What Recent Reports on APT-C-60's SpyGlace Reveal About File-Based Risks
Blog Post

What Recent Reports on APT-C-60's SpyGlace Reveal About File-Based Risks

File Security
Malware
Phishing Detection
Multiscanning
Deep CDR
Sandbox
September 8th 2026
Share this Post

 Tl;dr JPCERT/CC has documented three waves (Dec 2024, Nov 2025, July 2026) of SpyGlace backdoor attacks against Japanese organizations by the APT-C-60 group. No CVE is directly attributed to the SpyGlace delivery chain itself, and SpyGlace does not have a dedicated group entry in the MITRE ATT&CK knowledge base. Across all three waves, the attack involves files moving from inbox to full backdoor access, which is exactly why file-based inspection deserves more weight in the defense conversation than it currently gets. 

 

Key Takeaways 

  • APT-C-60 targets HR and recruitment staff specifically, because job application attachments are expected and rarely questioned. 

  • Delivery keeps changing: Google Drive links in 2024, direct VHDX attachments in 2025, and Proton Drive links to RAR archives (or direct attachments) in 2026. 

  • The 2026 wave introduces a new LOLBin, mshta.exe, and expands the abused infrastructure to GitHub, GitLab, jsDelivr, and Codeberg, on top of the previously seen StatCounter and Bitbucket. 

  • Every stage past the initial LNK click is invisible to the user, with no popup or ransom note displayed. 

  • The attack involves files at different stages across every wave, which is exactly why multi-layered file inspection involving multiscanning, Deep CDR, and sandboxing is warranted in today's threat landscape. 

 

Every company with an open job posting has an inbox that exists specifically to receive files from strangers. HR opens it; no questions asked. As it turns out, that inbox and that attachment are also a reliable path into a corporate network, which is exactly what APT-C-60 has been counting on. 

 

APT-C-60 is a cyber espionage group, deploying the SpyGlace backdoor once it gets in. The group has spent the past several years refining how it delivers SpyGlace onto Japanese organizations' devices, but the method hasn't changed much: pose as a job applicant, email a company's recruitment contact, and wait for someone in HR to open the attachment. 

 

JPCERT/CC has now documented this playing out three times: first in August 2024, again through Q3 2025, and most recently in a wave observed through 2026. Each wave shows a faster or more evasive path from opened attachment to working backdoor, which points to a group that keeps investing in delivery efficiency rather than changing its underlying approach. At the end of the day, SpyGlace is delivered via a chain built on files, each one designed to look innocent and trustworthy. Security teams responsible for email security, endpoint detection, or file inspection policy should pay attention. 

 

Who Is APT-C-60 and Who Are They After? 

 

APT-C-60 is a cyber espionage group active since at least 2021, with the SpyGlace backdoor first observed in June 2022. Their activity is concentrated on Japan and other East Asian regions; Japan specifically has been targeted in all three waves documented by JPCERT/CC, from 2024 through 2026. 

The group uses spear-phishing, going after recruitment, and HR contact points with weaponized job applications. The phishing documents have gotten more deliberate over time: the 2025 fake resumes listed real academic papers the "applicant" did not actually write, built to match the target organization's research area or industry. That level of detail indicates the group researches its targets before building the lure, rather than sending a generic resume to every recruiting inbox it finds. 

 

Who Is in Danger, and Are There Confirmed Victims? 

 

Any organization with a public-facing recruitment or HR contact is a plausible target, but the confirmed victims documented so far are Japanese organizations. JPCERT/CC confirmed one organization compromised in the August 2024 wave. 

The 2025 wave confirmed "at least 12 successful compromises across Japanese organizations" via victim files in a public GitHub repository. The 2026 report doesn't state a victim count directly, but its published appendix lists 24 distinct victim identifiers using the same naming convention, which would put the combined documented total across the two waves at 36 machines, before accounting for any redacted entries or victims JPCERT chose not to publish. 

 

APT-C-60's SpyGlace Attack Chain Explained 

 

The attack starts with an email impersonating a job seeker; the email email points to a RAR archive (a compressed file container) hiding an LNK file, which is just a Windows shortcut, but one built to run hidden commands when clicked. Clicking it triggers a legitimate Windows utility (mshta.exe) used as a LOLBin (Living-Off-the-Land Binary): a real, pre-installed system tool repurposed to run malicious code. The activity blends into normal system behavior. mshta.exe then runs embedded JavaScript that pulls down a downloader which retrieves the SpyGlace backdoor itself and hands the attacker persistent remote access.  

Every step uses legitimate tools, which is exactly what makes it hard to catch. 

 

 

How APT-C-60 Weaponizes Files 

 

Across all three waves, the VHDX or RAR archive, the LNK, the initial script, the decoy document, the first downloader, the second downloader payload, and the SpyGlace binary are all files, moving from inbox to full remote access, and each one is a point where inspection could interrupt the chain. APT-C-60's chain is built to defeat the controls that depend on prior knowledge: 

 

  • Domain reputation does not help against StatCounter, Bitbucket, GitHub, GitLab, jsDelivr, or Codeberg; these are all legitimate services. 

  • Signature-based detection struggles against a group that ships new SpyGlace builds on a regular cadence, as JPCERT/CC's GitHub timeline shows for versions 3.1.12 through 3.1.14 in June and July 2025, and again for 3.1.15, 3.1.17, and 3.1.18 in the 2026 wave. 

  • Hash-based blocklisting fails against payloads packaged per victim; each GitHub file is named and served using that specific machine's volume serial number and computer name, so no two victims ever share a hash worth blocklisting.  

What SpyGlace can't beat is inspection of the file itself, independent of where it came from or whether anyone has seen its hash before. A sandbox watching a VHDX get mounted, an LNK spawn mshta.exe, or a downloader assemble itself from staged .db files observes that behavior directly, regardless of whether the sample is new, and that behavior maps cleanly onto known MITRE ATT&CK techniques even though APT-C-60 itself has no formal group entry in the framework. A Deep CDR process applied to the fake resume strips whatever active content might be riding along with it, without needing to know the file is malicious first. 

Remediation Steps 

Every step in this chain relies on legitimate Windows behavior (archive extraction, LNK execution, mshta.exe scripting, COM object resolution) rather than a software vulnerability, so there is no CVE attached to any of the three waves. Additionally, at the time of writing, SpyGlace doesn't have a dedicated group entry in the MITRE ATT&CK knowledge base.  

With patching not applicable here, JPCERT/CC's published guidance centers on detection: monitoring the specific registry paths used for COM hijack persistence, flagging LNK files that spawn Git or mshta.exe processes, watching for GitHub, GitLab, jsDelivr, or Codeberg queries matching internal system naming patterns, and treating StatCounter traffic containing system information in the Referer header as an indicator of compromise. JPCERT/CC also published a Python implementation of SpyGlace's modified RC4 scheme, allowing defenders to decrypt captured command-and-control traffic directly. 

 

 File Inspection Needs the Same Attention as Vulnerability Management 

 Security teams spend a lot of energy on vulnerability management, and for good reason. However, SpyGlace proves that some of the most successful intrusions never touch a CVE; they only need someone opening a file without inspecting it first. 

That's exactly why phishing still works, even after decades of awareness campaigns. A convincing enough pretext will get a lot of employees to open an attachment or run a file, and no amount of awareness training gets that number to zero. APT-C-60's fake resumes, built around real research and matched to the target's field, are convincing. 

User awareness should be combined with layered security controls.  

MetaDefender™ Cloud lets you check a file, URL, domain, IP address, or hash against multiple reputation sources and scanning engines before trusting it, and it can inspect the LNK file itself rather than just the surrounding emai 

In practice, MetaDefender Cloud would reveal threat indicators for such an attack. While at the time of writing there were no samples available for SpyGlace, here's a comparable LNK-based campaign (Mustang Panda, 2025). Click here to see the full analysis. 


LNK sample the Mustang Panda campaign, targeting Vietnamese entities in 2025. 

If the initial compromise is going to happen regardless of how well an organization trains its staff, then the question is what happens after someone opens that attachment. APT-C-60's answer is a sequence of files moving from inbox to full remote access.   

Strong security policies, sitting at all workflow stages, should reveal weaponized behaviors in files. MetaDefender Cloud runs a submitted file, URL, or IP through +20 scanning engines at once (Metascan™ Multiscanning), strips potentially active content out of documents through the Deep CDR™ technologies, detonates suspicious files in a dynamic sandbox to observe what they actually do, and applies AI-based content detection on top. Since it's less likely that any single method would catch a chain like SPyGlace, MetaDefender Cloud layers multiple detection, inspection, and prevention techniques; it's free to use, with no account needed, whether you're inspecting a specific file or just pulling threat intelligence on an indicator you've already seen elsewhere. 

To use it, go to metadefender.com and paste the file, URL, or IP you want checked. 

 

 

File Security
Malware
Phishing Detection
Multiscanning
Deep CDR
Sandbox