Author: blogmirnet

Multiple vulnerabilities have been discovered in Apple products, the most severe of which could allow for arbitrary code execution. Successful exploitation of the most severe of these vulnerabilities could allow for arbitrary code execution in the context of the logged on user. Depending on the privileges associated with the user, an attacker could then install programs; view, change, or delete data; or create new accounts with full user rights. Users whose accounts are configured to have fewer user rights on the system could be less impacted than those who operate with administrative user rights.

THREAT INTELLEGENCE:

There are currently no reports of these vulnerabilities being exploited in the wild.

SYSTEMS AFFECTED:

• Versions prior to iOS 18.6 and iPadOS 18.6
• Versions prior to iPadOS 17.7.9
• Versions prior to macOS Sequoia 15.6
• Versions prior to macOS Sonoma 14.7.7
• Versions prior to macOS Ventura 13.7.7
• Versions prior to watchOS 11.6
• Versions prior to tvOS 18.6
• Versions prior to visionOS 2.6

RISK:
Government:

• Large and medium government entities: High
• Small government entities: Medium

Businesses:

• Large and medium business entities: High
• Small business entities: Medium

Home users: Low

TECHNICAL SUMMARY:
Multiple vulnerabilities have been discovered in Apple products, the most severe of which could allow for arbitrary code execution. Details of the vulnerabilities are as follows:

Tactic: Execution (TA0002):

Technique: Exploitation for Client Execution (T1203):

• Running an hdiutil command may unexpectedly execute arbitrary code. (CVE-2025-43187)
• An app may be able to execute arbitrary code out of its sandbox or with certain elevated privileges. (CVE-2025-24119)

Additional lower severity vulnerabilities include:

• Passcode may be read aloud by VoiceOver. (CVE-2025-31229)
• Privacy Indicators for microphone or camera access may not be correctly displayed. (CVE-2025-43217)
• Parsing a file may lead to an unexpected app termination. (CVE-2025-43186)
• A non-privileged user may be able to modify restricted network settings. (CVE-2025-43223)
• Processing a maliciously crafted audio file may lead to memory corruption. (CVE-2025-43277)
• Processing a maliciously crafted media file may lead to unexpected app termination or corrupt process memory. (CVE-2025-43210, CVE-2025-43224, CVE-2025-43221)
• An app may be able to access user-sensitive data. (CVE-2025-43230)
• Processing maliciously crafted web content may lead to an unexpected Safari crash. (CVE-2025-43209, CVE-2025-43214, CVE-2025-43213, CVE-2025-43212, CVE-2025-43216, CVE-2025-6558, CVE-2025-24188)
• Processing a maliciously crafted image may result in disclosure of process memory. (CVE-2025-43226, CVE-2025-43215)
• Processing a file may lead to memory corruption. (CVE-2025-43202, CVE-2025-7425)
• Processing maliciously crafted web content may lead to memory corruption. (CVE-2025-7424, CVE-2025-31278, CVE-2025-31277, CVE-2025-31273)
• Remote content may be loaded even when the ‘Load Remote Images’ setting is turned off. (CVE-2025-31276)
• Processing a maliciously crafted texture may lead to unexpected app termination. (CVE-2025-43234)
• Processing a maliciously crafted file may lead to unexpected app termination. (CVE-2025-31281, CVE-2025-43254, CVE-2025-43239)
• Visiting a malicious website may lead to address bar spoofing. (CVE-2025-43228)
• Processing maliciously crafted web content may disclose sensitive user information. (CVE-2025-43227)
• Processing web content may lead to a denial-of-service. (CVE-2025-43211)
• Processing maliciously crafted web content may disclose internal states of the app. (CVE-2025-43265)
• An attacker may be able to cause unexpected app termination. (CVE-2025-43222, CVE-2025-43236)
• An app may be able to access protected user data. (CVE-2025-43220, CVE-2025-43245, CVE-2025-43198, CVE-2025-43206, CVE-2025-43185)
• An app may be able to fingerprint the user. (CVE-2025-31279)
• A remote attacker may be able to cause unexpected system termination. (CVE-2025-24224)
• An app may be able to access sensitive user data. (CVE-2025-43225, CVE-2025-43195, CVE-2025-43267, CVE-2025-43197, CVE-2025-43246)
• An app may be able to read a persistent device identifier. (CVE-2025-24220)
• An app may be able to cause a denial-of-service. (CVE-2025-43191, CVE-2025-43235, CVE-2025-43193)
• An app may be able to cause unexpected system termination. (CVE-2025-43244, CVE-2025-43255, CVE-2025-43237, CVE-2025-43238)
• An app may be able to gain root privileges. (CVE-2025-31243, CVE-2025-43249, CVE-2025-43196, CVE-2025-43256)
• A malicious app may be able to launch arbitrary binaries on a trusted device. (CVE-2025-43253)
• A malicious app may be able to gain root privileges. (CVE-2025-43248, CVE-2025-43199, CVE-2025-43188, CVE-2025-43268)
• An app may be able to break out of its sandbox. (CVE-2025-43257, CVE-2025-43261, CVE-2025-43275, CVE-2025-43266, CVE-2025-43250)
• A sandboxed process may be able to circumvent sandbox restrictions. (CVE-2025-43273, CVE-2025-43274)
• iCloud Private Relay may not activate when more than one user is logged in at the same time. (CVE-2025-43276)
• Account-driven User Enrollment may still be possible with Lockdown Mode turned on. (CVE-2025-43192)
• A sandboxed process may be able to launch any installed app. (CVE-2025-31275)
• Processing a maliciously crafted image may corrupt process memory. (CVE-2025-43264, CVE-2025-43219)
• Processing a maliciously crafted file may lead to heap corruption. (CVE-2025-31280)
• Processing a maliciously crafted USD file may disclose memory contents. (CVE-2025-43218)
• An app may gain unauthorized access to Local Network. (CVE-2025-43270)
• An app may be able to hijack entitlements granted to other privileged apps. (CVE-2025-43260)
• A malicious app with root privileges may be able to modify the contents of system files. (CVE-2025-43247)
• An app may be able to modify protected parts of the file system. (CVE-2025-43194, CVE-2025-43243)
• An app may be able to bypass certain Privacy preferences. (CVE-2025-43232)
• An app may be able to read files outside of its sandbox. (CVE-2025-43241)
• A malicious app acting as a HTTPS proxy could get access to sensitive user data. (CVE-2025-43233)
• A local attacker may gain access to Keychain items. (CVE-2025-43251)
• A malicious app may be able to read kernel memory. (CVE-2025-43189)
• Processing maliciously crafted web content may lead to universal cross site scripting. (CVE-2025-43229)
• A download’s origin may be incorrectly associated. (CVE-2025-43240)
• An attacker with physical access to a locked device may be able to view sensitive user information. (CVE-2025-43259)
• A website may be able to access sensitive user data when resolving symlinks. (CVE-2025-43252)
• A shortcut may be able to bypass sensitive Shortcuts app settings. (CVE-2025-43184)

Successful exploitation of the most severe of these vulnerabilities could allow for arbitrary code execution in the context of the logged on user. Depending on the privileges associated with the user, an attacker could then install programs; view, change, or delete data; or create new accounts with full user rights. Users whose accounts are configured to have fewer user rights on the system could be less impacted than those who operate with administrative user rights.

RECOMMENDATIONS:

We recommend the following actions be taken:

• Apply the stable channel update provided by Apple to vulnerable systems immediately after appropriate testing. (M1051: Update Software)
  • Safeguard 7.1 : Establish and Maintain a Vulnerability Management Process: Establish and maintain a documented vulnerability management process for enterprise assets. Review and update documentation annually, or when significant enterprise changes occur that could impact this Safeguard.
  • Safeguard 7.2 : Establish and Maintain a Remediation Process: Establish and maintain a risk-based remediation strategy documented in a remediation process, with monthly, or more frequent, reviews.
  • Safeguard 7.6 : Perform Automated Vulnerability Scans of Externally-Exposed Enterprise Assets: Perform automated vulnerability scans of externally-exposed enterprise assets using a SCAP-compliant vulnerability scanning tool. Perform scans on a monthly, or more frequent, basis.
  • Safeguard 7.7 : Remediate Detected Vulnerabilities: Remediate detected vulnerabilities in software through processes and tooling on a monthly, or more frequent, basis, based on the remediation process.
  • Safeguard 16.13 Conduct Application Penetration Testing: Conduct application penetration testing. For critical applications, authenticated penetration testing is better suited to finding business logic vulnerabilities than code scanning and automated security testing. Penetration testing relies on the skill of the tester to manually manipulate an application as an authenticated and unauthenticated user.
  • Safeguard 18.1 : Establish and Maintain a Penetration Testing Program: Establish and maintain a penetration testing program appropriate to the size, complexity, and maturity of the enterprise. Penetration testing program characteristics include scope, such as network, web application, Application Programming Interface (API), hosted services, and physical premise controls; frequency; limitations, such as acceptable hours, and excluded attack types; point of contact information; remediation, such as how findings will be routed internally; and retrospective requirements.
  • Safeguard 18.2 : Perform Periodic External Penetration Tests: Perform periodic external penetration tests based on program requirements, no less than annually. External penetration testing must include enterprise and environmental reconnaissance to detect exploitable information. Penetration testing requires specialized skills and experience and must be conducted through a qualified party. The testing may be clear box or opaque box.
  • Safeguard 18.3 : Remediate Penetration Test Findings: Remediate penetration test findings based on the enterprise’s policy for remediation scope and prioritization.

• Apply the Principle of Least Privilege to all systems and services. Run all software as a non-privileged user (one without administrative privileges) to diminish the effects of a successful attack. (M1026: Privileged Account Management)
  • Safeguard 4.7: Manage Default Accounts on Enterprise Assets and Software: Manage default accounts on enterprise assets and software, such as root, administrator, and other pre-configured vendor accounts. Example implementations can include: disabling default accounts or making them unusable.
  • Safeguard 5.4: Restrict Administrator Privileges to Dedicated Administrator Accounts: Restrict administrator privileges to dedicated administrator accounts on enterprise assets. Conduct general computing activities, such as internet browsing, email, and productivity suite use, from the user’s primary, non-privileged account.

• Restrict use of certain websites, block downloads/attachments, block Javascript, restrict browser extensions, etc. (M1021: Restrict Web-Based Content)
  • Safeguard 2.3: Address Unauthorized Software: Ensure that unauthorized software is either removed from use on enterprise assets or receives a documented exception. Review monthly, or more frequently.
  • Safeguard 2.7: Allowlist Authorized Scripts: Use technical controls, such as digital signatures and version control, to ensure that only authorized scripts, such as specific .ps1, .py, etc., files, are allowed to execute. Block unauthorized scripts from executing. Reassessbi-annually, or more frequently.
  • Safeguard 9.3: Maintain and Enforce Network-Based URL Filters: Enforce and update network-based URL filters to limit an enterprise asset from connecting to potentially malicious or unapproved websites. Example implementations include category-based filtering, reputation-based filtering, or through the use of block lists. Enforce filters for all enterprise assets.
  • Safeguard 9.6: Block Unnecessary File Types: Block unnecessary file types attempting to enter the enterprise’s email gateway.

• Use capabilities to detect and block conditions that may lead to or be indicative of a software exploit occurring. (M1050: Exploit Protection)
  • Safeguard 10.5: Enable Anti-Exploitation Features: Enable anti-exploitation features on enterprise assets and software, where possible, such as Microsoft® Data Execution Prevention (DEP), Windows® Defender Exploit Guard (WDEG), or Apple® System Integrity Protection (SIP) and Gatekeeper™.

• Block execution of code on a system through application control, and/or script blocking. (M1038:Execution Prevention)
  • Safeguard 2.5 : Allowlist Authorized Software: Use technical controls, such as application allowlisting, to ensure that only authorized software can execute or be accessed. Reassess bi-annually, or more frequently.
  • Safeguard 2.6 : Allowlist Authorized Libraries: Use technical controls to ensure that only authorized software libraries, such as specific .dll, .ocx, .so, etc., files, are allowed to load into a system process. Block unauthorized libraries from loading into a system process. Reassess bi-annually, or more frequently.
  • Safeguard 2.7 : Allowlist Authorized Scripts: Use technical controls, such as digital signatures and version control, to ensure that only authorized scripts, such as specific .ps1, .py, etc., files, are allowed to execute. Block unauthorized scripts from executing. Reassess bi-annually, or more frequently.

• Use capabilities to prevent suspicious behavior patterns from occurring on endpoint systems. This could include suspicious process, file, API call, etc. behavior. (M1040: Behavior Prevention on Endpoint)
  • Safeguard 13.2 : Deploy a Host-Based Intrusion Detection Solution: Deploy a host-based intrusion detection solution on enterprise assets, where appropriate and/or supported.
  • Safeguard 13.7 : Deploy a Host-Based Intrusion Prevention Solution: Deploy a host-based intrusion prevention solution on enterprise assets, where appropriate and/or supported. Example implementations include use of an Endpoint Detection and Response (EDR) client or host-based IPS agent.

REFERENCES:

Apple:

https://support.apple.com/en-us/100100

https://support.apple.com/en-us/124147

https://support.apple.com/en-us/124148

https://support.apple.com/en-us/124149

https://support.apple.com/en-us/124150

https://support.apple.com/en-us/124151

https://support.apple.com/en-us/124155

https://support.apple.com/en-us/124153

https://support.apple.com/en-us/124154

CVE:

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-6558

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-7424

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-7425

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-24119

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-24188

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-24220

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-24224

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-31229

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-31243

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-31273

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-31275

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-31276

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-31277

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-31278

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-31279

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-31280

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-31281

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43184

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43185

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43186

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43187

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43188

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43189

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43191

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43192

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43193

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43194

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43195

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43196

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43197

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43198

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43199

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43202

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43206

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43209

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43210

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43211

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43212

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43213

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43214

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43215

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43216

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43217

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43218

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43219

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43220

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43221

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43222

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43223

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43224

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43225

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43226

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43227

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43228

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43229

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43230

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43232

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43233

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43234

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43235

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43236

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43237

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43238

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43239

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43240

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43241

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43243

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43244

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43245

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43246

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43247

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43248

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43249

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43250

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43251

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43252

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43253

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43254

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43255

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43256

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43257

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43259

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43260

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43261

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43264

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43265

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43266

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43267

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43268

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43270

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43273

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43274

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43275

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43276

https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-43277

Over the last several months, security researchers have observed threat actors targeting Microsoft 365 (M365) accounts in phishing campaigns that bypass security controls. The technique allows threat actors to spoof internal M365 users and deliver emails using Microsoft Exchange Online’s Direct Send function. Direct Send lacks proper authentication and is easily exploitable, making it a desirable tactic for threat actors. Microsoft allows emails to be sent using Direct Send by default if the emails are sent from the organization’s accepted domain. Microsoft 365 administrators can implement “Reject Direct Send” to block unauthenticated Direct Send traffic at the tenant level. Direct Send may also be disabled using PowerShell. In addition to disabling Direct Send, the NJCCIC recommends following Microsoft’s Email Security Best Practices for M365, including identifying SPF/DKIM/DMARC failures, use authenticated SMTP client submissions or SMTP relay with specific IP restrictions, implement strict DMARC policies, configure SPF to hardfail, and enforce MFA for M365 accounts. For additional information and guidance, review the Varonis blog post.

Thank you to everyone who participated in the Cyber AI Profile Workshop NIST hosted this past April! This work intends to support the cybersecurity and AI communities — and the input you provided during this workshop is critical. We are working to publish a Workshop Summary that captures themes and highlights from the event. In the interim, we would like to share a preview of what we heard.

Background on the Cyber AI Profile Workshop 

As NIST began exploring the idea of a Cyber AI Profile and writing the Cybersecurity and AI Workshop Concept Paper leading up to this workshop, stakeholders told us there are several cybersecurity topics that are top of mind as businesses adopt AI. The Cyber AI Profile aims to offer practical guidance to address those topics by applying the NIST Cybersecurity Framework to address three Focus Areas…Read the Blog

The NIST National Cybersecurity Center of Excellence (NCCoE) has released a second public draft of NIST Internal Report 8536, Supply Chain Traceability: Manufacturing Meta-Framework for public comment.  

We thank everyone who submitted comments on the initial draft. Your thoughtful feedback prompted substantial revisions. In response, we are publishing this second draft to provide an opportunity for further review and input before finalizing the report.

Background

This paper presents a framework to improve traceability across complex and distributed manufacturing ecosystems. It enables structured recording, linking, and querying of traceability data across trusted repositories. This initial research is intended to explore approaches that may support stakeholders in verifying product provenance, meeting contractual obligations, and assessing supply chain integrity.

This framework builds on previous NIST research (NIST IR 8419) and incorporates insight and feedback from industry, standards bodies, and academia. It is designed to enhance national security, economic resilience, and supply chain risk management, particularly across manufacturing and other critical infrastructure sectors.

We invite and encourage those interested to review and comment on this draft. 

Submit Your Comments 

The public comment period for this draft is open through September 1, 2025. Visit the project page for a copy of the draft and instructions for submitting comments. We value and welcome your input and look forward to your comments. 

Get Engaged 

You can continue to help shape and contribute to this and future projects by joining the NCCoE’s Blockchain Community of Interest. Visit our project page to join.

Comment Now

The NIST National Cybersecurity Center of Excellence (NCCoE) has re-issued NIST Internal Report (IR) 8579, Developing the NCCoE Chatbot: Technical and Security Learnings from the Initial Implementation. Originally published in June, the document was revised to improve the document’s demonstration of the enhanced abilities of an RAG-based LLM tool over a generic LLM.

The public comment period for the publication has been extended and will close on September 11, 2025.

The NCCoE identified a potential application for a chatbot to support its mission and developed a secure, internal-use chatbot to assist NCCoE staff with searching and summarizing cybersecurity guidelines tailored to specific audiences or use cases.

The chatbot was built using retrieval-augmented generation (RAG)-based LLM technology. This approach combines techniques from information retrieval and natural language generation, enabling the chatbot to provide more focused, contextually relevant responses by leveraging a repository of cybersecurity knowledge, including previous NCCoE publications. Compared to search engines, LLM-based chatbots provide more contextually relevant and precise responses by understanding the nuances of natural language queries.

This report provides a point-in-time examination of the NCCoE Chatbot, outlining the NCCoE’s approach to developing the tool, as well as the NCCoE’s response to specific security challenges. In addition, this report provides an overview of the chatbot and its supporting technologies so that other organizations might consider the benefits of their use.

We encourage you to review this document and provide comments by September 11, 2025. If you have any questions, please email the team at nlp-nccoe@nist.gov.

Comment Now!

Russian nation-state cyber actors remain among the most persistent, capable, and strategically aligned threats in the global cyber landscape. Their operations routinely target Western governments, critical infrastructure, defense contractors, and political institutions, using a combination of cyber espionage, advanced malware, and living-off-the-land techniques. These campaigns are characterized by prolonged presence, covert operations, and strategic alignment with Russia’s geopolitical goals, making them especially challenging to detect and mitigate.

A key takeaway from this analysis is that Russian cyber activity is not merely opportunistic; it is intentional and ongoing, designed to maintain persistent access for future disruption, espionage, or influence efforts. Russia’s ability to blend into trusted environments and exploit legitimate platforms significantly raises the threat to national security, essential services, and the private sector.

In the broader risk context, Russia exemplifies the intersection of cyber operations and hybrid warfare, where espionage, sabotage, and geopolitical strategy align. Organizations must prioritize detection and hardening against known Russian tactics, techniques, and procedures (TTPs), particularly those involving credential theft, cloud environment exploitation, and lateral movement, as these remain core elements of Russia’s offensive cyber strategy.

Key Points

• Who: Russian state-sponsored actors, including APT28 (Fancy Bear), APT29 (Cozy Bear), Sandworm, and Gamaredon.
• What: Conducted cyber espionage, prepositioned in networks, and launched disruptive operations aligned with Russian geopolitical goals.
• How: Used spearphishing, credential harvesting, zero-day exploits, cloud service abuse, and stealthy tools like living-off-the-land binaries (LOLBins) and custom malware.
• Why it matters: These operations are designed not only for data theft but also to maintain persistent access, evade detection, and retain the ability to disrupt critical infrastructure or manipulate information at strategic moments.

Risk Assessment

The NJCCIC has assessed that Russian state-sponsored cyber actors represent a persistent and evolving threat with the demonstrated capability and intent to conduct both intelligence-gathering and disruptive operations. Their activities seriously threaten national security, essential public services, and critical industries.

Recently, these groups have changed tactics by increasingly targeting cloud infrastructure and identity management platforms like Microsoft 365. This evolution includes deploying new malware variants designed to evade traditional detection methods, allowing them to infiltrate previously considered lower-risk environments.

Russian actors have a well-documented history of targeting entities across government, critical infrastructure, healthcare, defense, and election systems. The likelihood of exposure or compromise is significantly higher if environments include legacy technology or commonly used third-party platforms. The combination of their strategic intent, advanced techniques, and a sector’s relevance makes proactive defense and visibility into these threat vectors essential.

Timeline of Activity

Attribution	Start Date	End Date	Location	Sector	Activity
APT 28 (Fancy Bear)	2007	Present	Global (US, NATO, and Europe)	Government, Military, and Media	Espionage, election interference, credential theft, hack and leak operations
APT 29 (Cozy Bear)	2008	Present	Global	Government, Think Tanks, and NGOs	Cyber espionage, credential harvesting, and supply chain compromises
Turla (Venomous Bear)	2008	Present	Europe and the Middle East	Government, Military, and Research	Long-term espionage, custom malware, and hijacking satellite infrastructure
Berserk Bear (Energetic Bear)	2010	Present	Europe and the US	Energy and ICS	ICS reconnaissance, credential harvesting, and infrastructure targeting
Gamaredon (Primitive Bear)	2013	Present	Ukraine	Government	Phishing, malware, and espionage
Sandworm Team	2014	Present	Ukraine and Global	Energy, ICS, and Telecom	Destructive Malware (NotPetya), ICS Attacks, and Wiper Malware
Evil Corp (Indrik Spider)	2014	Present	Global	Finance, Retail, and Healthcare	Banking trojans and ransomware deployments
Star Blizzard	2017	Present	UK, US, and Ukraine	Academia and Government	Credential theft and phishing
Shuckworm	2017	Present	Ukraine	Government	Using old malware
Nobelium (Subset of APT 29)	2020	2021	Global	IT, Government, and Supply Chain	SolarWinds compromise, cloud service, and lateral movement
NoName057(16)	2022	Present	Europe, NATO-aligned States	Government and Media	DDoS ops focused on Finland, Latvia, and Poland. Activity spikes during major NATO summits and political votes.
KillNet (pro-Russian hacktivist)	2022	Present	NATO, EU, and US	Government, Healthcare, and Transportation	DDoS attack on US hospital, Polish rail, and airport websites.
Fancy Bear	2022	Present	Europe and the US	Critical Infrastructure	Compromise of CCTV at transit hubs to monitor Western Military Aid
RomCom	2024	Present	Europe and the US	Government, Defense, Energy, Pharma, and Legal	Zero-click exploits via Firefox (CVE-2024-9680) and Windows (CVE-2024-49039), deploying RATs

Capabilities

• Advanced Persistent Threats (APTs): Russia maintains multiple state-sponsored groups (APT28, APT29, Sandworm) capable of long-term, covert operations across sectors.
• Malware Toolsets: Known for deploying malware such as NotPetya, Snake, Drovorub, WellMess, and CosmicDuke.
• Tradecraft: Uses living-off-the-land binaries (LOLBins), legitimate credentials, and cloud platform abuse (Microsoft 365).
• Operational Focus: Supports espionage, prepositioning in critical infrastructure, disinformation, and destructive attacks aligned with geopolitical goals.
• Supply Chain Intrusions: Demonstrated ability to compromise software providers and abuse trusted relationships (SolarWinds, Ukrainian tax software).
• ICS/OT Disruption: Proven capacity to target and impact operational technology systems (BlackEnergy, Industroyer).
• Information Warfare: Coordinates cyber operations with disinformation and influence campaigns targeting elections and public opinion.

Key Intelligence Gaps

The NJCCIC has assessed that several intelligence gaps limit a full Russian cyber threat landscape assessment. One of the most pressing concerns is whether Russian threat actors are currently positioning themselves within networks to carry out future physical or disruptive attacks, particularly amid ongoing geopolitical tensions and hybrid warfare strategies.

Another critical intelligence gap concerns developing and deploying new malware strains, particularly malware designed to target industrial control systems (ICS), satellite infrastructure, or emerging space technologies, where visibility remains low.

The degree of coordination between Russian intelligence services and non-state criminal cyber groups (such as ransomware-as-a-service operators) also remains unclear. It is essential to determine how much operational freedom these groups have and whether they receive explicit support or protection from the state. There is also a lack of clarity about whether Russia’s agencies are working together or against each other. 

Additionally, there is limited insight into evolving Russian tactics for evading detection, especially in hybrid and multi-cloud environments, where traditional monitoring tools may fall short.

The impact of economic sanctions and wartime constraints on Russia’s cyber strategy and operational capabilities is also poorly understood, leaving a gap in forecasting future shifts in tactics or targeting priorities.

Known Threat Groups

Military Intelligence (GRU)

• APT28: Fancy Bear, Sofacy, STRONTIUM, Sednit
• Sandworm Team: BlackEnergy, Voodoo Bear, TeleBots
• Gamaredon Group: Primitive Bear
• InvisiMole: Occasionally linked with Gamaredon

Foreign Intelligence Service (SVR)

• APT29: Cozy Bear, The Dukes, Yttrium, Nobelium

Federal Security Service (FSB)

• Turla: Snake, Uroburos, Venomous Bear
• Berserk Bear: Energetic Bear, Crouching Yeti, Dragonfly
• Krypton: Suspected ties to Turla

Elevate your AI model performance by mastering RAG techniques   Improve customer experiences by making your AI models more accurate, reliable, and versatile with retrieval-augmented generation (RAG) architecture. Whether you’re a RAG novice or professional, expand your skill set with the Code With Confidence: Build Your Retrieval-Augmented Generation Skills video series. This six-part series starts with the basics of RAG and knowledge retrieval, search indexing, and vector search. Then, explore sessions on how to: Build the ultimate retrieval system for RAG.Optimize your vector index for scale.Index and retrieve multimodal and other data.Tap into best practices from real-world use cases.Customize AI models with RAG and fine-tuning techniques.

Watch the series

Explore more resources for AI customization

Quickly get started using RAG   Find out how RAG compares with fine-tuning and other AI customization techniques. Also, view a demo of how to streamline the process with Azure AI.   Watch the webinar >   Increase business value with fine-tuning   Discover how to use fine-tuning to train AI models on task-specific data, allowing you to efficiently scale AI capabilities that deliver better business outcomes.   Read the e-book >

Meet Researcher and Analyst in Microsoft 365 Copilot: July 2025 edition This month, we’re introducing Researcher and Analyst, two powerful new additions to Microsoft 365 Copilot. These features use advanced reasoning and data from your work and the web to analyze, infer, and synthesize information to deliver deeper and more accurate insights. Let’s take a look at what’s new and ready for you to try.

Introducing Researcher: Deep research, done in minutes Recommended for: Legal Marketing Sales​​​​​​​ Available on: Windows, Mac and Mobile Need to tackle complex research? Researcher is here to help. It works across your content and trusted sources to gather, connect, and make sense of information. Whether you’re exploring market trends, prepping for client meetings, or analyzing lengthy contracts, Researcher helps you make informed decisions faster. Try in Researcher: “What are the latest consumer trends in [industry or market] and how are competitors reacting? Include ways to improve our strategy based on [marketing_plan.docx].” “Build a sales brief on [business name] with recent updates and risks.” “Summarize [new law], compare to our policy, and flag compliance gaps.” ​​​​​​​Watch Researcher in action​​​​​​​

Introducing Analyst: Data analysis at your fingertips Recommended for: Finance Marketing Sales​​​​​​ Available on: Windows, Mac and Mobile Analyst helps you think like a data scientist, using advanced reasoning and powerful data analysis to turn complex data into clear insights. Use it to forecast demand, visualize customer trends, or surface trends across your files, so you can quickly move from data to decisions. Try in Analyst: “Forecast monthly expenses by department for the remainder of the year and chart the trend.” “Using [revenue_report.csv], calculate YoY growth, segment by customer tier, and flag any accounts with declining spend.” “Which channels drove the most conversions in [campaign_data.xlsx]? Summarize ROI and next steps.” ​​​​​​​Watch Analyst in action​​​​​​​ Researcher and Analyst are now generally available in Copilot. Availability may vary based on your organization’s settings.

Explore new features in Copilot   Personalize how you work with Copilot Available on: Web Copilot now adapts to your preferences and past interactions for more relevant responses.  Set your preferences: Tailor your experience with custom instructions that help guide Copilot’s responses. Personalized responses: Copilot remembers key details across chats, so you don’t have to provide context repeatedly. You’re in control: Memories update automatically and can be viewed or removed at any time. To create custom instructions or manage what Copilot remembers, select … Copilot chats and more > Settings. ​​​​​​​ Hear Copilot responses out loud Available on: Web, Windows Copilot can now read responses aloud so you can listen while you work.  Audio narration: Hear longer responses read aloud so you can stay focused without needing to read every word. To get started, select the speaker icon from any Copilot response.  ​​​​​​​ Try in Copilot : “Summarize my week in the style of a comedy roast” Try in Copilot : “Help me prepare for my day”

Helpful resources to make the most of ​​​​​​​​​​​​​​Copilot Upskill on Copilot with curated learning paths: Visit Copilot Academy for structured learning with resources and hands-on activities by Microsoft experts. Maximize skills with free Microsoft 365 Copilot training: Register now to gain hands-on experience, discover role-specific tips, and learn from real-world examples. Stay productive on-the-go: Download the Microsoft 365 Copilot app and sign in with your work account to have Copilot at your fingertips.

In March 2025, based on stakeholders’ input, NIST announced its AI Standards Zero Drafts pilot project to accelerate the creation of AI standards and expand participation in AI standards development.

After the pilot’s launch, NIST identified two initial topics based on community feedback: AI testing, evaluation, verification, and validation (TEVV) and documentation of AI models and datasets. NIST has now released a detailed outline proposing a direction and structure for the forthcoming TEVV zero draft. 

The proposed document is intended as an overarching framework that helps AI practitioners design appropriate TEVV approaches for specific systems and cases. It is envisioned as integrating with current and future ISO/IEC standards on AI testing while also providing a suitable backdrop to facilitate their application.

Read the Outline

Feedback on the outline is welcome. Input received by September 12, 2025, will be considered for NIST’s initial public draft of the text; input received later will be considered for incorporation into subsequent iterations.

In December 2024, NIST’s Crypto Publication Review Board initiated a review of the following Special Publications (SP):

• NIST SP 800-56Ar3, Recommendation for Pair-Wise Key-Establishment Schemes Using Discrete Logarithm Cryptography (2018)
• NIST SP 800-56Br2, Recommendation for Pair-Wise Key-Establishment Using Integer Factorization Cryptography (2019)
• NIST SP 800-56Cr2, Recommendation for Key-Derivation Methods in Key-Establishment Schemes (2020)

In response, NIST received public comments.

NIST proposes to:

• update SP 800-56Ar3
• reaffirm SP 800-56Br2
• revise SP 800-56Cr2

Submit comments on this decision by September 15, 2025 to cryptopubreviewboard@nist.gov with “Comments on SP 800-56 Decision Proposal” in the subject line. Comments received in response to this request will be posted on the Crypto Publication Review Project site after the due date. Submitters’ names and affiliations (when provided) will be included, while contact information will be removed. See the project site for additional information about the review process. 

Read More