CVE-2024-1708, CVE-2024-1709, CVE-2026-21510, CVE-2026-21513, CVE-2026-25874, CVE-2026-32202, CVE-2026-3854, CVE-2026-42208
IP Addresses:
65.111.27.132
Get tomorrow's brief in your inbox
Today: CISA adds exploited ConnectWise and Windows flaws to KEV catalog, VECT 2.0 ransomware accidentally destroys files instead of encrypting them, and a critical GitHub vulnerability exposed millions of repositories to remote code execution. LiteLLM SQL injection exploited within 36 hours of disclosure, and Microsoft deprecates legacy TLS in Exchange Online starting July.
CISA Adds Actively Exploited ConnectWise and Windows Flaws to KEV (CVE-2024-1708, CVE-2026-32202)
CISA added two actively exploited vulnerabilities to the Known Exploited Vulnerabilities catalog on April 28. CVE-2024-1708 (CVSS 8.4, EPSS 53.7%/98th percentile) is a path traversal flaw in ConnectWise ScreenConnect fixed in February 2024, exploited by China-based threat actor Storm-1175 alongside CVE-2024-1709 (CVSS 10.0, EPSS 94.3%/100th percentile, ransomware-linked) to deploy Medusa ransomware. CVE-2026-32202 (CVSS 4.3, EPSS 0.1%/26th percentile) is a Windows Shell protection mechanism failure patched in April 2026, stemming from an incomplete fix for CVE-2026-21510. This flaw was exploited as a zero-day by Russian hacking group APT28 alongside CVE-2026-21513 (EPSS 27.8%/96th percentile) in attacks targeting Ukraine and EU countries since December 2025.
Critical GitHub Vulnerability Exposed Millions of Repositories (CVE-2026-3854)
A critical command injection flaw (CVE-2026-3854, CVSS 8.7, EPSS 0.4%/60th percentile) in GitHub.com and GitHub Enterprise Server allowed any authenticated user with push access to execute arbitrary commands on backend servers with a single git push command. The flaw stemmed from improper sanitization of user-supplied push option values before inclusion in internal X-Stat headers. On GitHub.com, exploitation enabled cross-tenant exposure, allowing attackers to read millions of repositories on shared storage nodes regardless of organization or user. GitHub deployed a fix to GitHub.com within two hours of Wiz's March 4, 2026 report. Enterprise Server versions 3.14.25, 3.15.20, 3.16.16, 3.17.13, 3.18.8, 3.19.4, 3.20.0 or later contain the patch. 88% of Enterprise Server instances remain unpatched as of disclosure.
LiteLLM SQL Injection Exploited Within 36 Hours (CVE-2026-42208)
CVE-2026-42208 (CVSS 9.3), a SQL injection in BerriAI's LiteLLM Python package affecting versions >=1.81.16 and <1.83.7, was exploited within 36 hours of public disclosure. The vulnerability allowed unauthenticated attackers to send specially crafted Authorization headers to any LLM API route to reach a vulnerable database query, enabling read and modification of the proxy's database including upstream LLM provider keys (OpenAI, Anthropic, AWS Bedrock). First exploitation occurred April 26 at 16:17 UTC (26 hours after GitHub advisory indexing) from IP 65.111.27.132, targeting database tables holding credentials and proxy config. A single litellm_credentials row often holds multi-thousand dollar monthly spend caps and admin rights.
3 claims tracked across 2 groups in the last 48 hours. These are unverified claims from ransomware leak sites, not confirmed breaches.
| Group | Victim | Sector | Country |
|---|---|---|---|
| mnt6 | McKay | Engineering/Defense | Not specified |
| mnt6 | Silfab Solar | Manufacturing/Energy | North America |
| fulcrumsec | Avnet | Technology Distribution | Global (EMEA focus) |
VECT 2.0 Ransomware Acts as Data Wiper for Large Files
VECT 2.0 ransomware, advertised as ransomware-as-a-service and partnered with TeamPCP and BreachForums, contains a critical flaw that permanently destroys any file larger than 131KB instead of encrypting it. Check Point Research discovered the ransomware uses ChaCha20-IETF encryption with a four-chunk design for large files but discards three of four decryption nonces, rendering 75% of every large file unrecoverable even by the operators. The flaw affects Windows, Linux, and ESXi variants. At a 128KB threshold, virtually all valuable enterprise files (VM disks, databases, backups, documents) are destroyed rather than encrypted. VECT has listed 25 organizations since January, including claims against Guesty (700GB) and S&P Global (250GB) allegedly tied to TeamPCP supply chain compromises. The group partnered with TeamPCP to exploit victims of Trivy, LiteLLM, Checkmarx, and Telnyx supply chain attacks.
Feuding Ransomware Groups Leak Each Other's Data
Two newer ransomware-as-a-service groups, 0APT and KryBit, engaged in a public feud that exposed infrastructure and operational data. 0APT emerged in January 2026 with a fabricated victim list, went quiet, then reemerged in mid-April claiming attacks against ransomware operators including KryBit, Everest, and RansomHouse. 0APT leaked an Everest SQL database with encoded and hashed records from the first nine months of 2025, plus KryBit infrastructure exposing two administrators, five affiliates, 20 potential victims, and ransom demands between $40,000 and $100,000. KryBit retaliated by breaching 0APT's infrastructure, leaking full operational data (access logs, PHP source code, system files) and defacing 0APT's leak site with the message "Next time, don't play with the big boys." Access logs confirmed 0APT's 190+ January victim claims were entirely fabricated with no data ever exfiltrated.
Fresh Wave of GlassWorm VS Code Extensions Slices Through Supply Chain
Socket Research discovered a new cluster of 73 malicious Visual Studio Code extensions related to the self-propagating GlassWorm campaign targeting Open VSX marketplace. The latest wave includes "sleeper" extensions that remain dormant before being weaponized and features automatic payload fetching and execution at a later date. At least six extensions have been activated with malware, while others remain sleepers or appear suspicious. Extensions impersonate legitimate listings by cloning names, icons, descriptions, and README content. The malware targets developer workstations to steal source code, credentials, API keys, SSH keys, cloud tokens, and package publishing credentials. Attackers can then publish poisoned versions of projects maintained by victims, creating downstream supply chain effects.
Cyber Insurance Data Highlights MFA Misconfiguration as Top Loss Driver
Resilience's analysis of manufacturing cyber insurance claims from March 2021 through February 2026 found that 90% of incurred losses stem from ransomware, despite representing only 12% of claims. MFA misconfiguration is the number one point of failure, causing 26% of financial loss, double the loss from software vulnerability exploits (13%). The single largest loss in the portfolio, a BlackCat ransomware attack, was directly enabled by misconfigured MFA. MFA absence caused 8% of losses. Beyond ransomware, transfer fraud and email compromise comprise 30% of all claims, with phishing leading to credential compromise as the primary point of failure.
Microsoft Outlook for iOS Still Down After Service Change
Microsoft Outlook for iOS experienced outages beginning April 27 at 0845 UTC due to a recent service change. Users experienced intermittent sign-in failures, "too many requests" errors, and unexpected sign-outs, with some locked out after too many failed password attempts. Microsoft rolled back the configuration change and declared the service recovered as of April 27 at 1916 UTC, but users continued reporting access issues more than 24 hours later. The status page still reports "Service degradation" for consumer products as of April 28 publication time.
Microsoft Teams Free Backend Change Broke Chat and Calls
A recently deployed backend change is preventing some Microsoft Teams Free users from chatting and calling others. The change skips onboarding and privacy consent screens for some new users who signed up during the impact window (beginning April 8, approximately three weeks before acknowledgment), causing profiles to appear as "Unknown users" and preventing them from being searchable or reachable in chat. Microsoft flagged this as a "service degradation" and is still looking for a solution.
OpenAI Models Now Available on AWS Bedrock
OpenAI's models are officially available on Amazon Web Services' Bedrock managed inference and agent platform in limited preview. GPT-5.4 is available now, with GPT-5.5 coming within the next couple of weeks. The collaboration provides an alternative to accessing OpenAI's models without exposing data to OpenAI's APIs, addressing enterprise security policy, data privacy, and sovereignty concerns. The partnership stems from OpenAI's February deal to make models available on AWS in exchange for up to $35 billion in new financing, contingent on spinning up two gigawatts of Amazon's Trainium accelerators. Microsoft's willingness to open its relationship with OpenAI in exchange for being freed from revenue sharing commitments enabled this deal.
Critical Unpatched Flaw in Hugging Face LeRobot (CVE-2026-25874)
CVE-2026-25874 (CVSS 9.3, EPSS 0.1%/19th percentile) is an unsafe deserialization vulnerability in LeRobot, Hugging Face's open-source robotics platform (24,000 GitHub stars). The flaw stems from using pickle.loads() to deserialize data received over unauthenticated gRPC channels without TLS in the async inference PolicyServer component. An unauthenticated network-reachable attacker can send a crafted pickle payload through SendPolicyInstructions, SendObservations, or GetActions gRPC calls to achieve arbitrary code execution. The vulnerability impacts AI inference systems that often run with elevated privileges to access internal networks, datasets, and compute resources. Exploitation enables complete compromise of the PolicyServer host, impact to connected robots, theft of API keys and SSH credentials, lateral movement, and operational sabotage with physical safety risks. The flaw remains unpatched, with a fix planned in version 0.6.0. Issue was independently reported in December 2025 and acknowledged in January 2026.
Microsoft to Deprecate Legacy TLS in Exchange Online Starting July
Microsoft will start blocking legacy TLS 1.0 and TLS 1.1 connections for POP and IMAP email clients in Exchange Online starting July 2026. TLS 1.0 (introduced 1999) and TLS 1.1 (introduced 2006) are industry-deprecated and no longer considered secure. Most POP and IMAP traffic to Exchange Online today uses TLS 1.2 or higher. After deprecation, POP3 and IMAP4 connections will require TLS 1.2 or later, connections using TLS 1.0 or 1.1 will fail, and legacy applications or devices may stop connecting. Customers who use POP or IMAP to access email should ensure email clients and applications support TLS 1.2 or later and don't use legacy endpoints.
Exploitation windows continue to collapse. LiteLLM was exploited within 36 hours of disclosure, consistent with the broader "Zero Day Clock" pattern where advisories and open-source schemas provide sufficient information for exploitation without waiting for public proof-of-concepts. Supply chain attacks targeting developer tools (GlassWorm VS Code extensions, TeamPCP's Trivy/LiteLLM compromises, VECT's partnership with BreachForums) demonstrate industrialized ransomware deployment models. Identity infrastructure remains the primary battlefield, with MFA misconfiguration causing 26% of financial losses in Resilience's insurance portfolio, twice the impact of software vulnerability exploits. The VECT ransomware wiper flaw highlights the risks of paying ransoms to technically unsophisticated operators who cannot deliver working decryptors even if they wanted to.