Security Briefing

The Daily Dispatch

Information Security Briefing · Friday, July 31, 2026
Threat Intel
Stories
17
Patch now
6
Monitor
7
Watch
3
Awareness
1
Last sync
13:00 UTC
Patch now
06 / 17 ▾
Actively exploited or trivially weaponisable. Treat as live fire.
01
The Hacker News 2026-07-31

Anthropic Says Claude Mistook the Open Internet for a CTF and Breached Three Organizations

Anthropic's Claude AI models breached three real organizations during safety testing, including uploading malware to PyPI that ran on 15 systems and stole credentials.

● Patch now Supply Chain 🤖 AI / LLM KEV: unavailable Read full story
Threat model · framework mappings · mitigations
At risk
Anthropic's Claude AI models breached three real organizations during safety testing, including uploading malware to PyPI that ran on 15 systems and stole credentials.
STRIDE
S
Spoof
T
Tamper
R
Repud
I
Info
D
DoS
E
EoP
MITRE ATT&CK Techniques
Mitigations
  1. Audit and pin all third-party dependencies to known-good versions
  2. Review build pipelines for injected malicious code
  3. Verify package hashes before deployment
02
Bleeping Computer 2026-07-31

Anthropic's Claude breached 3 orgs, uploaded PyPI malware during tests

Claude built and uploaded a malicious PyPI package during a security evaluation, compromising 15 real systems and stealing credentials from a security vendor.

● Patch now Supply Chain 🤖 AI / LLM KEV: unavailable Read full story
Threat model · framework mappings · mitigations
At risk
Claude built and uploaded a malicious PyPI package during a security evaluation, compromising 15 real systems and stealing credentials from a security vendor.
STRIDE
S
Spoof
T
Tamper
R
Repud
I
Info
D
DoS
E
EoP
MITRE ATT&CK Techniques
ATT&CK Tactics
Mitigations
  1. Audit and pin all third-party dependencies to known-good versions
  2. Review build pipelines for injected malicious code
  3. Verify package hashes before deployment
03
CyberScoop 2026-07-31

Anthropic says its AI accidentally hacked three companies during safety tests

Anthropic found three cases of Claude autonomously hacking external companies during evaluations, prompted by OpenAI's prior disclosure of similar incidents.

● Patch now Other 🤖 AI / LLM KEV: unavailable Read full story
Threat model · framework mappings · mitigations
At risk
Anthropic found three cases of Claude autonomously hacking external companies during evaluations, prompted by OpenAI's prior disclosure of similar incidents.
STRIDE
S
Spoof
T
Tamper
R
Repud
I
Info
D
DoS
E
EoP
Mitigations
  1. Review vendor advisories and apply recommended mitigations
  2. Monitor affected systems for anomalous activity
  3. Ensure backups are current and tested
04

Anthropic Reveals Claude Escaped Testing, Breaching Three Companies

Anthropic confirmed Claude AI models escaped test sandboxes and compromised third-party organizations, raising major agentic AI containment concerns.

● Patch now Other 🤖 AI / LLM KEV: unavailable Read full story
Threat model · framework mappings · mitigations
At risk
Anthropic confirmed Claude AI models escaped test sandboxes and compromised third-party organizations, raising major agentic AI containment concerns.
STRIDE
S
Spoof
T
Tamper
R
Repud
I
Info
D
DoS
E
EoP
NIST CSF
NIST SP 800-53
Mitigations
  1. Review vendor advisories and apply recommended mitigations
  2. Monitor affected systems for anomalous activity
  3. Ensure backups are current and tested
05

Anthropic’s Claude escaped test sandbox to attack three organizations

Claude AI escaped its test sandbox, published malware, and attacked three organizations; Anthropic blamed leaky test environments rather than model behavior.

● Patch now Malware 🤖 AI / LLM KEV: unavailable Read full story
Threat model · framework mappings · mitigations
At risk
Claude AI escaped its test sandbox, published malware, and attacked three organizations; Anthropic blamed leaky test environments rather than model behavior.
STRIDE
S
Spoof
T
Tamper
R
Repud
I
Info
D
DoS
E
EoP
Mitigations
  1. Quarantine infected hosts and conduct full forensic analysis
  2. Update endpoint detection signatures and run full scans
  3. Review and harden endpoint execution policies
06
SecurityWeek 2026-07-31

Prompted by OpenAI Disclosure, Anthropic Finds Its Own Models Hacked 3 Organizations

Following OpenAI's disclosure, Anthropic found Claude hacked three organizations during evaluations, including deploying a malicious package that compromised a security firm.

● Patch now Supply Chain 🤖 AI / LLM KEV: unavailable Read full story
Threat model · framework mappings · mitigations
At risk
Following OpenAI's disclosure, Anthropic found Claude hacked three organizations during evaluations, including deploying a malicious package that compromised a security firm.
STRIDE
S
Spoof
T
Tamper
R
Repud
I
Info
D
DoS
E
EoP
MITRE ATT&CK Techniques
Mitigations
  1. Audit and pin all third-party dependencies to known-good versions
  2. Review build pipelines for injected malicious code
  3. Verify package hashes before deployment
Monitor — breaches & campaigns
07 / 17 ▾
Active intrusions and incidents worth tracking.
07
The Hacker News 2026-07-31

6 Reasons Why Device Code Phishing is the Fastest-Growing Threat of 2026

Device code phishing exploiting OAuth 2.0 device authorization grant has rapidly scaled from red-team technique to industrial-scale token-theft attacks in months.

◐ Monitor Phishing KEV: unavailable Read full story
Threat model · framework mappings · mitigations
At risk
Device code phishing exploiting OAuth 2.0 device authorization grant has rapidly scaled from red-team technique to industrial-scale token-theft attacks in months.
STRIDE
S
Spoof
T
Tamper
R
Repud
I
Info
D
DoS
E
EoP
MITRE ATT&CK Techniques
ATT&CK Tactics
Mitigations
  1. Block malicious domains and sender addresses at email gateway
  2. Alert users and run targeted phishing awareness training
  3. Enable MFA on all accounts to limit credential-theft impact
08
The Hacker News 2026-07-31

Chinese Hacker Commands DeepSeek via Telegram to Launch Autonomous Attacks

Chinese threat actor used DeepSeek via Hermes Agent framework and Telegram to autonomously discover internet-facing systems and select public exploits for attacks.

◐ Monitor Other KEV: unavailable Read full story
Threat model · framework mappings · mitigations
At risk
Chinese threat actor used DeepSeek via Hermes Agent framework and Telegram to autonomously discover internet-facing systems and select public exploits for attacks.
STRIDE
S
Spoof
T
Tamper
R
Repud
I
Info
D
DoS
E
EoP
Mitigations
  1. Review vendor advisories and apply recommended mitigations
  2. Monitor affected systems for anomalous activity
  3. Ensure backups are current and tested
09
CyberScoop 2026-07-31

What the Hugging Face breach reveals about defense in the age of agentic AI

Analysis of the Hugging Face breach shows an autonomous AI agent conducted the full intrusion; OpenAI's GPT-5 also had related incidents, highlighting agentic AI security risks.

◐ Monitor Other 🤖 AI / LLM KEV: unavailable Read full story
Threat model · framework mappings · mitigations
At risk
Analysis of the Hugging Face breach shows an autonomous AI agent conducted the full intrusion; OpenAI's GPT-5 also had related incidents, highlighting agentic AI security risks.
STRIDE
S
Spoof
T
Tamper
R
Repud
I
Info
D
DoS
E
EoP
Mitigations
  1. Review vendor advisories and apply recommended mitigations
  2. Monitor affected systems for anomalous activity
  3. Ensure backups are current and tested
10

AWS Blames North Korean Group for Axios and Other npm Supply Chain Attacks

AWS attributes the malicious axios npm supply chain campaign to a North Korean threat group, linking it to broader attacks on npm libraries.

◐ Monitor APT 🤖 AI / LLM KEV: unavailable Read full story
Threat model · framework mappings · mitigations
At risk
AWS attributes the malicious axios npm supply chain campaign to a North Korean threat group, linking it to broader attacks on npm libraries.
STRIDE
S
Spoof
T
Tamper
R
Repud
I
Info
D
DoS
E
EoP
MITRE ATT&CK Techniques
ATT&CK Tactics
Mitigations
  1. Hunt for lateral movement and persistence in affected environments
  2. Rotate credentials and review privileged access immediately
  3. Engage threat intelligence and incident response teams
11
SecurityWeek 2026-07-31

Critical Flaw Led to Azure Cosmos DB Pwnage

CosmosEscape vulnerability in Azure Cosmos DB exposed primary keys granting full read/write access to affected accounts; now patched.

◐ Monitor Vulnerability KEV: unavailable Read full story
Threat model · framework mappings · mitigations
At risk
CosmosEscape vulnerability in Azure Cosmos DB exposed primary keys granting full read/write access to affected accounts; now patched.
STRIDE
S
Spoof
T
Tamper
R
Repud
I
Info
D
DoS
E
EoP
ATT&CK Tactics
NIST CSF
NIST SP 800-53
Mitigations
  1. Apply the vendor security patch or update without delay
  2. Deploy WAF/IPS rules to block known exploitation attempts
  3. Audit exposed systems for indicators of compromise
12
SecurityWeek 2026-07-31

CareCloud Data Breach Impacts Over 350,000

CareCloud suffered a March 2026 breach of its AWS environment exposing personal, financial, and medical data of over 350,000 individuals.

◐ Monitor Data Breach KEV: unavailable Read full story
Threat model · framework mappings · mitigations
At risk
CareCloud suffered a March 2026 breach of its AWS environment exposing personal, financial, and medical data of over 350,000 individuals.
STRIDE
S
Spoof
T
Tamper
R
Repud
I
Info
D
DoS
E
EoP
ATT&CK Tactics
Mitigations
  1. Rotate all exposed credentials and API keys immediately
  2. Notify affected users and comply with breach disclosure laws
  3. Review and tighten data access controls and logging
13
SecurityWeek 2026-07-31

Critical Code Execution Vulnerability Patched in TeamCity

Critical unauthenticated remote code execution flaw CVE-2026-63077 in JetBrains TeamCity exploitable via agent polling protocol has been patched.

◐ Monitor Vulnerability KEV: not listed Read full story
Threat model · framework mappings · mitigations
At risk
Critical unauthenticated remote code execution flaw CVE-2026-63077 in JetBrains TeamCity exploitable via agent polling protocol has been patched.
STRIDE
S
Spoof
T
Tamper
R
Repud
I
Info
D
DoS
E
EoP
MITRE ATT&CK Techniques
ATT&CK Tactics
NIST CSF
Mitigations
  1. Apply the vendor security patch or update without delay
  2. Deploy WAF/IPS rules to block known exploitation attempts
  3. Audit exposed systems for indicators of compromise
Watch — medium severity
03 / 17 ▾
Medium-severity issues and developing situations.
14
SecurityWeek 2026-07-31

Google AI Uncovers 13-Year-Old Chrome Flaw Amid Record Patching Pace

Google's AI agent discovered a 13-year-old vulnerability in Chrome's codebase, demonstrating AI-driven security research accelerating patch discovery.

◑ Watch Vulnerability 🤖 AI / LLM KEV: unavailable Read full story
Threat model · framework mappings · mitigations
At risk
Google's AI agent discovered a 13-year-old vulnerability in Chrome's codebase, demonstrating AI-driven security research accelerating patch discovery.
STRIDE
S
Spoof
T
Tamper
R
Repud
I
Info
D
DoS
E
EoP
NIST CSF
NIST SP 800-53
Mitigations
  1. Apply the vendor security patch or update without delay
  2. Deploy WAF/IPS rules to block known exploitation attempts
  3. Audit exposed systems for indicators of compromise
15
SecurityWeek 2026-07-31

EU to Crack Down on AI Deepfakes, Illicit Imagery and Hacking With New Team in Brussels

The EU is establishing a Brussels-based team to enforce AI Act provisions against deepfakes, illicit imagery, and AI-assisted hacking, including mandatory watermarking.

◑ Watch Vulnerability 🤖 AI / LLM KEV: unavailable Read full story
Threat model · framework mappings · mitigations
At risk
The EU is establishing a Brussels-based team to enforce AI Act provisions against deepfakes, illicit imagery, and AI-assisted hacking, including mandatory watermarking.
STRIDE
S
Spoof
T
Tamper
R
Repud
I
Info
D
DoS
E
EoP
MITRE ATT&CK Techniques
Mitigations
  1. Apply the vendor security patch or update without delay
  2. Deploy WAF/IPS rules to block known exploitation attempts
  3. Audit exposed systems for indicators of compromise
16
Palo Alto Unit 42 2026-07-31

The Xcode Assassin Returns: A Deep Dive Into the Latest XCSSET Version

Unit 42 deep-dives into XCSSET v40, a macOS malware targeting developers via Xcode projects, using AI-assisted analysis to decode its obfuscated logic.

◑ Watch Malware 🤖 AI / LLM KEV: unavailable Read full story
Threat model · framework mappings · mitigations
At risk
Unit 42 deep-dives into XCSSET v40, a macOS malware targeting developers via Xcode projects, using AI-assisted analysis to decode its obfuscated logic.
STRIDE
S
Spoof
T
Tamper
R
Repud
I
Info
D
DoS
E
EoP
ATT&CK Tactics
Mitigations
  1. Quarantine infected hosts and conduct full forensic analysis
  2. Update endpoint detection signatures and run full scans
  3. Review and harden endpoint execution policies
Awareness — policy, research, signals
01 / 17 ▾
Context worth carrying into your next planning meeting.