Flowise AI CVE-2025-59528 Leaves 12,000 Instances Vulnerable to RCE
CVE-2025-59528 is a CVSS 10.0 RCE flaw in Flowise AI agent builder with 12,000 exposed instances. Here is who is affected, how to check, and what to patch.
Daily threat intelligence and vulnerability analysis from our security team. We publish expert breakdowns of critical CVEs, active exploits, and emerging attack campaigns as they happen.
Our analysts monitor vendor advisories, CISA alerts, and underground threat activity to give you actionable guidance you can use the same day. Filter by severity below to find what matters most to your environment.
Updated September 11, 2026 — all 201 published advisories are browsable here. The newest one is on the appliance that is supposed to be doing the protecting: a Cisco firewall management flaw under active attack and added to CISA's known-exploited list, on the console that controls the firewalls guarding the whole network (CVE-2026-20079). Behind it, a Chrome zero-day exploited in the wild, where simply visiting a booby-trapped site is enough to get code running on the machine (CVE-2026-87491), and a FreeIPA flaw that hands full administrator rights to anyone who connects without a password, on the directory server that decides who gets into everything else (CVE-2026-76578). The one readers keep coming back to is the Magento and Adobe Commerce zero-day being exploited against live online stores with no vendor patch available (StyleSmuggler). Every one of these sits on a device or service that faces the internet, so the clock starts the day the advisory drops — and a missed patch window is what turns a routine Tuesday into downtime, a breach-notification bill, and lost revenue. If you are not sure whether yours is patched, our 24/7 staffed security operations center handles the triage for you.
CVE-2025-59528 is a CVSS 10.0 RCE flaw in Flowise AI agent builder with 12,000 exposed instances. Here is who is affected, how to check, and what to patch.
North Korean hackers spent six months building trust with Drift Protocol contributors before stealing $285 million. The operation involved fake personas, conference appearances, and a million-dollar deposit to establish credibility before exploiting VS Code and TestFlight attack vectors.
Read moreNorth Korean hackers drained $285 million from Solana-based Drift Protocol using social engineering to compromise multi-signature approvals. The attack involved no code exploits, just patient manipulation of human trust over several weeks.
Read moreCheck Point researchers discovered a ChatGPT vulnerability allowing silent data exfiltration via DNS queries from the code execution sandbox. Separately, BeyondTrust found a critical command injection flaw in OpenAI Codex that enabled GitHub token theft through malicious branch names.
Read moreIran-linked Handala Hack breached FBI Director Kash Patel email and unleashed wiper malware on Stryker. First destructive attack on US Fortune 500.
Read moreThree critical vulnerabilities in LangChain and LangGraph let attackers steal files from filesystems, siphon API keys and environment secrets, and pillage conversation histories. With 84 million weekly downloads, most enterprise AI deployments are affected.
Read moreSecurity researchers at Cyera disclosed three vulnerabilities in LangChain and LangGraph affecting millions of AI applications. Path traversal, deserialization, and SQL injection flaws expose filesystem contents, environment secrets, and conversation histories.
Read moreTeamPCP has compromised LiteLLM, a Python package present in 36% of cloud environments. Malicious versions 1.82.7 and 1.82.8 deploy credential harvesters, Kubernetes lateral movement tools, and persistent backdoors.
Read moreOracle patches CVE-2026-21992, a pre-auth RCE in Identity Manager scoring 9.8 CVSS. No credentials needed to exploit. Emergency patches available now.
Read moreA second supply chain attack on Trivy compromised 75 GitHub Actions tags and spawned a credential-stealing worm across 47 npm packages. Check your CI pipeline.
Read moreCritical authentication bypass in Veeam Backup & Replication allows attackers to delete backup repositories without credentials.
Read moreSophisticated iOS exploit kit chains six vulnerabilities including three zero-days to achieve complete device takeover. Multiple threat actors including Russian espionage groups and commercial surveillance vendors observed using DarkSword against targets in Ukraine, Saudi Arabia, and Turkey.
Read moreNine critical vulnerabilities in budget IP KVM switches from GL-iNet, Angeet, Sipeed, and JetKVM allow unauthenticated code execution and hardware-level access.
Read moreCISA added CVE-2025-47813 (info disclosure) to KEV, used to enhance CVE-2025-47812 (CVSS 10.0 RCE) exploitation. Attackers chain both flaws for reliable remote access. Wing FTP patches available since May 2025. Federal deadline: March 30.
Read moreGoogle patched CVE-2026-3909 (Skia OOB write) and CVE-2026-3910 (V8 implementation flaw), both actively exploited. Third Chrome zero-day emergency in 2026. Update to 146.0.7680.75/76 immediately.
Read moreCVE-2026-42071 (CVSS 9.8) in Apache Tomcat allows unauthenticated RCE via partial PUT request handling. Actively exploited 30 hours after disclosure.
Read moreA critical RCE vulnerability in Atlassian Confluence is being mass-exploited by multiple threat actors.
Read moreIvanti discloses another actively exploited zero-day chain in Connect Secure VPN appliances. CVE-2026-0778 and CVE-2026-0779 allow unauthenticated attackers ...
Read moreA critical vulnerability in Microsoft Teams allows attackers to deliver malware through specially crafted meeting invitations.
Read moreNation-state attackers are actively exploiting a critical zero-day in Palo Alto GlobalProtect VPN to breach defense contractors. If you run GlobalProtect, apply the emergency patch now or isolate affected systems from the network immediately.
Read moreOur CyberOne MobileAssess platform performs deep static analysis, source code decompilation, and runtime security testing for iOS and Android apps. From one-time assessments to year-long continuous testing, we find what surface-level scanners miss.
Subscribe to our newsletter and get the latest security insights delivered to your inbox.