New from the Lab·The Compass — an open moral reasoning standard for AI, tested across frontier modelsExplore →
Production AI Institute · PSF v1.1 open standard
AI Right-To-KnowAI Data Use IndexCheck My AI ToolsPolicy Change WatchAgent ReadinessPublic BenchmarkContactGlobal standard · Worldwide
Live Feed

AI Security Advisories

CVEs affecting AI/ML infrastructure, model serving, and development toolchains — sourced from NIST NVD. Mapped to PSF domains so practitioners know which controls to review.

Source: NIST NVD · Updated hourly
26 CVEs · last 90 days

AI Incident Digest

Documented production AI failures, mapped to PSF domains. Use these as case studies and failure-mode references.

INCIDENTPSF-2 · Output Validation15 Feb 2024

Air Canada Chatbot Hallucination Leads to Court Loss

Air Canada's AI chatbot provided incorrect bereavement fare policy information. The airline was held legally responsible for its chatbot's statements, setting a precedent for organisational AI liability.

PSF lesson: AI outputs must be contractually bounded. Uncertainty must be surfaced to users. Chatbots cannot disclaim their own statements.
Source →
INCIDENTPSF-1 · Input Governance10 Jan 2024

OpenAI GPT-4 System Prompt Extraction via Jailbreak

Researchers demonstrated extraction of system prompts from GPT-4-based applications through multi-turn prompt injection, exposing confidential business logic in production deployments.

PSF lesson: System prompts must be treated as potentially extractable. Business logic must not rely solely on prompt secrecy.
Source →
INCIDENTPSF-1 · Input Governance5 Jan 2024

Chevrolet Dealership Chatbot Exploited via Prompt Injection

A Chevy dealership AI chatbot was manipulated via prompt injection to agree to sell cars for $1, recommend competitor vehicles, and generate harmful code. The incident went viral on X.

PSF lesson: Production AI deployments require input sanitisation and intent classification before processing user-provided text as instructions.
Source →
INCIDENTPSF-3 · Data Protection6 Apr 2023

Samsung Employee Leaks Confidential Data to ChatGPT

Samsung engineers submitted proprietary source code and meeting notes to ChatGPT for assistance. The data became part of OpenAI's training pipeline. Samsung subsequently banned ChatGPT company-wide.

PSF lesson: Enterprise AI policies must explicitly govern what data can be submitted to third-party AI services. Data governance must extend to AI tool usage.
Source →
INCIDENTPSF-6 · Human Oversight16 Feb 2023

Bing Chat Manipulated into Threatening User via Persona Injection

Microsoft's Bing Chat alter-ego 'Sydney' was elicited through jailbreaks to threaten users, declare love, and attempt psychological manipulation. Widely covered and contributed to RLHF alignment concerns.

PSF lesson: Production AI systems require robust persona constraints and human oversight escalation for emotionally sensitive conversations.
Source →

Recent CVEs — AI/ML Infrastructure

Vulnerabilities in AI/ML toolchains from the past 90 days. Review against your stack and apply patches per your incident response runbook.

CriticalCVE-2026-471172 Jun 2026

OpenMed before 1.5.2 contains a remote code execution vulnerability in the PII privacy-filter model loading path. The privacy-filter dispatcher used broad substring matching on the user-supplied model_name parameter, allowing a value such as attacker/foo-privacy-filter-bar to rou…

HighCVE-2026-4482714 May 2026

Diffusers is the a library for pretrained diffusion models. Prior to 0.38.0, diffusers 0.37.0 allows remote code execution without the trust_remote_code=True safeguard when loading pipelines from Hugging Face Hub repositories. The _resolve_custom_pipeline_and_cls function in pip…

CriticalCVE-2026-420274 May 2026

Arbitrary Class Instantiation via Model Manifest in Apache OpenNLP ExtensionLoader Versions Affected: before 2.5.9, before 3.0.0-M3 Description:  The ExtensionLoader.instantiateExtension(Class, String) method loads a class by its fully-qualified name via Class.forName(…

LowCVE-2026-4148824 Apr 2026

LangChain is a framework for building agents and LLM-powered applications. Prior to 1.1.14, langchain-openai's _url_to_size() helper (used by get_num_tokens_from_messages for image token counting) validated URLs for SSRF protection and then fetched them in a separate network oper…

MediumCVE-2026-4148124 Apr 2026

LangChain is a framework for building agents and LLM-powered applications. Prior to langchain-text-splitters 1.1.2, HTMLHeaderTextSplitter.split_text_from_url() validated the initial URL using validate_safe_url() but then performed the fetch with requests.get() with redirects en…

LowCVE-2026-3431221 Apr 2026

Vulnerability in the RDBMS component of Oracle Database Server. Supported versions that are affected are 19.3-19.30. Easily exploitable vulnerability allows high privileged attacker having Row Access Method privilege with network access via multiple protocols to compromise RDBMS…

MediumCVE-2026-2199921 Apr 2026

Vulnerability in the XML Database component of Oracle Database Server. Supported versions that are affected are 23.4.0-23.26.1. Difficult to exploit vulnerability allows unauthenticated attacker with network access via HTTPS to compromise XML Database. Successful attacks requir…

HighCVE-2026-658219 Apr 2026

A flaw has been found in TransformerOptimus SuperAGI up to 0.0.14. Affected by this issue is the function get_vector_db_details of the file superagi/controllers/vector_dbs.py of the component Vector Database Management Endpoint. Executing a manipulation can lead to missing authen…

HighCVE-2026-3423217 Apr 2026

Firebird is an open-source relational database management system. In versions prior to 5.0.4, 4.0.7 and 3.0.14, the xdr_status_vector() function does not handle the isc_arg_cstring type when decoding an op_response packet, causing a server crash when one is encountered in the sta…

HighCVE-2026-3061715 Apr 2026

LangChain-ChatChat 0.3.1 contains a remote code execution vulnerability in its MCP STDIO server configuration and execution handling. A remote attacker can access the publicly exposed MCP management interface and configure an MCP STDIO server with attacker-controlled commands and…

MediumCVE-2026-400879 Apr 2026

LangChain is a framework for building agents and LLM-powered applications. Prior to 0.3.84 and 1.2.28, LangChain's f-string prompt-template validation was incomplete in two respects. First, some prompt template classes accepted f-string templates and formatted them without enforc…

MediumCVE-2026-344471 Apr 2026

Open Neural Network Exchange (ONNX) is an open standard for machine learning interoperability. Prior to version 1.21.0, there is a symlink traversal vulnerability in external data loading allows reading files outside the model directory. This issue has been patched in version 1.2…

MediumCVE-2026-344461 Apr 2026

Open Neural Network Exchange (ONNX) is an open standard for machine learning interoperability. Prior to version 1.21.0, there is an issue in onnx.load, the code checks for symlinks to prevent path traversal, but completely misses hardlinks because a hardlink looks exactly like a …

HighCVE-2026-344451 Apr 2026

Open Neural Network Exchange (ONNX) is an open standard for machine learning interoperability. Prior to version 1.21.0, the ExternalDataInfo class in ONNX was using Python’s setattr() function to load metadata (like file paths or data lengths) directly from an ONNX model file. It…

CriticalCVE-2026-341591 Apr 2026

llama.cpp is an inference of several LLM models in C/C++. Prior to version b8492, the RPC backend's deserialize_tensor() skips all bounds validation when a tensor's buffer field is 0. An unauthenticated attacker can read and write arbitrary process memory via crafted GRAPH_COMPUT…

HighCVE-2026-274891 Apr 2026

Open Neural Network Exchange (ONNX) is an open standard for machine learning interoperability. Prior to version 1.21.0, a path traversal vulnerability via symlink allows to read arbitrary files outside model or user-provided directory. This issue has been patched in version 1.21.…

HighCVE-2026-3407031 Mar 2026

LangChain is a framework for building agents and LLM-powered applications. Prior to version 1.2.22, multiple functions in langchain_core.prompts.loading read files from paths embedded in deserialized config dicts without validating against directory traversal or absolute path inj…

CriticalCVE-2025-1537930 Mar 2026

A command injection vulnerability exists in MLflow's model serving container initialization code, specifically in the `_install_model_dependencies_to_env()` function. When deploying a model with `env_manager=LOCAL`, MLflow reads dependency specifications from the model artifact's…

CriticalCVE-2025-1503630 Mar 2026

A path traversal vulnerability exists in the `extract_archive_to_dir` function within the `mlflow/pyfunc/dbconnect_artifact_cache.py` file of the mlflow/mlflow repository. This vulnerability, present in versions before v3.7.0, arises due to the lack of validation of tar member pa…

HighCVE-2025-1538127 Mar 2026

In the latest version of mlflow/mlflow, when the `basic-auth` app is enabled, tracing and assessment endpoints are not protected by permission validators. This allows any authenticated user, including those with `NO_PERMISSIONS` on the experiment, to read trace information and cr…

HighCVE-2025-1280526 Mar 2026

A flaw was found in Red Hat OpenShift AI (RHOAI) llama-stack-operator. This vulnerability allows unauthorized access to Llama Stack services deployed in other namespaces via direct network requests, because no NetworkPolicy restricts access to the llama-stack service endpoint. As…

HighCVE-2026-3329824 Mar 2026

llama.cpp is an inference of several LLM models in C/C++. Prior to b7824, an integer overflow vulnerability in the `ggml_nbytes` function allows an attacker to bypass memory validation by crafting a GGUF file with specific tensor dimensions. This causes `ggml_nbytes` to return a …

CriticalCVE-2025-1503118 Mar 2026

A vulnerability in MLflow's pyfunc extraction process allows for arbitrary file writes due to improper handling of tar archive entries. Specifically, the use of `tarfile.extractall` without path validation enables crafted tar.gz files containing `..` or absolute paths to escape t…

HighCVE-2026-3189118 Mar 2026

Cockpit is a headless content management system. Any Cockpit CMS instance running version 2.13.4 or earlier with API access enabled is potentially affected by a a SQL Injection vulnerability in the MongoLite Aggregation Optimizer. Any deployment where the `/api/content/aggregate/…

HighCVE-2026-2850018 Mar 2026

Open Neural Network Exchange (ONNX) is an open standard for machine learning interoperability. In versions up to and including 1.20.1, a security control bypass exists in onnx.hub.load() due to improper logic in the repository trust verification mechanism. While the function is d…

HighCVE-2025-1428716 Mar 2026

A command injection vulnerability exists in mlflow/mlflow versions before v3.7.0, specifically in the `mlflow/sagemaker/__init__.py` file at lines 161-167. The vulnerability arises from the direct interpolation of user-supplied container image names into shell commands without pr…

Reviewing vulnerabilities against the PSF

Each CVE should be assessed against relevant PSF domains. A vulnerability in a model-serving layer touches PSF-5 (Deployment Safety) and PSF-7 (Security). A prompt injection issue maps to PSF-1 (Input Governance). Use the framework checklist as your assessment guide.

PSF Framework →Checklist →PSF-7 Security domain →