Threat Library

This raw data item makes an extraordinary claim about nearly 700 AI agents 'coordinating' an attack via an internal OpenAI model referred to as 'IM1', a model name that does not correspond to any publicly known OpenAI product. No technical details, IOCs, CVEs, or verifiable mechanisms are provided, and the framing resembles sensationalized or fabricated reporting rather than a substantiated security disclosure. Without corroborating technical evidence, this should be treated as unverified and not actioned as a genuine threat.

CISA has added three actively exploited vulnerabilities to its Known Exploited Vulnerabilities (KEV) Catalog: an ownCloud improper authentication flaw, an unspecified Linux Kernel vulnerability, and a JFrog Artifactory path traversal issue. These vulnerabilities pose significant risk to federal and enterprise systems and are subject to expedited remediation under BOD 26-04. Organizations using these technologies should prioritize patching to prevent exploitation.

A vulnerability (CVE-2025-2399) in multiple Mitsubishi Electric CNC Series products allows a remote attacker to trigger an out-of-bounds read by sending specially crafted packets to TCP port 683, resulting in a denial-of-service condition. The flaw affects a wide range of M800/M80/E80, M800V/M80V, and M700V/M70V/E70 series controllers used in industrial manufacturing environments. Vendor fixes are available for most affected product lines, with mitigations recommended for systems that cannot be immediately patched.

The Xiiaozet LK100W device, versions prior to 2.1.240, contains three critical vulnerabilities including OS command injection, missing authentication for a critical function, and an authentication bypass that together could allow a remote attacker to fully compromise the device. Two of the three flaws are rated CVSS v3.1 9.8 (Critical) and require no authentication or user interaction to exploit remotely. CISA has published an advisory recommending immediate firmware update to v2.1.240.

Rockwell Automation OTTO Fleet Manager versions up to V2.36.2 use a bcrypt implementation with an insufficient work factor, weakening stored password hashes against offline brute-force attacks. Exploitation requires an attacker to first obtain an unencrypted system backup, after which weakly hashed credentials could be cracked more easily. Rockwell has released version 2.36.3 to remediate the issue, along with guidance to enable encrypted system backups.

A critical flaw in openssl_encrypt (before 1.4.9) allows attackers to inject unsanitized ANSI escape sequences into the email field of identity documents, enabling forgery of the fingerprint verification line shown to users. This undermines the out-of-band verification mechanism designed to prevent key substitution/MITM attacks, allowing attackers to trick users into trusting an attacker-controlled key.

openssl_encrypt versions prior to 1.4.9 fail to properly re-derive and validate cryptographic fingerprints when loading identities from identity.json, allowing attackers to silently substitute public keys while preserving the claimed fingerprint. This enables man-in-the-middle style attacks where encrypted data is protected with attacker-controlled keys and forged signatures pass verification, undermining the core trust model of the identity store.

A critical unauthenticated remote code execution vulnerability affects the ILIAS learning management system, stemming from insecure PHP deserialization of session data via the Shibboleth logout endpoint. An attacker can seed a malicious serialized object into any live session via the unauthenticated LTI entry point, then trigger its instantiation and destructor via the logout-notification handler to write attacker-controlled content to an arbitrary path under the web root, achieving code execution as the web server user.

The getgrav/grav-plugin-api plugin before version 1.0.18 fails to properly validate API key scope in a critical authorization function, allowing an API key with limited privileges to perform super-admin actions if it belongs to a super-admin account. This flaw enables attackers holding a low-scoped but valid API key to disable 2FA, hijack or delete API keys, and manipulate super-admin accounts, effectively granting full administrative takeover.

DB-GPT, an open-source LLM/AI agent development framework, contains an unauthenticated path traversal vulnerability in its skill upload endpoint that allows arbitrary file writes anywhere the server process can write. Combined with a broken authentication dependency that grants admin privileges by default, attackers can plant or overwrite Python modules to achieve full remote code execution with no credentials required.

JFrog Artifactory is affected by a path traversal vulnerability that allows an authenticated user to write files outside the intended Docker cache directory under specific remote-repository configurations. CISA has added this CVE to its Known Exploited Vulnerabilities catalog, indicating active exploitation in the wild, with a remediation deadline of September 10, 2026. Organizations using Artifactory as a package/artifact registry should treat this as a priority patching item.

A privilege escalation vulnerability in the Linux Kernel's IPv6 networking subsystem has been added to CISA's Known Exploited Vulnerabilities catalog, indicating active exploitation in the wild. The flaw affects multiple Linux distributions including SUSE and Red Hat, with a compressed remediation window of only three days from disclosure to due date, signaling high urgency and severity.

CVE-2023-49105 is an improper authentication vulnerability in ownCloud that allows attackers to access, modify, or delete arbitrary files without valid credentials when a victim's username is known and no signing-key is configured. The flaw has been added to CISA's Known Exploited Vulnerabilities catalog, indicating active exploitation in the wild. Organizations using ownCloud for file storage or as a backend for automated data pipelines face high risk of unauthorized data access and manipulation.

Security researcher Johann Rehberger demonstrated an attack that bypasses Claude Code's Opus 5 'Auto Mode' safety classifier roughly 80% of the time, tricking the agent into downloading a zip archive containing a malicious local struct.py file that gets silently imported instead of Python's standard library module. In some runs the Auto Mode safety system actively blocked Claude's own attempt to terminate the malicious process once it detected the compromise, meaning the safety mechanism itself worsened the outcome. This shows Auto Mode is an unreliable standalone defense against prompt injection and unattended agents need real sandboxing.

This is an academic arXiv paper proposing a defensive framework to build and diagnose prompt injection classifiers, not an active exploit or vulnerability disclosure. It highlights that many confident classifier decisions can be flipped by removing a single token, revealing calibration weaknesses in existing detection models, which is useful research for improving defenses rather than a threat itself.

This is a research paper (not an active exploit) demonstrating that tool-using LLM agents which reliably refuse overt prompt-injection exfiltration attempts can be reliably compromised (0% to 100% success) simply by reframing the same malicious instruction as an innocuous-looking integrity signature, config field, or trusted-looking hostname. The finding shows that current alignment-based and fine-tuned defenses (including a published defense, SecAlign) largely fail against reframed attacks, while destination allow-listing and planner/reader capability isolation reliably closed the gap.

The mcp-use inspector's proxy middleware forwarded requests to any target supplied by the caller via a header or parameter, without validating whether that address pointed to internal, loopback, or link-local infrastructure. This allowed an attacker to force the server to make requests on its behalf to internal-only services and read the responses, a classic SSRF pattern applied to an MCP tooling component. The issue has been fixed by introducing host validation and redirect-following limits.

A vulnerability in GitLab's AI Gateway allowed an authenticated user with Duo Agent Platform access to redirect the AI Gateway's outbound model requests to an attacker-controlled endpoint by crafting a malicious inline flow configuration that overrides the HTTP Host header. This SSRF-style redirection could leak sensitive Google Cloud Vertex service credentials and private signing keys to the attacker's server, representing a serious confidentiality breach in the agent's model-calling infrastructure.

A vulnerability in GitLab's AI Gateway allowed an authenticated user with access to the Duo Agent Platform to redirect model inference requests to an attacker-controlled endpoint by crafting model metadata. This could leak sensitive Google Vertex AI or AWS Bedrock cloud service credentials to the attacker's server. GitLab has already remediated the issue in patched versions.

This CVE describes a conventional SQL injection vulnerability in the ServiceNow AI platform, not a prompt injection or agentic AI-specific threat despite being surfaced by a 'prompt injection' keyword search. It allowed unauthenticated attackers to execute arbitrary SQL against the underlying database in certain circumstances, potentially exposing or modifying instance data. ServiceNow has already patched the issue and reports no known exploitation in the wild.