hyperlink in the email, a recipient uses a phone or another device to scan a code shown inside the document. That doesn’t make the destination trustworthy. It can also make the handoff less obvious to someone reviewing the original email on a desktop, because the destination is encoded in the image rather than presented as a readable link.

The reporting establishes the campaign’s QR-to-login flow; it does not establish that every QR code in an attachment is harmful or that every phishing site will use the same impersonation. The practical question is whether the request makes sense and whether the destination is expected. A document that claims a file must be retrieved by scanning a code deserves scrutiny, particularly when the email arrived unexpectedly.

## What scanner results do—and don’t—show

Any.Run said the files could operate within the operating system yet remain undetected by many security solutions because those products did not apply suitable procedures for the file types. It also reported that VirusTotal results for the submitted files showed antivirus products returning “clean” or “Item Not Found,” which Any.Run attributed to their inability to analyze the files properly.

BleepingComputer described results from attachments shared with its newsroom: almost all had zero VirusTotal detections, while one cited sample was detected by two vendors. Those figures are observations about the samples and scans discussed in the coverage. They are not a universal test of every antivirus product, every scanning configuration or every malformed Word document.

There’s another caveat in BleepingComputer’s account: the files may have had no malicious code added to them and may simply have displayed a QR code. That possibility helps explain why a scanner looking primarily for executable behavior might not flag a document. It does not prove that every detection gap had the same cause, nor does a “clean” or “Item Not Found” result establish that the destination reached by scanning the QR code is safe.

The measured conclusion is modest: the reporting found that these particular malformed attachments were poorly represented by the cited scanner results, while the user-facing document still advanced a credential-phishing flow. Don’t turn that into the claim that all scanners fail on corrupted files. Don’t assume, either, that an attachment must contain executable malware to be part of a harmful attack.

## Where AI fits, and where it doesn’t

People searching for “AI cybersecurity threats” or asking how AI is used in cybersecurity may expect a story about machine-generated phishing. This report doesn’t support that angle. It describes a campaign using corrupted Word documents, recovery behavior and QR codes; it does not attribute those tactics to AI.

In this case, the useful security lesson is about file inspection and the transition from attachment to web login. Any.Run’s explanation concerns how the file types were analyzed by security solutions. BleepingComputer’s reporting describes the lures, the repaired documents and the credential page. Neither account, as presented here, demonstrates that artificial intelligence produced the emails or bypassed defenses.

## Practical implications for this attack chain

For staff, an unexpected benefits or payroll attachment that Word says is corrupted should not be made trustworthy by the recovery prompt. Treat the QR instruction as a second decision point, not as a routine way to retrieve a file. If the message appears legitimate, verify it through a known HR or payroll contact route rather than using contact details or links supplied in the message.

For security teams, the reported samples point to a specific review path: examine suspicious `.docx` and `.docx.bin` attachments, consider what users see after Word recovery, and assess whether QR-code destinations are being checked alongside the attachment. A scanner result marked clean or “Item Not Found” should be read in light of what the tool was able to analyze, not as a guarantee about a later website.

The reported attack succeeds by making each step feel like a reasonable fix for the one before it. The email offers a benefit; Word offers recovery; the document offers a QR code; the site offers a familiar login. Breaking that sequence early is more useful than debating whether the file looks broken.

![corrupted word files lead to qr-code phishing](https://assets.probackend.com/b7486daa-6f3a-4bcf-b140-d6c890b6f20a?token=ast_7b9ee8975a4b187969d396286e3f9717c42f2331)

![corrupted word files lead to qr-code phishing](https://assets.probackend.com/92d333a8-781b-481a-9b62-62893dbd4089?token=ast_7b9ee8975a4b187969d396286e3f9717c42f2331)

![corrupted word files lead to qr-code phishing](https://assets.probackend.com/f301a93a-d532-4f4c-be0f-bc032b50455e?token=ast_7b9ee8975a4b187969d396286e3f9717c42f2331)

![ai cybersecurity threats](https://assets.probackend.com/6676c3a0-64b3-476d-bdf8-f45a6afc539c?token=ast_7b9ee8975a4b187969d396286e3f9717c42f2331)

![ai cybersecurity threats](https://assets.probackend.com/c9a682e7-38d6-4238-a1d7-3d0e0a5d2bb4?token=ast_7b9ee8975a4b187969d396286e3f9717c42f2331)
