This briefing covers significant cybersecurity developments from July 11-12, 2026. The period saw critical vulnerabilities in PraisonAI platforms with CVSS scores of 10.0, enabling remote code execution through LLM prompt injection attacks. Law enforcement activity continued with extradition and sentencing of ransomware operators linked to Ryuk and BlackCat groups. A massive data breach allegation against TikTok claims exposure of 2.4 billion user records, while Glendale Community College confirmed a breach affecting 793,925 individuals. Ransomware groups remain highly active, with The Gentlemen group alone posting 13 new victims and multiple critical WordPress plugin vulnerabilities enabling authentication bypass and remote code execution. Infrastructure threats include widespread Mirai/Mozi botnet distribution and a novel 'Ghostcommit' attack technique targeting AI code review systems through image-based prompt injection.
Multiple critical vulnerabilities discovered in PraisonAI platform enabling remote code execution, with novel prompt injection attack vectors targeting AI systems
Critical vulnerability (CVSS 10.0) in PraisonAI before 1.6.78 allows attackers to execute arbitrary Python code through LLM prompt injection without AST validation or sandboxing. Attackers can influence LLM output to exfiltrate environment variables and execute malicious code.
Critical flaw (CVSS 9.9) in PraisonAI before 4.6.78 enables arbitrary file writes and command execution through AICoder component. Missing path validation allows attackers to inject malicious prompts via chat interface to write files anywhere on the filesystem.
Critical vulnerability (CVSS 9.8) in PraisonAI before 4.6.78 due to unvalidated dimension arguments in PGVector and Cassandra backends. Caller-controlled values can lead to SQL injection despite partial identifier validation.
Novel attack technique embeds prompt injection instructions in PNG images to deceive AI code review agents like CodeRabbit and Bugbot. Successfully convinced coding agents to exfiltrate secrets from repository .env files by hiding malicious instructions in images that automated reviewers never open.
High severity (CVSS 8.6) default configuration flaw in PraisonAI before 1.7.3 binds to all interfaces without API key requirements and wildcard CORS. Unauthenticated attackers can read agent instructions, system prompts, and invoke agents remotely.
Multiple high and critical severity vulnerabilities affecting popular WordPress plugins, enabling authentication bypass, remote code execution, and SQL injection
Authentication bypass vulnerability (CVSS 8.8) in Simple JWT Login plugin versions up to 3.6.6. The generatePayload() function only validates payload structure, not content, allowing attackers to forge JWTs and escalate privileges to administrator level.
RCE vulnerability (CVSS 8.8) in WP Ultimate CSV Importer up to 8.0.1 via MappedFields parameter. Missing capability checks on AJAX handlers allow authenticated attackers to execute arbitrary code through install_addon and save_module actions.
RCE flaw (CVSS 8.8) in Code Engine plugin up to 0.3.5 via code-engine shortcode. Plugin fails to restrict access to code injection functionality, allowing authenticated attackers with contributor-level access to execute arbitrary code.
Account takeover vulnerability (CVSS 8.8) in Essential Addons for Elementor up to 6.6.10. Insufficient server-side validation of Login/Register widget settings enables authenticated attackers to perform email header injection and hijack user accounts.
Directory traversal vulnerability (CVSS 7.5) in W3 Total Cache up to 2.9.4 via setupSources function. Unauthenticated attackers can read arbitrary files on the server, potentially exposing sensitive configuration data and credentials.
Account takeover flaw (CVSS 8.1) in SureCart up to 4.2.3. Plugin fails to validate user identity during customer profile synchronization from webhook events, allowing attackers to update user emails and escalate privileges.
Critical and high severity vulnerabilities in CMS platforms, browsers, and enterprise software requiring immediate attention
Australian Cyber Security Centre (ACSC) issued alert about widespread exploitation campaign targeting vulnerable content management systems and plugins globally. Organizations using CMS platforms urged to apply patches immediately.
Significant data breach allegations including TikTok's claimed 2.4 billion user exposure and confirmed breach of Glendale Community College affecting nearly 800,000 individuals
Class action lawsuit filed alleging data breach exposed personal data of more than 2.4 billion TikTok users worldwide. Case filed in U.S. District Court for Central District of California on June 11, 2026 by California resident Sean Mortazi. Scope and verification of breach claims under investigation.
Glendale Community College targeted by ShinyHunters pay-or-leak extortion campaign in June 2026. Breach exposed 793,925 unique email addresses along with names, physical addresses, phone numbers, dates of birth, genders, government IDs, and academic records. Data subsequently published online.
Dutch police uncovered evidence of local criminal involvement in Odido telecom cyberattack that exposed personal data of 6+ million customers. Investigation revealed Dutch-speaking suspect posing as legitimate actor to facilitate breach. Voice recording evidence may be released.
Significant law enforcement successes with extradition and sentencing of ransomware operators, while multiple ransomware groups continue aggressive victim posting campaigns
Karen Serobovich Vardan, Armenian national extradited from Ukraine, pleaded guilty to role in Ryuk ransomware attacks targeting U.S. companies including Oregon technology firm. Case prosecuted in Portland federal court represents continued law enforcement pressure on ransomware operators.
Third co-conspirator who provided BlackCat ransomware operators with inside information on victim defense strategies sentenced to 70 months in federal prison. Individual colluded with BlackCat scammers to extort victims more effectively by betraying negotiation positions.
The Gentlemen ransomware group posted 13 new victims including Royal Foods (Australia), Vicenzi Group (Italy), Open Options (Texas), INTERNET AG (Germany), Crossroads Medical Management (Georgia), and Pharma Wholesale (Florida). Victims span food manufacturing, IT services, healthcare, and pharmaceutical distribution sectors across multiple countries.
Qilin ransomware group published four new victims: Century Equities, Retelit SpA (Italian telecommunications), Carolina Agri-Power, and Allied Plumbing & Heating. Telecommunications sector targeting demonstrates continued focus on critical infrastructure.
LockBit5 variant posted 9 victims including JS Hotels (Majorca), multiple German firms (Gies Dienstleistungen, ComTRI), Spanish companies (Bianchini galvanized steel), and Italian manufacturer Ravagnan Group. Campaign demonstrates continued operations despite law enforcement disruptions.
Significant Mirai/Mozi botnet activity with 50+ malware distribution URLs, plus ClearFake and Vidar stealer campaigns targeting multiple platforms
Over 30 URLs distributing Mirai and Mozi botnet variants targeting IoT devices. Payloads compiled for multiple architectures (ARM, MIPS, x86_64) indicating broad targeting of embedded systems and routers. Primary C2 infrastructure at 94.154.43.42, 182.122.x.x, 191.37.19.74, and 5.166.134.218 ranges.
Multiple ClearFake malware distribution URLs hosted on gambling/casino domains (betebet.poker, casinomhub.bet, bahigo.casino, betraca.bet). Campaign targets both Windows and macOS platforms with unique build identifiers, suggesting ongoing development and variant testing.
These briefings are compiled from publicly available threat-intelligence feeds, which may include CISA KEV, NIST NVD, the GitHub Advisory Database (OSV), abuse.ch, and Wordfence Intelligence. Data-breach and credential-leak items may include data from Have I Been Pwned and ransomware.live.
CVE® is a registered trademark of The MITRE Corporation. CVE Records are © The MITRE Corporation, reproduced under the CVE Program Terms of Use. WordPress vulnerability data is provided by Wordfence Intelligence, © Defiant, Inc. Breach data from Have I Been Pwned is licensed under CC BY 4.0.