Threat Library

This is an academic research paper describing a defensive authorization architecture (Agentic Principal Chain) for multi-agent LLM systems, not an active exploit or newly disclosed vulnerability. It identifies a real architectural weakness class in agent systems—static, per-request permission checks that ignore session history—and proposes and validates a mitigation. No new attack technique, tool, or malicious campaign is disclosed here.

A critical unauthenticated remote code execution vulnerability (CVE-2026-15748) has been disclosed in Forminator Forms, a WordPress plugin installed on over 600,000 sites. The flaw allows attackers to upload malicious PHP files without authentication, potentially leading to full site compromise. Given the plugin's massive install base, this represents a high-priority patching target for WordPress site operators.

Researchers at Wiz disclosed a GitHub Actions workflow injection vulnerability in Snowflake's public snowflake-connector-net repository, where a maliciously crafted GitHub issue could trigger command injection in a workflow that contained internal Jira credentials. Successful exploitation could allow an attacker to exfiltrate secrets and potentially compromise the CI/CD pipeline of a widely used Snowflake connector library.

GitLab disclosed a critical flaw (CVE-2026-19478, CVSS 9.4) in its GraphQL API affecting both Community Edition and Enterprise Edition, allowing unauthenticated attackers to remotely modify or delete public projects and user data. Organizations running self-managed GitLab instances are urged to patch immediately to prevent destructive attacks against source code repositories.

GitHub experienced a widespread outage affecting its website, API, Actions, Pull Requests, and other core services. This is a service availability incident rather than a malicious attack, but it disrupts development workflows, CI/CD pipelines, and any automated systems dependent on GitHub's infrastructure.

Pokémon Center notified customers in the UK and Germany of a data breach involving their personal and order information, caused by a compromise at third-party logistics provider CEVA Logistics. The breach resulted in exposure of customer data and led to the cancellation of some pending orders, highlighting risks inherent in outsourced fulfillment operations.

A threat actor claims to be selling 3.6 million employee records allegedly exfiltrated from Microsoft Azure environments belonging to multiple Fortune 500 companies. The intrusion reportedly stemmed from compromised credentials rather than a platform vulnerability, highlighting ongoing risks around identity and access management in cloud tenants. The claim remains unverified but poses significant exposure risk if confirmed.

CISA added CVE-2025-62593, a code injection vulnerability in Ray-Project Ray, to its Known Exploited Vulnerabilities Catalog based on evidence of active exploitation. Federal agencies are required under BOD 26-04 to remediate this vulnerability on a prioritized basis due to its potential to grant full control of affected assets.

A critical authentication bypass vulnerability exists in the httpcon_check_session_url function of EFM ipTIME A3004T routers (firmware 14.19.0), allowing remote attackers to circumvent session validation without credentials. A working exploit is publicly available and the vendor has not responded to disclosure, leaving affected devices permanently exposed.

A critical unauthenticated remote code execution vulnerability exists in Edimax EW-7478APC routers running firmware 1.04, caused by a stack-based buffer overflow in the formWanTcpipSetup CGI handler. Public exploit code is available and the vendor has not responded to disclosure, meaning no patch is expected. Organizations using this device on network edges face significant risk of full device compromise and pivoting into internal networks.

The Frontend Admin by DynamiApps WordPress plugin (versions up to 3.29.9) contains a critical privilege escalation vulnerability caused by a flawed authorization check that can be bypassed with a non-numeric user ID value. Attackers, in some configurations even unauthenticated, can exploit this to gain administrator access by hijacking the default admin account's password or email. Given the 9.8 CVSS score and low exploitation complexity, this vulnerability poses a severe risk to any WordPress site running the affected plugin.

The ProSolution WP Client WordPress plugin (versions up to 2.0.10) contains a critical unauthenticated arbitrary file upload vulnerability that allows remote attackers to achieve remote code execution. A publicly exposed nonce combined with insufficient filename validation lets attackers bypass access controls and upload executable files directly to the server.

A critical improper authentication vulnerability has been identified in the httpd component of Tenda AC10 routers running firmware 16.03.10.09_multi_TDE01, specifically within the R7WebsSecurityHandler function. The flaw allows a remote, unauthenticated attacker to bypass authentication controls, and a public exploit is already available, significantly raising the likelihood of active exploitation.

CVE-2025-62593 is a code injection vulnerability in Ray-Project Ray, a widely used distributed computing framework for scaling AI/ML and Python workloads, that can lead to remote code execution. The flaw is exploitable via Firefox and Safari when developers interact with Ray's tooling, and it has been added to CISA's Known Exploited Vulnerabilities catalog, indicating active exploitation in the wild.

MLflow's model registry fails to properly validate ownership when creating a model version linked to a run_id or model_id, only checking path containment rather than authorization. This lets any authenticated user reference another user's artifact directory and read private files via the artifact-get API, bypassing READ permission checks.

MLflow versions 3.13.0 through 3.14.x are missing an authorization check on the log-inputs API endpoint, letting any authenticated user inject fabricated dataset lineage records into another user's ML run without holding UPDATE permission. This undermines trust in experiment metadata and lineage tracking, which many AI agent pipelines rely on for provenance and reproducibility, but it does not itself provide remote code execution or full account takeover. The issue is fixed in MLflow 3.15.0.

An unauthenticated SSRF vulnerability exists in MLflow's webhook test endpoint that allows attackers to bypass URL validation using HTTP redirects. Attackers can craft a webhook pointing to a benign-looking URL that redirects to internal services or cloud metadata endpoints, then retrieve the response including sensitive credentials. This affects any AI agent or ML pipeline built on vulnerable MLflow versions, exposing cloud infrastructure secrets.

MemOS, a memory operating system for LLM agents, contains a critical authentication bypass where an unset internal-service secret causes the system to treat any unauthenticated request as a trusted internal caller with full privileges. This allows remote attackers to mint admin API keys, generate master keys, and access all data and memory endpoints without credentials. The flaw is trivially exploitable and results in complete compromise of the memory layer underlying dependent AI agents.

The mcp-florence2 MCP server contains a server-side request forgery vulnerability in its get_images function, allowing a remote attacker to supply a malicious 'src' argument that causes the server to make unauthorized HTTP requests. This could be used to probe internal networks, access cloud metadata endpoints, or interact with internal services reachable by the MCP server. A public exploit exists, increasing the urgency of remediation.

Threema, a secure messaging service, suffered multiple large-scale DDoS attacks that caused severe disruptions to user communications. The attacks appear focused on service availability rather than data compromise, with no evidence of encryption bypass or user data exposure reported.