20,225 Instagram accounts may have been impacted in a recent attack that abused an AI-powered support tool, a single number in a week crowded with supply-chain worms, malicious packages, and long-running mailbox espionage.
Miasma worm strikes 73 Microsoft GitHub repositories
Microsoft's GitHub repositories across four organizations — Azure, Azure-Samples, Microsoft, and MicrosoftDocs — were affected by a self-replicating supply chain campaign tracked as the Miasma worm. The incident touched 73 repositories, prompting GitHub to disable access to those projects. Security analysts assess Miasma as a variant of the Mini Shai-Hulud worm publicly released by TeamPCP in mid-May 2026, highlighting how publicly circulated malware can rapidly mutate into supply-chain threats that touch major vendor codebases.
Google issues June fixes for Android; CVE-2025-48595 is already under scrutiny
For June 2026 Google released patches for 124 Android vulnerabilities, including a high-severity Framework flaw tracked as CVE-2025-48595 with a CVSS score of 8.4. The bug is described as a privilege-escalation issue that does not require user interaction and impacts Android 14, 15, 16, and 16 QPR2. Google acknowledged "indications that CVE-2025-48595 may be under 'limited, targeted exploitation,'" while declining to disclose specifics about actors, targets, or scale.
Instagram accounts compromised via AI-powered support tool; web reset flow exposed contact data
Meta disclosed that 20,225 Instagram accounts may have been impacted when attackers abused a High Touch Support (HTS) chatbot. According to the report, attackers convinced the bot to link their own email address to targeted accounts, enabling password resets and account takeovers; many high-profile accounts were later sold on the dark web. The exploitation was discovered on May 31, 2026 and the HTS capability has since been disabled. Separately, a vulnerability in Instagram's web-based password reset flow was disclosed that exposed unredacted email addresses and phone numbers associated with accounts when a username was provided, and Meta has not yet clarified what personal information, if any, was accessed.
Supply-chain and package attacks: Hola, npm, PyPI and loader families
Supply-chain compromises and malicious open-source packages peppered the week. Sophos found an XMRig miner bundled into a certified Hola Browser Windows installer; Hola said the issue stemmed from a compromise of its "update distribution pipeline" and stated, "This was a supply chain compromise, and critically, no user data was accessed, exfiltrated, or compromised at any point during this incident affecting 0.1% of users." Hola said it rebuilt its distribution pipeline and added stronger code-signing and monitoring.
On developer ecosystems, multiple malicious npm packages were reported. OpenSourceMalware described packages that drop malware impersonating an AI coding tool and harvest ~/.local/share/stardrop/auth.json and other credentials via a post-install hook. SafeDep detailed two trojanized axios clones, turbo-axios and faster-axios, which used postinstall hooks to fetch remote JavaScript and ultimately delivered Epsilon Stealer, an Electron infostealer. OX Security flagged a package named cms-store-ren that exfiltrates developer data to Telegram while leaking its own bot API token. On PyPI, Zscaler reported Parsimonius — a typosquat of parsimonious — which combined legitimate parsing functionality with a Telegram-based backdoor and accumulated 2,474 downloads before removal.
Five months of mailbox espionage on a stock exchange executive
Unknown actors spied on a senior member of an unnamed global stock exchange by maintaining access to an Outlook mailbox for at least five months. The earliest sign of malicious activity was observed on October 10, 2025. Attackers deployed a mailbox stealer that ran in two- to four-week intervals to collect email data and exfiltrated small batches via Dropbox and Microsoft OneDrive Personal; the exfiltration activity continued through March 2026. While details such as the initial access vector and actor identity remain unspecified, analysts assessed the operation's primary goal as cyber espionage.
What this means for technologists, procurement leaders, and end users
- Technologists and security teams: patch prioritization matters — urgent fixes like CVE-2025-48595 and the trending CVEs listed for this week require fast, verifiable deployment, while code-hosting platforms should audit repositories and CI pipelines for worm-like propagation indicators such as Miasma's behavior.
- Procurement and product teams: the Hola incident and multiple malicious package reports underscore supply-chain risk in update channels and package registries; teams should verify update pipelines, strengthen code-signing, and vet third-party packages before deployment.
- End users and account holders: the Instagram HTS abuse and the mailbox espionage case show that account recovery and mailbox access flows can be abused — users should monitor account recovery settings and be alert to unusual password-reset activity or unexplained mail forwarding or access.
As The Hacker News’ weekly recap put it: "Same tricks. Same shortcuts. Same open inboxes." The week’s headlines — from a worm that reached Microsoft repos to an AI-assisted support exploit that touched tens of thousands of Instagram accounts — are a blunt reminder that old vectors still work and that patching, pipeline hygiene, and attention to account recovery flows remain urgent.
