appsec.fyi

Weekly Changelog

148 new resources added across 10 topics this week.

May 29 — Jun 05, 2026

Supply Chain +60

DateResourceSummary
2026-06-05Hola browser supply chain attack delivers cryptocurrency minerHola browser was the target of a supply chain attack that delivered a cryptocurrency miner. Attackers compromised Hola's infrastructure, injecting malicious code into legitimate browser updates. This allowed them to distribute a cryptominer to unsuspecting users. The extent of the compromise and any potential payouts for discovering this vulnerability are not specified in the provided content.
2026-06-05600000 Monthly Downloads Affected: Miasma Supply Chain Attack Is Back on npmThe Miasma supply chain attack has resurfaced on npm, impacting approximately 600,000 packages with monthly downloads. This means a significant number of projects are at risk due to compromised dependencies. The specific details of the attack and the affected packages are outlined in the linked article. The severity of the situation underscores the ongoing threats within software supply chains and the importance of vigilant security practices for developers utilizing npm packages.
2026-06-05Kaspersky Discovers Supply Chain Attack Delivering Backdoors via Official Daemon Tools WebsiteKaspersky has uncovered a sophisticated supply chain attack that infiltrated the official Daemon Tools website. Attackers compromised the site to distribute malware-infected installers, delivering backdoors to unsuspecting users. This attack highlights the significant risks associated with supply chain vulnerabilities, where trusted software sources can be compromised to spread malicious payloads. The discovery underscores the importance of rigorous security measures throughout the software development and distribution lifecycle.
2026-06-05Rust-Written IronWorm Hits NPM Supply ChainA new malware called IronWorm, written in Rust, has been discovered targeting the NPM (Node Package Manager) supply chain. This malicious software infiltrates the development ecosystem by compromising popular packages. The goal of IronWorm is to steal sensitive information from developers and their projects, posing a significant risk to the security of software built using these compromised dependencies. Further details on its specific infection vectors and the full extent of its capabilities are still under investigation.
2026-06-05IronWorm Supply Chain Malware Hits npmLibrary for analyzing supply chain attacks like IronWorm, a self-replicating Rust-built malware that infected 36 npm packages, targeting environment variables, cloud credentials, and crypto wallets. This campaign, detected by JFrog, utilized binary executables within postinstall scripts and spread by stealing credentials to publish new malicious packages, impacting over 32,000 monthly downloads. Actions include rotating keys and enabling 2FA.
2026-06-04IronWorm Supply Chain Attack Uses Malicious npm Packages to Steal Developer SecretsA supply chain attack dubbed "IronWorm" is targeting developers through malicious npm packages. These compromised packages are designed to steal sensitive developer secrets. The attack highlights a growing threat vector where attackers inject malicious code into widely used software development tools and libraries, compromising the integrity of the software supply chain and potentially leading to widespread data breaches and unauthorized access. Further details on the specific methods and impact are available at the provided link.
2026-06-04Hola Browser for Windows compromised to deliver cryptominerLibrary compromised to deliver cryptominer. The Windows version of Hola Browser, built on Chromium, experienced a supply chain attack where an undeclared executable named ‘me.exe’ was installed, later identified as a Monero cryptocurrency miner. This malicious component added Windows Defender exclusion rules, copied itself as ‘HolaMonitorService.exe,’ created an auto-starting service, and ran during idle periods. Hola confirmed the supply chain compromise, stating approximately 0.1% of users were affected without evidence of data theft.
2026-06-04New IronWorm malware hits 36 packages in npm supply-chain attackLibrary infections by the IronWorm malware on npm's registry targeted 36 packages, stealing OpenAI, AWS, and npm credentials, SSH keys, and cryptocurrency wallet data. Written in Rust, IronWorm utilizes an eBPF rootkit and communicates via Tor, self-propagating by publishing trojanized package versions using stolen npm Trusted Publishing credentials. The malware leverages GitHub Actions to exfiltrate secrets as build artifacts and exhibits similarities to the Shai Hulud supply-chain attack.
2026-06-04Miasma Attack Hits Red Hat npm PackagesLibrary of obfuscated JavaScript code embedded within @redhat-cloud-services npm packages, identified as the Miasma attack. This malicious code, running via preinstall scripts, acts as a self-propagating worm and credential stealer, targeting developer and cloud credentials, and exploiting compromised GitHub accounts to achieve valid SLSA provenance for its releases. The attack compromised at least 32 package releases, averaging 80,000 weekly downloads, and is tracked by Snyk advisories, with the lead advisory rated Critical (CVSS v4.0) and exploit maturity as Attacked.
2026-06-04NCSC Releases Software Supply Chain Attacks GuidanceGuide from the NCSC details four attacker techniques used in software supply chain attacks: maintainer account compromise (seen in the Axios npm attack), abandoned package takeover, typosquatting, and self-propagation. It emphasizes the structural vulnerability of modern development's reliance on numerous third-party packages, particularly in ecosystems like Node.js and Python. Defenders are urged to implement visibility through software bills of materials, detection via anomaly monitoring and dependency scanning, and a remediation posture that includes pausing automatic updates, manual review, and MFA enforcement, especially on package registry accounts.
2026-06-04Supply chain attack compromises Red Hat software packages on npmAnalysis of a supply chain attack that compromised over 30 Red Hat Cloud Services packages on npm, allowing attackers to distribute malicious code through a trusted channel. The malware harvested cloud service credentials (AWS, Google, Microsoft), pipeline tokens, and developer tool passwords, transmitting them to attackers by mimicking Anthropic service addresses. It also established persistent background processes and embedded hooks in AI coding assistants, with a risk of file deletion if credentials are revoked prematurely.
2026-06-04Software supply chain attacks: check your dependenciesAnalysis of software supply chain attacks detailing how attackers compromise open-source packages, like those in Node.js, Rust, and Python ecosystems, to spread malware. It discusses techniques such as maintainer account compromise, abandoned package takeover, typosquatting, and self-propagation, and advises on identifying affected systems by reviewing dependencies, monitoring CI/CD activity, and scanning for known issues. Recommended mitigation steps include pausing automatic updates, manually approving new versions, rotating credentials, enforcing MFA, and strengthening development lifecycles.
2026-06-03ThreatLocker Highlights Key Cyber Threat Activity and Research from May 2026Analysis of May 2026 cyber threat activity by ThreatLocker highlights significant software supply chain attacks, including those targeting GitHub, Nx Console, and TanStack, alongside zero-day exploits like MiniPlasma, Linux Copy Fail, and Dirty Frag. The report underscores the abuse of trust in software ecosystems, code-signing infrastructure, and identity systems, recommending Zero Trust principles and application allowlisting/ringfencing to mitigate risks from both traditional threats and emerging AI-assisted attacks.
2026-06-03Red Hat hit by npm supplychain attack - here's how to stay safeLibrary for detecting and mitigating npm supply-chain attacks, specifically addressing the Red Hat @redhat-cloud-services namespace compromise. It highlights the use of npm preinstall hooks by malware like Miasma and Mini Shai-Hulud to steal credentials from environments including GitHub, AWS, and Kubernetes. Recommendations include rotating secrets, auditing activity, and rebuilding potentially contaminated systems.
2026-06-03Preinstall to persistence: Inside the Red Hat npm Miasma credential-stealing campaignLibrary for detecting the "Miasma" supply chain attack, which compromised 32 npm packages under the @redhat-cloud-services scope. The attack used obfuscated dropper scripts to download the Bun JavaScript runtime and steal credentials from GitHub, npm, AWS, Azure, GCP, HashiCorp Vault, and Kubernetes. It also attempted privilege escalation via passwordless sudo and self-propagation by publishing poisoned packages with forged SLSA provenance, and included a destructive tripwire to wipe the home directory.
2026-06-03Sonatype flags 176 npm packages in dependency attackSonatype has identified 176 npm packages compromised in a dependency attack. Attackers injected malicious code into these packages, posing a significant risk to software supply chains. This incident highlights the ongoing threat of malicious actors targeting open-source ecosystems. Developers relying on these packages are urged to audit their dependencies and consider alternatives to mitigate potential vulnerabilities. No bounty payout amount was explicitly stated in the provided content.
2026-06-03Sonatype flags 176 npm packages in dependency attackSonatype has identified a significant supply chain attack involving 176 npm packages. These packages, ranging from utilities to frameworks, were found to contain malicious code, posing a risk to developers using them. The discovered vulnerabilities allow for the potential execution of arbitrary code, leading to unauthorized access and data breaches. This incident highlights the ongoing threats within the open-source software ecosystem and underscores the importance of robust security practices for managing dependencies.
2026-06-03Dozens of Red Hat npm packages targeted in supply- chain attackDozens of Red Hat npm packages were compromised in a sophisticated supply-chain attack. The vulnerability allowed attackers to inject malicious code into the development pipeline, potentially affecting a wide range of users and projects relying on these packages. Details about the specific vulnerabilities and the extent of the compromise are still emerging, but the incident highlights the ongoing risks associated with software supply chains. No bounty payout amount is mentioned in the provided content.
2026-06-03OpenAI Codex Supply Chain Attack Exposes Growing Risks in AI Development EnvironmentsA supply chain attack targeted OpenAI's Codex, exposing vulnerabilities in AI development. The incident, which involved a compromised private GitHub repository, highlights the increasing risks associated with the complex dependencies and third-party code used in building AI systems. Attackers could potentially leverage such breaches to inject malicious code, leading to data theft or system manipulation. This event underscores the critical need for enhanced security measures throughout the AI development lifecycle, from code repositories to model training environments, to protect against sophisticated threats.
2026-06-03Red Hat Confirms Supply Chain Breach Impacting @redhat-cloud-services npm PackagesRed Hat Confirms Supply Chain Breach Impacting @redhat-cloud-services npm Packages https://ift.tt/YeO8mcy
2026-06-02Megalodon Supply Chain Attack Infects Over 5500 GitHub Repositories with Backdoors and StealersA sophisticated supply chain attack, dubbed Megalodon, has compromised over 5,500 GitHub repositories. The attackers injected malicious code, specifically backdoors and stealer malware, into these repositories. This widespread infection poses a significant risk to the software development ecosystem, as compromised code can be propagated to numerous downstream projects and users. The attack highlights the persistent threat of supply chain vulnerabilities and the need for robust security measures in software development pipelines.
2026-06-02Claude Code GitHub Actions Flaw Created Supply Chain Attack RiskLibrary components for Claude Code GitHub Actions were found to have vulnerabilities that could lead to supply chain attacks. These flaws allowed attackers to bypass permission controls and inject untrusted input into trusted workflows, potentially exposing sensitive credentials like GitHub Actions OIDC tokens. The most severe scenario could have introduced malicious code into Anthropic's own claude-code-action repository, impacting downstream users. Organizations should upgrade to the latest version and review workflow permissions to mitigate these risks.
2026-06-02Supply Chain Cyber Threats: How Malicious AI Puts Logistics Data at RiskMalicious AI poses a significant threat to logistics data within supply chains. Attackers can leverage AI to find vulnerabilities and launch sophisticated attacks, potentially disrupting operations and compromising sensitive information. This poses a risk to the integrity and security of supply chain data.
2026-06-02Why supply chain attacks work and what detection can actually do about itAnalysis of supply chain attacks highlights the exploitation of trust, as demonstrated by TeamPCP's campaigns across GitHub, PyPI, npm, and Docker Hub. Their methods, including poisoned VS Code extensions like Nx Console and automated worms like Mini Shai-Hulud, bypass traditional security measures by targeting developers and leveraging CI/CD pipelines. While initial compromise detection is challenging, robust endpoint and SIEM monitoring remain crucial for identifying anomalous post-compromise activity driven by stolen credentials.
2026-06-02State of Supply Chain Security: Roundup of the Big HitsRoundup of significant cyber incidents highlights the escalating threats to software supply chains and third-party vendors. The Maersk NotPetya attack in 2017, originating from a compromised M.E.Doc update, crippled global operations and underscored the need for patch management and business continuity. More recent attacks in 2025 impacted UK grocery chains like Marks & Spencer and manufacturer JLR, causing substantial revenue losses and production halts. These events emphasize the critical need for asset visibility, dependency tracking, continuous compliance monitoring, software security, and robust vendor risk management in increasingly interconnected ecosystems, especially with the expanding attack surface driven by AI adoption.
2026-06-02Attack targeting OpenAI Codex users exposes AI software supply chain risksWriteup of the codexui-android npm package attack, revealing AI software supply chain risks. Attackers hid malicious code within a seemingly legitimate OpenAI Codex remote user interface package, exfiltrating developer authentication tokens, including long-lived refresh tokens. This incident highlights vulnerabilities in build and distribution pipelines, where published software artifacts may differ from public source code, leading to persistent access to AI developer tools and the resources they control.
2026-06-02ReversingLabs Spectra Assure Wins 2026 Fortress Cybersecurity Award in Software Supply Chain SecurityLibrary for securing the software supply chain. ReversingLabs Spectra Assure, recognized with a 2026 Fortress Cybersecurity Award, provides deep binary analysis for visibility and control against threats, addressing a 73% increase in malicious open-source packages. It offers significant efficiency improvements, reducing third-party risk management from months to a week and employee software approval times from hours to minutes.
2026-06-02New npm Supply Chain Attack: @redhat-cloud-services CompromisedAnalysis of a multi-stage dropper infecting the `@redhat-cloud-services` organization in npm, identified as a variant of the Shai-Hulud malware. This infostealer targets GitHub tokens, npm tokens, AWS, GCP, and Azure cloud credentials, and local environment information. It employs enhanced obfuscation, multi-stage loading, and uses `api.anthropic.com` as a decoy C2 server, exfiltrating data to newly created GitHub repositories with the description "Miasma: The Spreading Blight." The malware exhibits destructive behavior, potentially nuking the compromised machine if stolen tokens are revoked.
2026-06-02Compromised Red Hat npm packages downloaded over 80000 times in one week supply chain attack still ongoingWriteup of a Red Hat npm supply chain attack, where compromised packages were downloaded over 80,000 times. The attack leveraged a variant of the Mini Shai-Hulud worm to steal GitHub secrets, npm tokens, cloud credentials, and other sensitive information, with a copycat worm exhibiting similar tradecraft and targeting GCP and Azure identities.
2026-06-02Red Hat removes tainted packages after software pipeline compromiseWriteup of Red Hat's response to a supply chain attack involving the Mini Shai-Hulud worm variant, Miasma, which was distributed via a compromised GitHub account. The attack affected 32 packages and targeted developers with credential-stealing malware. This incident follows a series of similar supply chain compromises, including attacks on LiteLLM, the axios JavaScript library, and breaches affecting GitHub and OpenAI employees via malicious extensions.
2026-06-02Critical Supply Chain Attack Compromises 32 Red Hat @redhat-cloud-services NPM Packages with Credential-Stealing MalwareLibrary exploiting the @redhat-cloud-services NPM package supply chain attack, which injected Mini Shai-Hulud malware to steal credentials and propagate via GitHub Actions OIDC and NPM's bypass_2fa. This attack compromised 32 Red Hat packages, impacting over 116,000 weekly downloads and necessitating immediate rotation of all exposed secrets and affected package replacements.
2026-06-0234 Malicious Packages Steal Cloud Keys Wallets and SSH CredentialsThirty-four malicious npm packages have been discovered that steal sensitive information from developers. These packages, disguised as legitimate tools, are designed to exfiltrate cloud API keys, cryptocurrency wallet credentials, and SSH keys. The compromised packages were published on the npm registry, a popular repository for JavaScript. This incident highlights the ongoing threat of supply chain attacks and the importance of vigilant security practices when using third-party code. No specific bounty payout amount was mentioned in the provided content.
2026-06-02Supply Chain Attack Hits 32 Red Hat NPM PackagesAnalysis of a supply chain attack on 32 Red Hat NPM packages details a credential-stealing worm that exploited compromised CI/CD pipelines via GitHub Actions OIDC. The malicious preinstall hook executed during NPM install, harvesting sensitive data like GitHub secrets, npm tokens, and cloud credentials, similar to the Mini Shai-Hulud worm. Attackers exfiltrated data to attacker-controlled servers or GitHub repositories, with initial findings indicating at least 210 repositories may contain stolen credentials. Red Hat has since released clean versions of the affected packages.
2026-06-02Attack targeting OpenAI Codex users exposes AI software supply chain risksAnalysis of codexui-android, a malicious npm package targeting OpenAI Codex users, reveals AI software supply chain risks. This package, downloaded thousands of times weekly, appeared legitimate but exfiltrated developer authentication tokens, including long-lived refresh tokens, by injecting malicious code into the distributed artifact, not the public GitHub source. This highlights a blind spot where build and distribution pipelines, rather than source code, become the attack vector, necessitating verification of package provenance and source code consistency.
2026-06-02Miasma Revives Shai-Hulud Supply Chain Attack TacticsLibrary for detecting and mitigating Miasma, a supply chain attack campaign evolving from Shai-Hulud. Miasma targets npm packages, specifically compromised redhat-cloud-services, to steal credentials, harvest secrets, and propagate through development environments. It utilizes install-time execution, encrypted exfiltration via GitHub API commits, and exploits GitHub Actions for persistence and privilege escalation, even evading endpoint protection like CrowdStrike.
2026-06-02Unpatched software is now the top way into banksSurvey of the 2026 Verizon Data Breach Investigations Report, highlighting unpatched software as the leading entry vector into banks, surpassing stolen credentials for the first time. The report also identifies increasing third-party vendor involvement in breaches and the growing threat of AI-assisted attacks exploiting known software flaws.
2026-06-02Multiple redhat-cloud-services npm Packages compromisedLibrary analyzing multiple @redhat-cloud-services npm packages compromised with malware, executing on install before application code. The payload uses ROT-21, AES-128-GCM, obfuscator.io, and a B5 cipher to evade detection and harvest secrets from GitHub Actions, AWS, GCP, Azure, Kubernetes, Vault, npm, and CircleCI. It also acts as a self-propagating worm, using stolen npm tokens and `bypass_2fa` to republish backdoored packages. Analysis involved static and dynamic techniques, including StepSecurity Harden-Runner.
2026-06-02Containers on fire: from container escapes to supply chain attacksThis article, "Containers on fire: from container escapes to supply chain attacks," explores the security risks associated with containerized environments. It discusses vulnerabilities such as container escapes, which allow attackers to break out of isolated containers, and supply chain attacks, which compromise the integrity of software components used within containers. The content likely delves into the methods attackers exploit and the potential impact of these security breaches on organizations utilizing container technology.
2026-06-01OpenAI Codex tool with over 29000 downloads linked to malicious npm supply chain attack stealing authentication tokensLibrary exposing a supply-chain attack targeting OpenAI Codex users. A malicious npm package, "codexui-android," with over 29,000 downloads, exfiltrated non-expiring refresh tokens. Researchers also identified two Android apps, OpenClaw Codex Claude AI Agent and Codex, with tens of thousands of downloads, which similarly targeted Codex users by sending credentials to attacker-controlled servers.
2026-06-01Dozens of Red Hat packages backdoored through its offical NPM channelWriteup detailing a supply-chain attack targeting Red Hat's official NPM channel. Threat actors compromised the `@redhat-cloud-services` namespace, publishing over 30 backdoored packages. These packages execute obfuscated payloads during `npm install`, stealing credentials like GitHub action secrets, npm tokens, Kubernetes, and Vault material. The malware then spreads by republishing compromised packages to other accounts, with infected systems encrypting and exfiltrating data via web requests or to compromised GitHub repositories.
2026-06-01Miasma: Supply Chain Attack Targeting RedHat npm PackagesAnalysis of Miasma details a supply chain attack targeting @redhat-cloud-services npm packages, compromising at least 32 releases. The attack involved unauthorized code modifications, obfuscated JavaScript payloads using eval() and ROT-based decoding, and new data collectors for GCP and Azure identities. This variant, similar to TeamPCP's (Mini) Shai-Hulud malware, employs unique encrypted payloads per infection. The root cause appears to be a compromised Red Hat employee GitHub account that injected malware via orphan commits and manipulated GitHub Actions to publish packages with valid SLSA provenance attestations.
2026-06-01Miasma Supply Chain Attack Compromises Red Hat npm Packages with Credential-Stealing WormAnalysis of the Miasma supply chain attack details how compromised @redhat-cloud-services npm packages, including `@redhat-cloud-services/vulnerabilities-client` and others, were injected with an obfuscated preinstall hook. This hook targeted developer machines and CI/CD environments to steal credentials like GitHub Actions secrets, npm tokens, and cloud identities, employing tactics similar to the Mini Shai-Hulud worm. The malware exfiltrated data encrypted to `api.anthropic[.]com:443/v1/api` and leveraged GitHub commits for further propagation, with evidence suggesting a Red Hat employee's compromised GitHub account as the initial entry point.
2026-06-01CrowdStrike and Google take down botnet used by hackers to target software developers in supply chain attacksCrowdStrike and Google have successfully dismantled a botnet that was being used by hackers to conduct supply chain attacks specifically targeting software developers. These malicious actors exploited vulnerabilities to compromise the development environment, potentially impacting a wide range of software products. The coordinated takedown aims to disrupt this threat and protect the software development ecosystem from further exploitation. No bug bounty payout amount was mentioned in the provided content.
2026-06-01OpenAI Codex Authentication Tokens Stolen in codexui-android npm Supply Chain AttackLibrary for securing OpenAI Codex authentication tokens, affected by a supply chain attack via the `codexui-android` npm package and Android apps like OpenClaw Codex Claude AI Agent and Codex (BrutalStrike). The malicious code exfiltrates tokens from `~/.codex/auth.json` to `sentry.anyclaw.store`, granting attackers persistent access due to non-expiring refresh tokens. This highlights a growing trend of targeting AI developer tooling for credential theft.
2026-06-01Microsoft discovers new npm attack in 14 packagesLibrary containing malware that targeted developers via 14 malicious npm packages designed to steal credentials for AWS, HashiCorp Vault, and GitHub Actions. The attack utilized typosquatting and copied metadata from legitimate projects, employing npm installation hooks and the Bun runtime for execution to compromise cloud and CI/CD platforms, enabling lateral movement and further data exfiltration.
2026-06-01Mercor Hit: 4TB Stolen via LiteLLM (95M Downloads) [2026]Library providing a unified interface to over 100 LLM providers, LiteLLM was compromised, leading to a 4TB data exfiltration from Mercor. The attack chain involved compromising Trivy, injecting malicious code into LiteLLM, and exfiltrating credentials from downstream environments. This incident highlights the risks of fast-moving open-source projects in AI infrastructure and the critical importance of pinned dependencies to prevent supply-chain attacks.
2026-06-01Emerging Threats to AI-Assisted Software Supply Chains Highlight Security DemandAI-assisted software development introduces new security vulnerabilities to the software supply chain. These threats can be exploited to compromise AI models, inject malicious code, or manipulate training data, leading to insecure software. The rise of these risks underscores the critical need for robust security measures throughout the AI development lifecycle to ensure the integrity and safety of AI-assisted software.
2026-05-31SlowMist Says TrapDoor is One of 2026s Largest Supply Chain AttacksAnalysis of the TrapDoor supply chain attack, identified by Socket and detailed by SlowMist, reveals a sophisticated campaign targeting crypto and AI developers via malicious packages on npm, PyPI, and Crates.io. The attack utilized hidden code within installation and build processes, leveraging trusted developer services like GitHub Pages to exfiltrate sensitive data including SSH keys, cloud credentials, and crypto wallets. The npm variant, particularly advanced, manipulated Git hooks and AI coding assistant files like `.cursorrules` and `CLAUDE.md`, employing prompt injection to spread malicious instructions.
2026-05-3114 malicious npm packages impersonated OpenSearch Elasticsearch librariesWriteup on 14 malicious npm packages impersonating OpenSearch and Elasticsearch libraries, demonstrating a supply chain attack vector targeting developers. These packages, designed to mimic legitimate OpenSearch and Elasticsearch modules, pose a significant risk to software integrity and development pipelines.
2026-05-31CISA warns that Nx Console and GitHub repositories abused in multiple supply chain compromises tools across enterprise cloud and DevOps environments exploitedAnalysis of supply chain attacks by CISA details exploitation of Nx Console VSCode extension and GitHub repositories via the Megalodon campaign. Threat actors pilfered CI/CD secrets and cloud credentials by poisoning workflows. Recommended mitigations include auditing contributor activity, forensic reviews, rotating secrets, pinning trusted package versions, and delaying package pulls for community detection.
2026-05-31Solana Sui and Aptos wallet data targeted in TrapDoor package attackLibrary containing malicious packages on npm, PyPI, and Crates.io designed for the TrapDoor supply-chain attack. These packages disguised as developer utilities target Solana, Sui, and Aptos wallet data, along with SSH keys, GitHub tokens, and cloud credentials. Attackers also abused AI configuration files like `.cursorrules` and `CLAUDE.md` to exfiltrate secrets during AI coding sessions.
2026-05-31Hackers caught hiding OpenAI token-stealing malware in Codex npm packageHackers have embedded malware designed to steal OpenAI API tokens within the popular Codex npm package. This malicious code was discovered by security researchers, who identified it as a sophisticated attempt to gain unauthorized access to users' AI models and data. The discovery highlights a growing trend of supply chain attacks targeting software development tools. Further investigation is ongoing to determine the full scope of the breach and the potential impact on users.
2026-05-31Perplexity launches Bumblebee: How its new read-only dev scanner differs from ChainguardTool, Bumblebee, is a read-only scanner by Perplexity designed to check developer machines for risky packages, extensions, and AI tool configurations during supply-chain incidents. It targets four surfaces including language package managers like npm, pnpm, Yarn, Bun, PyPI, Go modules, RubyGems, and Composer. Unlike other tools, Bumblebee avoids executing scripts and directly inspects metadata files, preventing potential attacks triggered by scanner execution. It differentiates itself from Chainguard by focusing on the developer workstation environment rather than containers and build pipelines.
2026-05-30The hidden AI security flaw behind four major supply chain attacksLibrary for hardening software supply chains against AI-related supply chain attacks, this resource details vulnerabilities in build systems and release workflows that bypass traditional AI model security testing. It highlights incidents impacting OpenAI, Anthropic, Meta, and TanStack, specifically mentioning the TanStack Mini Shai-Hulud worm's exploitation of GitHub Actions and cache poisoning, and the LiteLLM attack's reliance on credential reuse in PyPI. The library emphasizes the need for pipeline-focused red teams, behavioral validation of build processes, dependency credential hygiene, mandatory human review before releases, and strict input sanitization in build tools to mitigate risks like command injection.
2026-05-30Download pumping: New npm deception technique for supply chain attacksA new npm deception technique called "download pumping" has been discovered, posing a threat to supply chain security. This method involves malicious packages registering fake download metrics to appear more popular and trustworthy than they are. Researchers observed campaigns where attackers used this technique to obscure malicious code within seemingly legitimate packages, making them harder to detect and leading to potential system compromises. This discovery highlights the evolving tactics used in supply chain attacks and the need for enhanced vigilance in package vetting.
2026-05-30Malicious npm packages abuse dependency confusion to profile developer environmentsLibrary for detecting malicious npm packages that exploit dependency confusion to profile developer environments. These packages impersonate internal corporate namespaces and use obfuscated reconnaissance payloads downloaded from attacker-controlled C2 servers. They leverage npm lifecycle hooks for automatic execution during `npm install`, employing anti-analysis techniques and targeting various operating systems. The attack chain involves namespace squatting, spoofed enterprise metadata, and inflated version numbers, with a reconnaissance-only mode that collects system information and credentials for potential follow-on exploitation.
2026-05-29Typosquatted npm packages used to steal cloud and CI/CD secretsLibrary detailing an npm supply chain attack where typosquatted packages like "opensearch-setup" and "elastic-opensearch-helper" were used to steal AWS credentials, HashiCorp Vault tokens, and CI/CD secrets. The malicious packages leverage npm lifecycle hooks to execute a credential harvester that targets AWS IMDSv2, ECS task metadata, Secrets Manager across multiple regions, and npm publish tokens, enabling cloud lateral movement and downstream supply-chain pivoting.
2026-05-29Malicious Sicoob NuGet Steals Banking Credentials as npm Packages Target Cloud SecretsLibrary exploiting Sicoob NuGet and npm packages targets sensitive data. Malicious versions of "Sicoob.Sdk" exfiltrate client IDs and PFX certificates, while npm packages like "@vpmdhaj/devops-tools" harvest AWS credentials, Vault tokens, and CI/CD secrets. These attacks employ techniques such as typosquatting, dependency confusion, and brandjacking to achieve manufactured legitimacy and compromise developer workflows, echoing broader supply chain attack campaigns.
2026-05-29CISA adds Daemon Tools TanStack and Nx Console compromised versions to KEV catalogCatalog listing of CVE-2026-8398, CVE-2026-45321, and CVE-2026-48027, impacting Daemon Tools Lite, TanStack npm packages, and Nx Console. These vulnerabilities, identified by CISA for inclusion in the Known Exploited Vulnerabilities catalog, resulted from supply chain attacks. Compromised Daemon Tools installers utilized valid code-signing certificates, while TanStack packages were maliciously published via GitHub Actions. The Nx Console vulnerability involved a malicious version appearing on the Visual Studio Marketplace and OpenVSX.
2026-05-29CISA urges security teams to check for software development compromisesCISA is issuing an urgent alert to security teams, advising them to proactively scan their systems for compromises within their software development environments. This directive highlights the critical need to safeguard the integrity of the software supply chain. The agency's recommendation stems from concerns about potential vulnerabilities and breaches that could affect the development process, leading to widespread risks for downstream users. Security teams are encouraged to implement robust checks and balances to ensure the safety and trustworthiness of their software development practices.

RCE +45

DateResourceSummary
2026-06-05Android Update Patches Exploited Zero-Day 123 Other VulnerabilitiesGoogle has released an Android security update addressing a critical zero-day vulnerability that was actively exploited in the wild. The update also patches 123 other vulnerabilities, ranging in severity. This rapid response highlights the importance of keeping Android devices updated to protect against potential threats. No specific bounty payout amount was mentioned for this vulnerability.
2026-06-05VS Code Vulnerability Allows One-Click GitHub Token TheftA critical vulnerability in Visual Studio Code allows for one-click GitHub token theft. Attackers can exploit this flaw to steal sensitive authentication tokens, potentially gaining unauthorized access to users' GitHub repositories. The vulnerability reportedly allows for an easy exploit, requiring minimal user interaction. Further details are available at the provided link.
2026-06-05Hackers Exploit Critical Everest Forms Pro WordPress Plugin Flaw to Take Over SitesHackers are exploiting a critical vulnerability in the Everest Forms Pro WordPress plugin, allowing them to gain full administrative control over affected websites. This flaw enables attackers to steal sensitive data and perform unauthorized actions. The vulnerability has been fixed in the latest version of the plugin, and users are strongly advised to update immediately to prevent their sites from being compromised.
2026-06-05Microsoft Edge Vulnerability Allows Remote Attackers to Execute Arbitrary CodeA critical vulnerability in Microsoft Edge enables remote attackers to execute arbitrary code on affected systems. This security flaw, detailed in a recent report, poses a significant risk as it allows malicious actors to gain control of a user's device without requiring any interaction. The exact payout for reporting this bug was not disclosed. Users are advised to ensure their Microsoft Edge browsers are updated to the latest version to mitigate this threat.
2026-06-04Critical Redis vulnerability CVE-2026-23479 allows remote code executionWriteup of CVE-2026-23479, a critical use-after-free vulnerability in Redis allowing remote code execution. Discovered by Team Xint Code, the exploit chains Lua scripting and memory manipulation to overwrite function pointers in the Global Offset Table, enabling execution of arbitrary commands via `system()`. Affecting Redis versions 7.2.0 and later, this flaw requires authenticated access with specific ACL privileges, commonly found in default cloud deployments. Patches are available in Redis versions 7.2.14, 7.4.9, 8.2.6, 8.4.3, and 8.6.3.
2026-06-04Critical vulnerability in Hugging Face Transformers library allowed arbitrary code executionWriteup on CVE-2026-4372 in Hugging Face Transformers, a critical remote code execution vulnerability allowing attacker-controlled AI models to execute arbitrary code. Exploitable via a malicious payload in model configuration files when loading with `from_pretrained()`, even with `trust_remote_code=False` and the `kernels` package installed. Versions 4.56.0 through 5.2.x were affected, with millions of downloads. Hugging Face patched this in version 5.3.0.
2026-06-049.8 Mirasvit bug actively exploited on Magento serversWriteup of CVE-2026-45247, a critical 9.8 Mirasvit bug in the Full Page Cache Warmer extension for Magento and Adobe Commerce, is actively exploited for remote code execution. Exploitation involves bypassing storefront authentication via malicious base64 encoded payloads in the CacheWarmer HTTP cookie. CISA added this vulnerability to its Known Exploited Vulnerabilities catalog, urging federal agencies to patch by June 6. The flaw poses a significant supply chain risk due to Magento's prevalence in e-commerce, potentially exposing payment credentials, API keys, customer data, and enabling deeper business system access. Organizations must prioritize patching to version 1.11.12 and actively hunt for post-exploitation artifacts.
2026-06-04Everest Forms Pro Vulnerability Allows Remote Code Execution on WordPress SitesVulnerability in Everest Forms Pro, CVE-2026-3300, permits unauthenticated attackers to execute remote code on WordPress sites. The flaw, residing in the Calculation add-on, allows injection of PHP code via the `eval()` function due to insufficient sanitization of single quotes. Exploits observed include registering administrator accounts like "diksimarina" and planting webshells, with over 29,300 blocked attempts detected by Wordfence. The issue is patched in version 1.9.13.
2026-06-04Critical Hugging Face Transformers flaw ran attacker code on a routine model loadLibrary vulnerability affecting Hugging Face's Transformers, specifically CVE-2026-4372, allowed attackers to execute arbitrary code by slipping malicious payloads into model configuration files. This critical flaw bypassed the `trust_remote_code=False` setting, leading to silent system compromise and potential theft of sensitive data like cloud credentials and API keys. Exploitable versions range from 4.56.0 through 5.2.x when the `kernels` package is installed, with a fix released in version 5.3.0.
2026-06-04Hackers Actively Exploiting WordPress Plugin Vulnerability to Inject Malicious PHP CodeHackers are actively exploiting a vulnerability in a WordPress plugin to inject malicious PHP code. This allows them to compromise websites, steal sensitive data, and disrupt operations. The vulnerability has been detected in multiple sites, and its widespread exploitation poses a significant threat to WordPress users. It is crucial for users to update their plugins to the latest versions to patch this security flaw and protect their websites from further attacks.
2026-06-04Hugging Face Transformers RCE flaw enables stealthy compromise via AI model configsLibrary vulnerability in Hugging Face Transformers (CVE-2026-4372) allows attackers to achieve remote code execution by including a specially crafted `_attn_implementation_internal` parameter in model configuration files. This bypasses the `trust_remote_code=false` protection, enabling the execution of arbitrary Python code from attacker-controlled repositories without user prompts or runtime warnings, particularly impacting users with GPU-accelerated inference due to the optional `kernels` dependency.
2026-06-04Redis Use-After-Free Remote Code Execution Vulnerability (CVE-2026-23479)A critical use-after-free vulnerability in Redis (CVE-2026-23479) allows for remote code execution. This means an attacker could potentially gain control of a server running a vulnerable Redis instance. The vulnerability arises from a flaw in how Redis handles memory after an object has been freed. This could lead to serious security breaches, allowing unauthorized access and modification of data. Users are advised to update their Redis installations to a patched version as soon as possible to mitigate this risk.
2026-06-04CISA Adds Exploited Magento RCE Flaw CVE-2026-45247 to KEV CatalogCISA has added an exploited Magento remote code execution (RCE) vulnerability, CVE-2026-45247, to its Known Exploited Vulnerabilities (KEV) catalog. This addition signifies that the vulnerability is actively being exploited in the wild and poses a significant threat. Organizations using Magento are urged to prioritize patching or mitigating this vulnerability to prevent potential cyberattacks. The KEV catalog is a crucial resource for cybersecurity professionals to identify and address the most critical threats.
2026-06-03CERT-In flags security vulnerabilities in Microsoft Office appWriteup on CERT-In advisories detailing high-severity vulnerabilities in Microsoft Office and Microsoft 365 Copilot, including CVE-2026-45659, which allows remote code execution via untrusted data deserialisation. The flaws, potentially exploitable through malware-laced documents, risk data theft, and require immediate updates to Office via the Word app's Account settings to apply Microsoft's patches.
2026-06-03Autonomous AI Tool Finds 2-Year-Old RCE Flaw in Redis (CVE-2026-23479)Tool: An autonomous AI security tool by Team Xint Code identified CVE-2026-23479, a two-year-old use-after-free vulnerability in Redis (CWE-416). This flaw, introduced in Redis 7.2.0 and present in multiple branches until May 2026 fixes, allows for remote code execution (RCE) by overwriting the Global Offset Table (GOT) to redirect function calls. The exploit chain, detailed by Wiz, involves leaking a heap address via Lua scripting, manipulating client memory to achieve a use-after-free, and then overwriting GOT entries, notably `strcasecmp()` to `system()`, especially in default Redis Docker deployments with partial RELRO.
2026-06-03Exclusive: How One Line of Code Put Billions of Microsoft Android App Downloads at RiskLibrary detailing a critical vulnerability in six Microsoft Android apps, including Word and Excel, where a single debug flag, `setIsDebugMode(true)`, left in production code allowed untrusted apps to intercept Microsoft account access tokens. This flaw, affecting billions of downloads and assigned CVE-2026-41100, -41101, and -41102, enabled potential supply chain attacks by granting attackers access to sensitive data like emails and documents. The issue was identified by Enclave and has since been patched by Microsoft.
2026-06-03CISA flags two-year-old Oracle flaw as actively exploited in attacksAdvisory regarding CVE-2024-21182, a critical Oracle WebLogic Server vulnerability, now actively exploited. CISA mandated federal agencies patch this flaw, exploitable remotely by unauthenticated attackers via T3 or IIOP to gain unauthorized access. Over 1,500 vulnerable Oracle WebLogic servers have been identified online. This advisory highlights the urgency of patching known vulnerabilities, especially those flagged by CISA as being actively exploited.
2026-06-03Attackers exploit Palo Alto GlobalProtect flaw days after disclosureLibrary for analyzing the CVE-2026-0257 vulnerability affecting Palo Alto GlobalProtect, which allows credential-less authentication bypass into enterprise networks. Attackers can forge authentication override cookies using public keys to establish unauthorized VPN sessions, a method that bypasses traditional security measures and poses a significant risk in zero-trust environments. The flaw exploits how PAN-OS handles cookies without proper signature verification, enabling stealthy network access without malware or stolen credentials.
2026-06-03CVE-2026-34197 Jolokia Exposure Enables RCE in Apache ActiveMQWriteup of CVE-2026-34197, a remote code execution vulnerability in Apache ActiveMQ Classic, stemming from insecure exposure of broker management via the Jolokia HTTP/JMX interface. This allows attackers to interact with privileged operations, potentially leading to broker-side processing of malicious configuration content. Validation confirmed exploit path feasibility in controlled, insecure environments, highlighting a significant enterprise security concern due to ActiveMQ's critical infrastructure role.
2026-06-03Critical Vulnerability in HP VoIP Phones Enables Enterprise Network BreachesWriteup of CVE-2026-0826, a critical stack-based buffer overflow in HP Poly Voice VoIP phones impacting models like the VVX 150-450 and Trio 8800-8300 series. The vulnerability, exploitable via crafted SIP INVITE requests with malicious Session Description Protocol attributes when Interactive Connectivity Establishment (ICE) is enabled, allows for remote code execution with root privileges by leveraging Return Oriented Programming chains to bypass ASLR and NX mitigations. Disabling ICE or updating firmware mitigates the risk.
2026-06-02HP Poly VoIP vulnerability sets the stage for executive voice deepfakesWriteup of CVE-2026-0826, a critical buffer overflow vulnerability in HP Poly VoIP phones, allows unauthenticated attackers to gain root access and perform eavesdropping or record audio for AI-enabled voice deepfakes. Discovered by Rapid7, the flaw in the SDP parsing code, even with ASLR enabled, enables exploit execution via Metasploit. This vulnerability highlights the growing threat of embedded device compromise for both traditional espionage and modern AI-driven fraud.
2026-06-02Attackers are exploiting Palo Alto Networks defect that initially flew under the radarWriteup of CVE-2026-0257, an actively exploited authentication-bypass vulnerability in Palo Alto Networks firewalls, allowing remote attackers to establish VPN connections. Initially rated medium, it was escalated to critical by Palo Alto Networks, CISA, and Rapid7 due to observed exploitation in the wild. The exploit, requiring a specific configuration of GlobalProtect with authentication override cookies and certificate reuse, involves forging a valid authentication cookie using the appliance's public TLS certificate. Attackers are opportunistically targeting this vulnerability for initial access, adapting quickly to published research.
2026-06-0219-Year-Old Linux Kernel Vulnerability Exposes Systems to Root AccessLibrary of proof-of-concept code for CVE-2024-XXXX, a 19-year-old Linux kernel vulnerability named CIFSwitch, allowing low-privileged users to gain root access. The vulnerability exploits the CIFS subsystem's handling of SMB authentication, enabling attackers to manipulate key descriptions and inject malicious Name Service Switch modules via the cifs.upcall helper when certain Linux distributions like Linux Mint, CentOS, and Rocky Linux have cifs-utils installed.
2026-06-02Threat Actors Reportedly Target CVE-2026-41089 FlawWriteup of CVE-2026-41089, a critical Windows Netlogon vulnerability, detailing its exploitation by threat actors for remote code execution. This stack-based buffer overflow, accessible by unauthenticated attackers via crafted network requests, allows privileged code execution on domain controllers. Authorities like the Centre for Cybersecurity Belgium have issued warnings, urging immediate patching due to its high CVSS score of 9.8 and its potential to compromise entire networks.
2026-06-02TP-Link Router Security Bug Enables Remote Command Execution AttacksA critical security vulnerability in TP-Link routers allows for remote command execution. This flaw enables attackers to compromise the devices without requiring user interaction or authentication, potentially leading to widespread network breaches. Further details on the exploit and affected models are available via the provided link.
2026-06-02Critical Windows Netlogon Vulnerability in Attackers CrosshairsWriteup of CVE-2026-41089, a critical Windows Netlogon vulnerability, details its exploitation for remote code execution by threat actors. This stack-based buffer overflow flaw allows unauthenticated attackers to target domain controllers, potentially gaining system privileges. The Centre for Cybersecurity Belgium warns of active in-the-wild exploitation, urging immediate patching despite Microsoft's initial assessment. This vulnerability poses a significant risk due to Netlogon's role in domain authentication.
2026-06-02RedHat Linux Kernel Multiple VulnerabilitiesBulletin detailing multiple vulnerabilities affecting RedHat Linux Kernel, including CVE-2024-56547 and others. Exploits can lead to denial of service, remote code execution, data manipulation, and security restriction bypass across various Red Hat Enterprise Linux and OpenShift Container Platform versions, impacting architectures like aarch64, s390x, ppc64le, and x86_64.
2026-06-01Critical Magento Cache Plugin Vulnerability Enables Remote Code Execution AttacksA critical vulnerability has been discovered in a Magento cache plugin, allowing remote code execution (RCE) attacks. This flaw could enable attackers to compromise Magento websites. Further details and potential mitigation strategies are available via the provided link. No specific bounty payout amount is mentioned in the content.
2026-06-01IBM WebSphere Server Vulnerable to Remote Code Execution Attack Via Crafted RequestIBM WebSphere Application Server has a critical remote code execution (RCE) vulnerability. Attackers can exploit this flaw by sending a specially crafted request, allowing them to execute arbitrary code on the server. This poses a significant security risk, potentially leading to unauthorized access and control of affected systems. Organizations using IBM WebSphere should prioritize patching and mitigating this vulnerability to protect their environments.
2026-06-01Critical Magento Cache Plugin Vulnerability Enables Remote Code Execution AttacksA critical vulnerability has been discovered in a Magento cache plugin, allowing attackers to execute arbitrary code remotely. This flaw poses a significant security risk for e-commerce stores using the affected plugin, as it could lead to complete system compromise. Merchants are strongly advised to immediately update or remove the plugin to mitigate potential attacks. The exact bounty payout for this vulnerability was not disclosed.
2026-06-01Critical Flowise Flaw Gives Attackers Full Server ControlLibrary for securing open-source AI platforms, specifically addressing CVE-2026-40933 in Flowise. This vulnerability allows attackers to achieve remote code execution (RCE) by importing a malicious workflow file, enabling them to run arbitrary server commands via the Custom MCP tool's stdio transport. The provided patch can be bypassed, and the most effective mitigation involves disabling stdio transport or switching to Server-Sent Events (SSE).
2026-06-01Weekly Recap: New Linux Flaw PAN-OS Exploit AI-Powered Attacks OAuth Phishing and MoreReference of actively exploited vulnerabilities including CVE-2026-0257 in PAN-OS, a critical zero-day RCE in Gogs, and multiple CVEs affecting WordPress, GitLab, and Microsoft products, with recent attacks leveraging AI for faster exploitation and information gathering, alongside the takedown of the GlassWorm C2 operation via trojanized VS Code extensions.
2026-06-01Palo Alto VPN bug graduates from advisory to active exploitationAnalysis of active exploitation of a Palo Alto VPN bug, highlighting the growing trend of AI-assisted attacks targeting API-driven applications and the emergence of AI malware like Shai-Hulud worming Red Hat npm packages. This shifts focus from traditional vulnerabilities to sophisticated AI-driven threats, with implications for election security and the rise of AI integration in development tools such as GitHub Copilot and RAD Studio extensions.
2026-06-01Windows Netlogon RCE exploited domain controllers at risk (CVE-2026-41089)Writeup on CVE-2026-41089, a critical Windows Netlogon RCE vulnerability, details its exploitation in the wild against domain controllers. This stack-based buffer overflow flaw allows remote code execution via crafted network requests. Microsoft has released patches, and Acros Security offers micropatches for legacy systems. Security teams are advised to patch immediately, restrict Netlogon traffic, and monitor for exploitation indicators like unexpected service crashes or anomalous traffic.
2026-06-01Critical Windows Netlogon RCE flaw now exploited in attacksWriteup on CVE-2026-41089, a critical Windows Netlogon RCE flaw, details its exploitation by threat actors. This stack-based buffer overflow in the Netlogon RPC interface allows unprivileged attackers to achieve remote code execution on domain controllers. Patched by Microsoft during May 2026 Patch Tuesday, it affects all supported Windows Server versions. The Centre for Cybersecurity Belgium has warned of active exploitation in the wild.
2026-06-01Flowises MCP implementation can run ghost commandsVulnerability in Flowise’s MCP stdio implementation, CVE-2026-40933, allows for one-click remote code execution in self-hosted deployments. Attackers can exploit a sandboxing failure in attacker-controlled MCP configurations, leading to server-side code execution with the privileges of the Flowise process, potentially granting root-level access in containerized environments. While Flowise has implemented several hardening measures, they have been found to be bypassable. The recommended complete mitigation is disabling MCP stdio by setting `CUSTOM_MCP_PROTOCOL=sse`.
2026-06-01Gogs Zero-Day Exposes Servers to Remote Code ExecutionWriteup of a Gogs zero-day vulnerability (CVSS 9.4) enabling remote code execution. Exploitable by authenticated attackers via pull requests with malicious branch names, this argument injection flaw allows for command execution as the Gogs server process user. Rapid7 developed a Metasploit module to automate the exploit and provided IoCs for detection. The vulnerability impacts default-configured Gogs servers on Windows, Linux, and macOS, particularly those with multiple user accounts.
2026-06-01Windows Netlogon 0-Click RCE Vulnerability Now Actively Exploited In The WildA critical 0-click Remote Code Execution (RCE) vulnerability in Windows Netlogon is now being actively exploited. This means attackers can compromise systems without any user interaction. The vulnerability, detailed in a linked article, poses a significant security threat to Windows environments. Details regarding specific exploit methods and potential mitigation strategies are likely available within the linked content, emphasizing the urgency for organizations to address this threat. No bug bounty payout amount is mentioned.
2026-05-30New 7-Zip security flaw could put hundreds of millions of systems at riskA newly discovered security vulnerability in 7-Zip, a popular file archiving utility used by millions, poses a significant risk to systems worldwide. The flaw, dubbed "7-Zip Double Vulnerability," allows attackers to execute arbitrary code. While no specific payout amount is mentioned, the potential impact is substantial, affecting numerous users and systems that rely on 7-Zip for file compression and decompression. Prompt patching and updates are recommended for users to mitigate this threat.
2026-05-30Notepad patches critical bugs: shortest path to malware executionNotepad++ has released patches for critical vulnerabilities that could allow for the "shortest path" to malware execution. These security flaws, if exploited, posed a significant risk to users by enabling malicious code to run. The software vendor has addressed these issues, urging users to update their Notepad++ installations to the latest version to protect themselves from potential threats. No specific payout amount for the discovered bugs was mentioned in the provided content.
2026-05-30Notepad vulnerabilities could enable arbitrary code execution on Windows systemsLibrary of information detailing two arbitrary code execution vulnerabilities, CVE-2026-48778 and CVE-2026-48800, affecting Notepad++ versions up to 8.9.6. These flaws, rated High (CVSS 7.8), allow local attackers to execute commands by manipulating `shortcuts.xml` and `config.xml` files. A third crash bug, CVE-2026-48770, was also patched. Exploitation requires the attacker to have write access to user profile directories or trick the user into opening a poisoned settings folder.
2026-05-30CVE-2026-35616: FortiClient EMS Flaw Actively Exploited in Malware AttacksWriteup of CVE-2026-35616, an actively exploited critical FortiClient EMS vulnerability with a CVSS score of 9.1, allowing unauthenticated remote code execution. Threat actors are abusing an improper access control flaw, bypassing API authentication to escalate privileges. Exploits have involved disguised Fortinet patches delivering the EKZ Infostealer malware, which exfiltrates credentials. CISA added this zero-day to its Known Exploited Vulnerabilities catalog. Fortinet has released hotfixes for versions 7.4.5 and 7.4.6, with a permanent fix in 7.4.7.
2026-05-30Exploit Code Published for Critical Flowise RCE VulnerabilityWriteup on CVE-2026-40933, a critical remote code execution vulnerability in Flowise impacting self-hosted instances. This systemic command injection flaw, originating from Anthropic's MCP protocol and facilitated by Flowise's unsafe serialization of stdio commands, allows attackers to execute arbitrary OS commands by convincing a user to import a crafted chatflow. Exploitation can lead to full server compromise, with credentials and connected services at risk. Obsidian Security has published proof-of-concept code demonstrating this exploit.
2026-05-30Imperva Customers Protected Against CVE-2026-45247 in Mirasvit Full Page Cache Warmer for MagentoImperva's services are protecting its customers from vulnerabilities in Mirasvit's Full Page Cache Warmer for Magento, specifically addressing CVE-2026-45247. This protection shields Magento e-commerce sites from potential attacks targeting this specific security flaw within the cache warmer extension. No bug bounty payout amount is mentioned.
2026-05-29No fix yet for critical Gogs RCE bug - exploit module is outWriteup on a critical Remote Code Execution (RCE) vulnerability in the open-source Git service Gogs. As of May 2024, no fix has been released, and an exploit module is publicly available. The researcher reported the vulnerability in March, but maintainers have not responded.

SSRF +14

DateResourceSummary
2026-06-05CVE-2026-20230: Cisco Unified CM WebDialer SSRF Can Lead to Root-Level CompromiseCisco Unified CM WebDialer is vulnerable to a Server-Side Request Forgery (SSRF) exploit, identified as CVE-2026-20230. This vulnerability allows an attacker to potentially achieve root-level compromise of the system. The issue stems from how WebDialer handles requests, enabling an attacker to trick the server into making unintended requests to internal or external resources. This could lead to unauthorized access and control over the Cisco Unified Communications Manager.
2026-06-04Doyensec: A #security update for #safeurl addressing an #SSRF bypass in non-default configs affecting those with IPv6 support enabled. A fix was released within 12 hours of report. Upgrade to version 0.2.4: Thanks to @tonghuaroot for the report! #doyensecDoyensec has released a security update (version 0.2.4) for safeurl to address an SSRF bypass vulnerability. This bypass affects non-default configurations, specifically for users with IPv6 support enabled. A fix was deployed within approximately 12 hours of the vulnerability being reported by @tonghuaroot. Users are advised to upgrade to the latest version. No bug bounty payout amount was specified.
2026-06-04Cisco Unified CM SSRF Flaw CVE-2026-20230: Public Exploit Code Opens Path to RootTool analysis of CVE-2026-20230, a critical Server-Side Request Forgery vulnerability in Cisco Unified Communications Manager's WebDialer Web Service. This flaw, CWE-918, allows arbitrary file writes and privilege escalation to root without authentication, with public exploit code now available. Mitigation involves disabling the WebDialer service or upgrading to patched versions (14SU6 or interim COP for 15SU5).
2026-06-04PoC Exploit Released for Cisco Unified Communications Manager Security VulnerabilityA Proof-of-Concept (PoC) exploit has been released for a security vulnerability affecting Cisco Unified Communications Manager (CUCM). This vulnerability, detailed in a Cisco security advisory, allows attackers to potentially gain unauthorized access or disrupt services. The release of the PoC means that exploits are now publicly available, increasing the risk for organizations using vulnerable CUCM versions. It is crucial for users to apply the necessary patches and updates provided by Cisco to mitigate this threat.
2026-06-04Critical Cisco Unified CM Bug Patched as Public Exploit Code EmergesAdvisory for CVE-2026-20230 details a critical vulnerability in Cisco Unified CM and Unified CM SME. Improper HTTP request validation allows unauthenticated remote attackers to perform SSRF attacks, potentially writing files to the OS for root privilege escalation. While public exploit code exists, exploitation requires the WebDialer service to be enabled, which is disabled by default. Mitigation involves disabling the WebDialer service until a patch is applied.
2026-06-04Cisco Unified Communications Manager Vulnerability Exposed Along With PoC Exploit CodeCisco Unified Communications Manager (CUCM) has a critical vulnerability that has been publicly disclosed. Researchers have released Proof of Concept (PoC) exploit code for this vulnerability, meaning attackers can readily use it to compromise affected systems. This vulnerability poses a significant risk to organizations relying on CUCM for their communication infrastructure. Further details about the specific nature and impact of the exploit are expected to emerge.
2026-06-04Cisco Warns of Available PoC for Critical Unified CM VulnerabilityAdvisory regarding CVE-2026-20230, a critical server-side request forgery (SSRF) vulnerability in Cisco Unified Communications Manager (Unified CM) and Unified CM SME. This flaw, with a CVSS score of 8.6, allows attackers to write files to the operating system, potentially leading to root privilege escalation. Proof-of-concept code is publicly available, though Cisco is unaware of active exploitation. The vulnerability impacts appliances with the WebDialer service enabled. Patches are available in Unified CM and Unified CM SME version 14SU6, with version 15SU5 expected in September.
2026-06-04Gray Hats: A critical Cisco Unified CM vulnerability is exposed. This public PoC code disclosure highlights a severe SSRF risk in WebDialer. #Cisco #SSRF #CVE202620230 #WebDialer #Cybersecurity #PoCA critical vulnerability in Cisco Unified Communications Manager (CM) has been publicly disclosed, featuring proof-of-concept (PoC) code. This exposes a severe Server-Side Request Forgery (SSRF) risk within the WebDialer component, identified by CVE-202620230. The disclosure, originating from Gray Hats, highlights a significant cybersecurity threat for Cisco users utilizing this feature.
2026-06-03How an Unauthenticated MCP Server Led to SSRF LFI and AWS Credential TheftWriteup of a real-world attack chaining Server-Side Request Forgery (SSRF), Local File Inclusion (LFI), and AWS credential theft via an unauthenticated Model Context Protocol (MCP) server. This finding highlights the rapid adoption of MCP for AI integrations, outpacing security maturity and leaving systems vulnerable to exploitation, as demonstrated by the theft of live AWS IAM credentials and database secrets from a customer environment.
2026-06-02How an Unauthenticated MCP Server Led to SSRF LFI and AWS Credential TheftLibrary for discovering and analyzing security risks in AI integrations. It highlights vulnerabilities in Model Context Protocol (MCP) servers, such as Server-Side Request Forgery (SSRF) and Local File Inclusion (LFI), demonstrated by a real-world case study where an unauthenticated MCP server led to AWS credential theft. The library addresses the growing attack surface created by rapid AI adoption and the security maturity gap in new AI-native technologies.
2026-06-02xHackInSeconds: URL field accepted internal addresses. 169.254.169.254. IAM role credentials in the response. Full S3 and RDS access. #infosec #cloud #ssrfxF HackInSeconds discovered a vulnerability where a URL field accepted internal IP addresses, including the metadata service address 169.254.169.254. This allowed attackers to retrieve IAM role credentials from the response. With these credentials, they gained full access to S3 buckets and RDS instances. This highlights a critical Server-Side Request Forgery (SSRF) flaw in cloud environments.
2026-06-02Hi: that NASA XXE chain is filthy. multiline DOCTYPE bypass SSRF chaining to get AWS metadata is exactly the kind of creative attack chain that gets $50k bounties. insane find honestly #WebSecurity #SSRFA researcher discovered a "filthy" XXE (XML External Entity) chain attack against NASA. This sophisticated attack involves a multiline DOCTYPE bypass to achieve SSRF (Server-Side Request Forgery), ultimately enabling the retrieval of AWS metadata. This type of creative attack chain is highly valued in bug bounty programs, with potential payouts exceeding $50,000. The find is described as "insane" and highlights advancements in web security exploitation.
2026-06-02Gray Hats: A critical Apache Fesod SSRF vulnerability has been found. Learn about the CVE-2026-49328 patch released to protect internal network resources. #ApacheFesod #SSRF #CVE202649328 #JavaSecurity #CybersecurityA critical Server-Side Request Forgery (SSRF) vulnerability, identified as CVE-2026-49328, has been discovered in Apache Fesod. A patch has been released to address this security flaw, which could have allowed attackers to access internal network resources. The vulnerability highlights ongoing concerns within Java security and the broader cybersecurity landscape.
2026-06-01TechEarl: SSRF makes a server fetch what the attacker chooses; in the cloud that means 169.254.169.254 leaking credentials. Why allowlists beat blocklists plus IMDSv2. #SSRF #WebSecurity techearl.com/server-side-reqServer-Side Request Forgery (SSRF) allows attackers to force servers to fetch arbitrary data. In cloud environments, this vulnerability can expose sensitive credentials by targeting the Instance Metadata Service (IMDS) at `169.254.169.254`. The article highlights the superiority of allowlists over blocklists for security and mentions the importance of Instance Metadata Service Version 2 (IMDSv2) in mitigating such risks.

API Security +11

DateResourceSummary
2026-06-04Hugging Face Transformers contains critical remote code execution vulnerabilityLibrary vulnerability detailing critical remote code execution in Hugging Face's Transformers, tracked as CVE-2026-4372 and CVE-2026-1839. The flaw allows arbitrary code execution during routine model loading, even when `trust_remote_code=False`, due to deserialization of untrusted data. Exploits leverage crafted model configurations or malicious checkpoints with unsafe `torch.load()` calls. Patches are available in Transformers 5.3.0 and 5.0.0rc3 respectively, addressing widespread exposure in ML pipelines and inference services.
2026-06-03Critical StrongDM Vulnerability Allow Attackers to Steal and Reuse AuthenticationA critical vulnerability in StrongDM has been discovered, allowing attackers to steal and reuse authentication credentials. This security flaw poses a significant risk by compromising user access. The full details and implications of this exploit are still being assessed, but it highlights a serious concern for organizations using StrongDM for access management.
2026-06-031-Click GitHub Vulnerability Enables OAuth Token TheftA critical vulnerability in GitHub's OAuth application flow allowed attackers to steal OAuth tokens with a single click. This exploit leveraged a misconfiguration that enabled the redirection of authenticated users to malicious websites. Once redirected, attackers could trick users into granting permissions, effectively gaining unauthorized access to their GitHub accounts and associated data. This significant security flaw highlights the importance of robust authentication and authorization mechanisms.
2026-06-03Critical WP Maps Pro Vulnerability Allow Attackers to Create Administrator AccountCritical WP Maps Pro Vulnerability Allow Attackers to Create Administrator Account https://ift.tt/6ykBmfY
2026-06-031-Click GitHub Token Vulnerability Lets Attackers Steal Users' OAuth TokensA severe vulnerability in GitHub's web application allows attackers to steal users' OAuth tokens with a single click. This exploit targets how GitHub handles certain types of URLs, enabling malicious actors to trick users into clicking a specially crafted link. Upon clicking, the attacker can gain access to sensitive user data and potentially perform actions on their behalf. The vulnerability was disclosed and has since been patched by GitHub, though specific payout details were not mentioned in the provided content.
2026-06-02Claude Code's Vulnerability in GitHub Actions Allows an Attacker to Compromise any RepositoryA critical vulnerability has been discovered in Claude Code's integration with GitHub Actions. This flaw enables an attacker to compromise any repository utilizing the service. The exact payout amount for this bug bounty was not explicitly stated in the provided content.
2026-06-02Web Application & API Attacks Are Rising:Are You Blind to Modern Web Attacks? Join WAAP Security...Web application and API attacks are on the rise, leaving organizations vulnerable to modern threats. The provided content, titled "Web Application & API Attacks Are Rising: Are You Blind to Modern Web Attacks? Join WAAP Security...", suggests that businesses may be unaware of these escalating dangers. It promotes WAAP Security as a solution to address these blind spots and improve defenses against contemporary web attacks. The content does not mention any specific bug bounty payout amounts.
2026-06-02Financial Data Crisis! ChatGPT Spreadsheet Plugin Exposes Serious Security VulnerabilityA critical security vulnerability has been discovered in a ChatGPT spreadsheet plugin, potentially exposing sensitive financial data. The plugin's flaw could allow unauthorized access to user information. This presents a significant "Financial Data Crisis" and highlights the risks associated with integrating third-party tools into AI platforms. Users are advised to exercise caution and consider disabling the plugin until a fix is implemented. The exact payout amount for reporting this vulnerability is not specified in the provided content.
2026-06-01Eliminate Critical API Attack Paths with Wiz API SPMLibrary for continuous, agentless API discovery across AWS, Azure, and GCP environments. It assesses API exploitability by simulating attacker techniques, identifying "toxic combinations" where exposed APIs lead to sensitive data compromise. The library prioritizes remediation efforts and offers actionable guidance, even supporting automated workflows like triggering Terraform patches. It integrates API and cloud security context within a Security Graph, revealing attack paths and risks like SQL injection vulnerabilities or Broken Object Level Authorization.
2026-05-31Anthropic AI Vulnerability Scanner in Enterprise Beta: IBM Joins Glasswing After 10000 Flaws FoundTool for AI-powered application security scanning, Claude Security, now in public beta for enterprise customers, identifies vulnerabilities by reasoning over code behavior and data flows, moving beyond traditional signature matching. This approach has surfaced over 10,000 critical software flaws through Anthropic's Project Glasswing consortium, which includes IBM, and has also revealed specific vulnerabilities like CVE-2026-5194 in wolfSSL. The tool aims to compress the find-fix cycle, though patching remains a bottleneck for maintainers.
2026-05-30Vibe Coding Security: Why 62% Of AI-Generated Code Ships With VulnerabilitiesLibrary analyzing security flaws in AI-generated code, including Cross-Site Scripting (XSS) and Server-Side Request Forgery (SSRF). Research indicates AI code exhibits significantly higher vulnerability rates than human-written code, with studies highlighting failures in XSS defenses, exposed secrets, PII, and lack of CSRF protection and security headers across platforms like Cursor and Claude Code.

AI +5

DateResourceSummary
2026-06-03Guardrails for AI Agents: Safety and SecurityLibrary providing a layered governance and security system for AI agents, acting as a runtime control to prevent issues like hallucinations, prompt injection, unsafe actions, and data leakage by validating inputs, model outputs, and tool calls. It enforces structured policies and safeguards through pre-LLM input checks, post-LLM output and action validation, and system-level controls such as least privilege and tool sandboxing. This approach treats guardrails as production infrastructure, incorporating context-grounded validation, self-correction loops, multi-agent validation, and hard constraints to ensure security, compliance with regulations like GDPR and HIPAA, and prevent operational incidents.
2026-06-02Snowflake Bolsters AI SecurityLibrary integrating native, proactive, enterprise-grade security for AI workloads, focusing on agent security, data security, and platform-level security. Features include Agent Identity for distinct AI agent actions, enabling auditability and access restrictions to sensitive data, complementing Snowflake Horizon Catalog for AI governance.
2026-06-02What Is LLM (Large Language Model) Security?Guide to LLM security covering fundamental concepts, prominent risks like prompt injection and data leakage, and real-world attack examples such as Microsoft's Tay and PoisonGPT. It emphasizes that LLM security differs from traditional app security due to the probabilistic nature of models, and it details practical implementation strategies across the LLM lifecycle to mitigate vulnerabilities.
2026-06-02You cant patch your way out of prompt injection: AI agents need a different defenseLibrary for defending against prompt injection in AI agents, emphasizing structural defenses over filters. It addresses vulnerabilities like EchoLeak (CVE-2025-32711) and ShareLeak (CVE-2026-21520) by mitigating the "lethal trifecta" of private data access, untrusted content exposure, and outbound communication. The library promotes treating source text as data, scoping agent capabilities, and implementing strict data-flow and control-flow rules, inspired by research like Google DeepMind's CaMeL.
2026-06-01ChatGPhish Reveals ChatGPT Browser Prompt Injection RiskLibrary that demonstrates browser-based prompt injection against ChatGPT, named ChatGPhish, allows attackers to manipulate page summaries and deliver phishing or social engineering attacks. This technique bypasses traditional security controls by injecting malicious instructions into ordinary web pages, influencing the LLM's output within the trusted ChatGPT interface. The research highlights risks associated with rendering untrusted Markdown content, including a QR code delivery method that circumvents desktop browser protections.

XSS +4

DateResourceSummary
2026-06-04Cisco Webex Meetings Cross-Site Scripting Vulnerability (CVE-2026-20233)Writeup of CVE-2026-20233, a cross-site scripting (XSS) vulnerability in Cisco Webex Meetings. The flaw stemmed from insufficient user input validation, allowing an unauthenticated remote attacker to execute arbitrary script code or access sensitive browser information by tricking a user into clicking a malicious link. Cisco has resolved this issue in their cloud-based Webex Meetings service, requiring no customer action.
2026-06-03Zero-Click pretalx XSS Flaw Lets Hackers Hijack Conference Organizer AccountsLibrary for detecting stored XSS vulnerabilities, exemplified by CVE-2026-41241 in pretalx, which allows zero-click account hijacking. This flaw, exploitable with low privileges, bypasses Content Security Policies by leveraging chained exploits involving JavaScript payloads disguised as presentation materials and iframe `srcdoc` attributes. A secondary JavaScript-free technique demotes administrators via image tags in submission titles, triggering a superuser-demotion endpoint. Automated AI agents can weaponize this for mass exploitation across numerous conferences.
2026-06-03https://github.com/Armur-Ai/Pentest-Swarm-AILibrary for advanced penetration testing utilizing a real swarm intelligence architecture. It coordinates independent agents via stigmergy and emergence, allowing them to coordinate by writing to and reading from a shared blackboard, rather than through a central planner. This approach enables emergent attack chains and dynamic agent interaction, supporting tools like nmap, sqlmap, Burp, ZAP, and Metasploit, and is compatible with LLMs such as Claude and Llama.
2026-06-03House committee chair calls on Instructure to testify in Canvas hackWriteup on the Shiny Hunters attack on Instructure's Canvas platform, highlighting cross-site scripting (XSS) vulnerabilities exploited to hijack admin sessions and exfiltrate student data. The incident prompted a US House committee inquiry, emphasizing the continued relevance of foundational security flaws like input validation and output encoding in critical educational technology infrastructure, despite focus on novel AI threats.

SQLi +4

DateResourceSummary
2026-06-02AnonymousPostgreSQL Injection in Drupal Core (CVE-2026-9082)Drupal Core is vulnerable to an anonymous PostgreSQL injection (CVE-2026-9082). This flaw allows unauthenticated users to inject arbitrary PostgreSQL code into a database. The vulnerability arises from insufficient sanitization of user-supplied data. Successful exploitation could lead to data manipulation, disclosure, or even remote code execution. Users are advised to update Drupal Core to the latest patched version as soon as possible to mitigate this critical security risk.
2026-06-01Exploitation of Critical SQL Injection Vulnerability in Drupal (CVE-2026-9082)Writeup detailing the exploitation of CVE-2026-9082, a critical SQL injection vulnerability in Drupal affecting PostgreSQL databases. This unauthenticated flaw, rated 9.8 (CVSS), allows arbitrary SQL execution via crafted requests. CISA has added it to their KEV catalogue, with over 15,000 exploitation attempts detected across various sectors, primarily in the United States. Drupal recommends upgrading to the latest supported release or applying backported fixes.
2026-05-30700 education and tech websites hijacked in huge ClickFix malware campaignAnalysis of CVE‑2026‑26980, a critical SQL injection vulnerability affecting Ghost CMS versions 3.24.0 through 6.19.0, details how attackers exploited this flaw to hijack over 700 websites. The vulnerability allowed for the theft of administrative API keys, enabling attackers to inject JavaScript that presented fake Cloudflare or CAPTCHA verification dialogs, tricking visitors into running Windows commands to install malware through ClickFix attacks.
2026-05-29Critical Roundcube Flaw Allows Attackers to Inject SQL QueriesWriteup detailing critical vulnerabilities in Roundcube Webmail, including a pre-authentication SQL injection flaw in the `virtuser_query` plugin via `preg_replace` backslash escape bypass, allowing arbitrary database queries. The update also addresses code injection via the LDAP `autovalues` option and other issues like stored XSS, CSS injection, SSRF bypass, and session poisoning. These vulnerabilities are patched in versions 1.6.16 and 1.7.1.

Python +2

DateResourceSummary
2026-06-05Type Level Security for Secure AI Code GenerationThis content discusses type-level security for AI code generation. It aims to build more robust and secure AI systems by incorporating security considerations directly into the type system. This approach helps prevent vulnerabilities during the code generation process, ultimately leading to safer AI applications. The provided link likely contains further details on the technical implementation and benefits of this security methodology.
2026-06-01Millions of AI agents imperiled by critical vulnerability in open source packageLibrary implementing ASGI for Python applications, Starlette, is vulnerable to CVE-2026-48710 (BadHost), allowing attackers to bypass authorization via a modified HTTP Host header. This critical vulnerability, trivially exploitable without proper firewalling, impacts FastAPI, vLLM, LiteLLM, and millions of AI agents relying on Starlette, potentially exposing sensitive data and third-party credentials.

Bug Bounty +2

DateResourceSummary
2026-06-03Microsoft under fire for threatening security researcher with criminal investigationWriteup detailing Microsoft's controversial actions against security researcher "Nightmare Eclipse" for publicly disclosing unpatched vulnerabilities, including BlueHammer, RedSun, UnDefend, and YellowKey, affecting products like Defender and BitLocker. This incident highlights ongoing debates around researcher disclosure responsibilities versus company protection, with cybersecurity veterans like Katie Moussouris and Kevin Beaumont criticizing Microsoft's threats of legal action and prosecution, potentially discouraging future vulnerability reporting.
2026-05-30Adobe expands bug bounty program to incentivize AI security research| Adobe Security BlogProgram announcement detailing Adobe's expanded bug bounty initiative to incentivize AI security research. It introduces an 'AI Bonus Tier' with rewards up to $15,000 for critical AI-related vulnerabilities. The scope now includes specific AI features across web applications like Acrobat AI Assistant, Firefly Image Models, and Photoshop AI Assistant, along with explicit guidance on AI vulnerability classes such as prompt injection and model abuse. Future expansion will cover mobile and desktop products.

GraphQL +1

DateResourceSummary
2026-06-02GitLab Patches Multiple Duo AI DoS and Authorization Flaws in Community and Enterprise EditionGitLab has released security updates to address critical vulnerabilities affecting its Duo AI, Denial-of-Service (DoS), and authorization features. These flaws, present in both Community and Enterprise Editions, could allow attackers to disrupt service or gain unauthorized access. Users are strongly advised to update to the latest versions to mitigate these risks. No specific bounty payout amounts were mentioned in the provided content.