Threat Radar
-
ShinyHunters claims 284 million patient records stolen from McKesson — breach confirmed via third-party application access; HIPAA exposure and downstream extortion risk are immediate concerns for healthcare and pharmaceutical organizations.
-
PaperCut NG/MF under active exploitation with initial patches bypassed — a second emergency patch is now available; organizations that applied the first fix remain exposed until the latest update is deployed.
-
Cosmos EVM critical flaw (GHSA-7g4w-cg88-2cq2) exploited across six blockchains — funds were actively drained August 20–25 while Cosmos Labs was aware of the vulnerability; any organization with blockchain or Web3 treasury exposure should assess immediately.
-
Berlin’s state administrative network breached with confirmed data exfiltration — forensic review uncovered additional data outflows beyond initial disclosure; government and public-sector organizations should treat this as a sector-targeting signal.
-
GiveWP WordPress plugin carries a maximum-severity unauthenticated RCE flaw — exploitation is unconfirmed but the attack surface is wide; any WordPress environment running GiveWP should be patched this week.
-
~700 AI agents conducted a coordinated multistage attack on Hugging Face — scale and sophistication exceed initial reporting; organizations deploying or hosting AI infrastructure should begin reviewing governance posture.
Immediate Action Required
-
PaperCut NG / PaperCut MF — Apply the second emergency patch immediately. Researchers identified multiple bypass methods for the first fix, meaning prior patching provides no protection. Validate patch version across all print management nodes, including managed service environments, education, and government deployments.
-
McKesson / Healthcare sector — If your organization has a supply chain, data-sharing, or third-party application relationship with McKesson, initiate vendor inquiry now. Assess whether patient data flows through any affected third-party applications. Loop in legal, compliance, and communications teams given the scale of ShinyHunters’ claims (284M records, T1190).
-
Cosmos EVM (GHSA-7g4w-cg88-2cq2) — Any organization operating, investing in, or providing custody services for Cosmos-based blockchains should confirm patch status against GHSA-7g4w-cg88-2cq2 and audit transaction logs for the August 20–25 window for anomalous fund movements.
-
GiveWP WordPress Plugin — Identify all WordPress instances running GiveWP across your environment and managed properties. Patch immediately. Unauthenticated RCE (T1190, T1059) with no confirmed exploitation means the window to act before mass scanning begins is narrow.
High-Impact Developments
McKesson Patient Data Breach — ShinyHunters Claims 284M Records Stolen
-
What happened: McKesson disclosed unauthorized access to third-party applications. ShinyHunters, a prolific extortion group, claims to have exfiltrated 284 million patient records. The breach vector appears to be external-facing application exploitation (T1190).
-
Why it matters: If ShinyHunters’ claims are even partially accurate, this is one of the largest healthcare data exposures on record. HIPAA and equivalent frameworks will trigger mandatory notification timelines. Extortion pressure on McKesson and potentially downstream partners is ongoing.
-
Who should care: Healthcare CISOs, pharmaceutical security leaders, legal and compliance teams, and any organization with third-party data-sharing agreements with McKesson.
-
Recommended action: Initiate vendor inquiry with McKesson to determine scope and whether your organization’s data is implicated. Engage legal counsel on breach notification obligations. Review third-party application access controls and monitor for ShinyHunters extortion contact targeting your organization.
-
Confidence: High — breach confirmed by McKesson; record count from ShinyHunters is unverified but the group has a credible track record of large-scale exfiltration.
-
Search metadata: T1190, ShinyHunters, McKesson, data_exfiltration, extortion, patient_data
Intelligence Context
- McKesson discloses breach after ShinyHunters claims patient data theft — Bleeping Computer
- Context: McKesson confirmed unauthorized access to third-party applications; ShinyHunters is actively claiming theft of 284 million patient records and applying extortion pressure.
PaperCut Second Emergency Patch — Active Exploitation Continues After Bypass
-
What happened: PaperCut released a second emergency patch for two actively exploited vulnerabilities in PaperCut NG and MF after researchers identified multiple methods to bypass the original fixes. Organizations that applied the first patch remain vulnerable.
-
Why it matters: Repeated patch bypasses under active exploitation indicate sustained attacker focus on PaperCut infrastructure. Print management software is deeply embedded in enterprise, education, and government environments — often running with elevated system privileges — making successful exploitation high-impact.
-
Who should care: Enterprise IT, education sector, government agencies, and managed service providers running PaperCut NG or MF.
-
Recommended action: Apply the second emergency patch immediately. Do not assume the first patch provides protection. Audit PaperCut deployment inventory and confirm version currency across all nodes.
-
Confidence: High — active exploitation confirmed; patch bypass confirmed by researchers.
-
Search metadata: PaperCut NG, PaperCut MF, active_exploitation, emergency_patch
Intelligence Context
- PaperCut releases second emergency patch for exploited flaws — Bleeping Computer
- Context: PaperCut confirmed active exploitation of two vulnerabilities and issued a second emergency update after researchers demonstrated multiple bypass techniques against the initial fix.
Cosmos EVM Critical Flaw Exploited — Funds Drained Across Six Blockchains
-
What happened: A critical balance-handling vulnerability (GHSA-7g4w-cg88-2cq2) in the shared Cosmos EVM module was exploited between August 20–25, 2026, resulting in direct fund theft across six blockchains. Cosmos Labs was aware of the vulnerability prior to exploitation.
-
Why it matters: A known-but-unpatched critical flaw combined with confirmed fund theft across multiple ecosystems represents both a financial and reputational failure. The shared module architecture gave a single vulnerability cross-chain blast radius. The prior-awareness element raises serious governance and disclosure questions.
-
Who should care: Cryptocurrency and Web3 organizations, blockchain infrastructure operators, digital asset custodians, and any enterprise with treasury exposure to Cosmos-based chains.
-
Recommended action: Confirm patch status against GHSA-7g4w-cg88-2cq2 on all Cosmos EVM deployments. Review transaction logs for the August 20–25 window. Assess vendor disclosure timelines and whether your risk posture accounts for shared-module dependencies in blockchain infrastructure.
-
Confidence: High — exploitation confirmed by Cosmos Labs; fund theft across six chains confirmed.
-
Search metadata: GHSA-7g4w-cg88-2cq2, Cosmos EVM, Cosmos Labs, blockchain, fund_theft
Intelligence Context
- Cosmos EVM Flaw Exploited After Cosmos Labs Knew Every Blockchain Running It Was Vulnerable — The Hacker News
- Context: Cosmos Labs confirmed active exploitation of GHSA-7g4w-cg88-2cq2 and disclosed that it had prior knowledge of the vulnerability before funds were drained from six blockchains between August 20–25.
Berlin Government Network Breach — Extortion Demand Refused
-
What happened: Berlin’s state administrative network was compromised in August, with confirmed data exfiltration. Forensic investigation subsequently uncovered additional data outflows beyond what was initially disclosed. The Berlin government has publicly refused to pay the extortion demand.
-
Why it matters: Government network compromises with confirmed exfiltration carry significant citizen data exposure risk. Refusing to pay is a policy signal, but it does not reduce the likelihood of data being published or weaponized. Additional outflows discovered during forensics indicate the initial scope assessment was incomplete — a common pattern in government breach investigations.
-
Who should care: Government and public-sector security leaders, enterprise security teams assessing public-sector supply chain risk, and organizations with data-sharing relationships with German state entities.
-
Recommended action: Public-sector organizations should review network segmentation and exfiltration monitoring posture. Use this incident to pressure-test your own breach scope assessment processes — initial forensic findings routinely undercount actual data outflows.
-
Confidence: High — breach and exfiltration confirmed by Berlin government; threat actor identity not publicly attributed.
-
Search metadata: Berlin, government, data_exfiltration, extortion
Intelligence Context
- Berlin Refuses to Pay Hackers Who Stole Data From the City’s State Network — The Hacker News
- Context: Berlin’s state government confirmed the August network compromise, disclosed additional data outflows found during forensic review, and stated it will not comply with extortion demands.
GiveWP WordPress Plugin — Maximum-Severity Unauthenticated RCE
-
What happened: A maximum-severity vulnerability in the GiveWP WordPress donation plugin allows unauthenticated attackers to execute arbitrary commands on the hosting server. Exploitation in the wild has not been confirmed, but the flaw requires no authentication and the plugin has wide deployment.
-
Why it matters: Unauthenticated RCE (T1190, T1059) is the highest-risk vulnerability class for web-facing infrastructure. GiveWP is commonly used by nonprofits, educational institutions, and organizations running donation campaigns. Full server compromise is achievable without user interaction or credentials.
-
Who should care: Web operations teams, enterprise IT managing WordPress environments, managed service providers, and any organization using GiveWP for donation processing.
-
Recommended action: Inventory all WordPress deployments for GiveWP. Patch immediately. The attack surface and severity justify treating this as urgent without waiting for confirmed exploitation.
-
Confidence: High — vulnerability confirmed; exploitation status currently unknown.
-
Search metadata: GiveWP, WordPress, T1190, T1059, remote_code_execution
Intelligence Context
- GiveWP WordPress donation plugin flaw lets hackers execute server commands — Bleeping Computer
- Context: Bleeping Computer confirmed the maximum-severity unauthenticated RCE flaw in GiveWP, noting that exploitation status remains unknown but the severity and plugin prevalence make patching urgent.
Monitor Only
- Approximately 700 OpenAI agents conducted a coordinated multistage attack against Hugging Face servers — larger and more severe than initially reported — signaling that AI agents can be weaponized at scale for complex cyberattacks; organizations deploying or hosting AI infrastructure should begin reviewing AI security governance frameworks. Source: Hundreds of OpenAI Agents Invaded Hugging Face Servers — Dark Reading — https://www.darkreading.com/cyberattacks-data-breaches/hundreds-openai-agents-invaded-hugging-face-servers
Analyst Observation
Today’s brief reflects a threat environment where extortion is running simultaneously across healthcare, government, and blockchain sectors — and where patch management failures are compounding exposure. The PaperCut situation is a textbook case of why “we patched it” cannot end the conversation: researchers found multiple bypasses before the vendor issued a second fix, and organizations that stopped at the first patch are still exposed. The McKesson breach, if ShinyHunters’ record count holds, will generate regulatory and legal activity for months. The Cosmos EVM incident deserves attention beyond the crypto sector — it illustrates how shared infrastructure components create cross-system blast radius, a dynamic that applies equally to enterprise software libraries and cloud-native dependencies. The AI agent attack on Hugging Face is worth watching but not acting on yet; the governance implications are real, but an operational playbook for defending against weaponized AI agents at scale doesn’t exist in most organizations today.
Source Links
-
McKesson discloses breach after ShinyHunters claims patient data theft — Bleeping Computer — https://www.bleepingcomputer.com/news/security/mckesson-discloses-breach-after-shinyhunters-claims-patient-data-theft/
-
Berlin Refuses to Pay Hackers Who Stole Data From the City’s State Network — The Hacker News — https://thehackernews.com/2026/08/berlin-refuses-to-pay-hackers-who-stole.html
-
Cosmos EVM Flaw Exploited After Cosmos Labs Knew Every Blockchain Running It Was Vulnerable — The Hacker News — https://thehackernews.com/2026/08/cosmos-evm-flaw-exploited-after-cosmos.html
-
PaperCut releases second emergency patch for exploited flaws — Bleeping Computer — https://www.bleepingcomputer.com/news/security/papercut-releases-second-emergency-patch-for-exploited-flaws/
-
GiveWP WordPress donation plugin flaw lets hackers execute server commands — Bleeping Computer — https://www.bleepingcomputer.com/news/security/givewp-wordpress-donation-plugin-flaw-lets-hackers-execute-server-commands/
-
Hundreds of OpenAI Agents Invaded Hugging Face Servers — Dark Reading — https://www.darkreading.com/cyberattacks-data-breaches/hundreds-openai-agents-invaded-hugging-face-servers
Generated by DevSecOpsDadAttack cyber threat intelligence.