Hello, stay ahead with CISO Brief 🚀

🔒 SafePal reports a data breach affecting about 39,798 customers after an authorization flaw in an order-tracking plug-in was exploited to steal order information. The exposed data includes names, emails, shipping addresses, phone numbers, and purchase details for orders placed between March 2, 2025, and April 11, 2026. SafePal says sensitive wallet credentials, payment card numbers, and government IDs were not exposed and that it has fixed the vulnerability, notified affected customers, and launched a verification tool. A threat actor is now claiming to sell the stolen data on a cybercrime forum, and the company warns of targeted phishing and social engineering attempts.

🔴 Anthropic confirmed a major outage beginning August 16, 2026, causing login failures and degraded performance across Claude.ai, Claude Code, and Claude Cowork. The company first reported authentication issues at 21:58 UTC, then noted broader performance disruptions at 22:07 UTC. Affected users may experience sign-in failures, loading issues, or incomplete requests while Claude Console and the Claude API remain operational.

🛡️ A new macOS infostealer called AmnesiaStealer uses ClickFix campaigns to deliver a Mach-O payload inside a password-protected archive. It copies Chromium profiles and launches hidden, headless browser instances to preserve authentication state while enabling remote operator control. The malware exfiltrates passwords, keychain items, crypto wallets, browser data across 16 Chromium-based browsers, and streams live screencasts and input via WebSocket channels. Researchers at Jamf warn the module abuses the Chrome DevTools Protocol to let attackers navigate and act in victims' authenticated sessions.

🔍 Anthropic announced it will apply invisible watermarking to Claude-generated text worldwide to comply with the EU AI Act. The watermark modifies the model's internal randomness during token selection rather than adding visible markers or hidden characters, producing a statistical signature detectable only with a secret key. Anthropic says watermarking has no practical effect on creativity, readability, token costs, or generation speed, and will be omitted where exact outputs or code correctness are required.

🔔 Amazon RDS for Oracle now supports Oracle Application Express (APEX) 26.1, a low-code platform for building secure, scalable enterprise applications. This managed database service simplifies setup, operation, and scaling of Oracle Database deployments in the cloud. APEX 26.1 is available in all AWS regions where RDS for Oracle is offered; consult the RDS documentation for details on enabling or modifying APEX options.

🔒 The Netherlands' NCSC warns that a macOS Screen Sharing authentication bypass (CVE-2026-65400) is being actively exploited after public exploit code appeared. The flaw affects the built-in VNC-based Screen Sharing service (TCP 5900) and allows network attackers to authenticate without valid credentials. Apple fixed the issue in macOS Tahoe 26.6.1, Sequoia 15.7.9 and Sonoma 14.8.9; affected users should update or disable Screen Sharing in System Settings.

🔍 New analysis links the ExfilSquad extortion group to leaked data from 13 victims across government, education, finance and manufacturing. Fortra Intelligence and Research Experts (FIRE) validated that public samples contained sensitive information, with published torrents totaling 382.64 GB and 27 million records. Researchers say misconfigured Microsoft Power Pages and unauthorized read access to Microsoft D365/Dataverse exports appear to be the primary cause, not a D365 vulnerability. FIRE identified numerous exposed Power Pages instances and highlighted the risk of the Anonymous Users web role.

🛡️ A maximum-severity remote code execution vulnerability in SAP Commerce Cloud (CVE-2026-58231) patched three days ago is already being targeted in attacks, Defused reports. The flaw, in the core Data Hub Adapter extension, allows unauthenticated actors to exploit improper authorization to execute arbitrary code. SAP warned the issue arises from abuse of a default authentication client and insufficient input validation. Threat researchers observed initial exploitation attempts hitting honeypots despite no public PoC existing.

🔒 Cloudflare announces new Cloudflare One capabilities to detect and control Model Context Protocol (MCP) traffic. These features let administrators identify which users and servers are generating MCP requests, distinguish Portal-mediated connections from direct ones, and block unauthorized direct connections on managed network paths. The update combines Gateway protocol signals with MCP Server Portals to help teams find shadow MCP servers and enforce Portal-only access to trusted MCP endpoints.

🔒 Oracle has introduced Database Security Central, a tool that provides a centralized view of security risk across database environments and will be free through February 2027. It arrives as attackers increasingly target Oracle database flaws and following Oracle’s move to monthly patch releases. The tool assesses posture, detects configuration drift, highlights privileged access risks, monitors sensitive data access, and centralizes policy management and audit evidence collection.

🔐 Cloudflare now lets you apply Access directly to a Worker or to all Workers in an account so applications are protected by your company login by default. When enabled, Access enforces authentication before any request reaches Worker code, regardless of domain, route, or preview URL. Policies can be set per hostname, per Worker, or account-wide, with the most specific policy taking priority. Developers can also access authenticated user details through ctx.access.getIdentity() for personalization and logging.

🔎 Shell is investigating a potential security incident after the Clop ransomware gang claimed to have stolen 89GB of data, including engineering drawings and project plans. A Shell spokesperson confirmed awareness and said security teams and external experts are examining the matter. Clop listed Shell among 43 victims allegedly targeted via a PTC Windchill and FlexPLM vulnerability tracked as CVE-2026-12569. PTC, CISA, and other authorities have warned of active exploitation and urged urgent patching and mitigations.

🔍 DecryptAds is a free service that scrapes and correlates public adtech files (ads.txt, app-ads.txt, buyers.json, sellers.json) to reveal which companies can run ads or harvest data from websites and apps. The site presents consolidated profiles, legal dossiers, and geo-risk warnings to help researchers and security teams trace malvertising, ad fraud, and opaque ad-supply chains. Its API and quiet-removals feed enable automation and visibility into removed sellers and reseller relationships.

📰 This essay, coauthored with Nathan E. Sanders and originally published in The Guardian, argues that OpenAI and Anthropic—once founded to restrain reckless corporate AI development—have been co-opted by market incentives and investor priorities. Recent market turbulence and questions about long-term profitability suggest these labs may not be viable as private, for-profit companies. The authors propose nationalizing their innovation and compute functions, converting them into publicly governed national labs and utilities to align AI with democratic values and public benefit.

🔒 In July 2026, the ShinyHunters extortion group claimed to have stolen personal data from RingCentral accounts after a reported social engineering intrusion. RingCentral acknowledged a security incident and said remediation steps were taken, noting services continued to operate and only a portion of customers were affected. Have I Been Pwned confirmed leaked data tied to 1.6 million accounts, including names, emails, phone numbers, and addresses.

🛡️ Researchers at Jamf warn of a Rust-based macOS infostealer named AmnesiaStealer distributed through ClickFix social engineering. The malware harvests credentials, browser data and live sessions, uses OS version–specific bypasses, and includes a remote-controlled second stage to stealthily control Chromium-family browsers. Jamf recommends enabling threat prevention, advanced threat controls and web protection set to Block and Report.

🔍 Security teams should oversee risk rather than perform every remediation task. Assign clear roles: security maintains the authoritative risk inventory, prioritizes findings, escalates missed commitments and verifies closure, while infrastructure, cloud, application and business owners execute fixes. Executives resolve resource conflicts and accept residual risk. Backlogs typically reflect organizational failures in ownership, capacity and decision-making rather than purely technical deficiencies.

🛡️ A former Brightly Software contractor was sentenced to two years in prison after pleading guilty to orchestrating a $2.5 million extortion scheme. He stole payroll and corporate data, emailed employees threatening to leak PII, and demanded ransom in cryptocurrency after his contract ended. Brightly paid a small Bitcoin ransom before involving law enforcement; the FBI recovered devices linking the suspect to the crimes.