AI Security Dashboard
Curated vulnerabilities, advisories, and breaches affecting AI/ML systems.
Get the weekly digest
Top AI security stories every Monday. Free, no spam. Want it daily? See Daily Briefing.
CVE-2026-77596: Kubeflow Pipelines enables users to build and deploy portable, scalable machine learning workflows. Prior to 2.18.0, the
Kubeflow Pipelines enables users to build and deploy portable, scalable machine learning workflows. Prior to 2.18.0, the multi-user HTTP endpoints POST /apis/v2beta1/pipelines/upload and POST /apis/v2...
CVE-2026-102895: GPUStack is an open-source GPU cluster manager for AI model serving and GPU instance provisioning. Prior to 2.2.0, get_c
GPUStack is an open-source GPU cluster manager for AI model serving and GPU instance provisioning. Prior to 2.2.0, get_current_user in gpustack/api/auth.py treated a loopback request source as the bui...
CVE-2026-102894: GPUStack is an open-source GPU cluster manager for AI model serving and GPU instance provisioning. From 0.7.1 until 2.2.
GPUStack is an open-source GPU cluster manager for AI model serving and GPU instance provisioning. From 0.7.1 until 2.2.2, the /auth/saml/callback endpoint decoded SAML XML, used its NameID and attrib...
CVE-2026-120338: Prefab UI's Generative UI sandbox runs caller-supplied Python inside Pyodide, Python compiled to WebAssembly, in a Deno
Prefab UI's Generative UI sandbox runs caller-supplied Python inside Pyodide, Python compiled to WebAssembly, in a Deno subprocess. The feature is documented as isolation for untrusted code, including...
CVE-2026-63377: TAKO AI Agent for Okta is a local-first AI agent that helps Okta administrators query tenant data using natural language
TAKO AI Agent for Okta is a local-first AI agent that helps Okta administrators query tenant data using natural language. Prior to 3.0.1-beta, `POST /api/auth/setup` in `src/api/routers/auth.py` allow...
CVE-2026-62863: MoLing is a computer-use and browser-use based MCP server. Prior to 0.5.0, when MoLing runs in SSE mode using `-l`, its
MoLing is a computer-use and browser-use based MCP server. Prior to 0.5.0, when MoLing runs in SSE mode using `-l`, its `/sse` and `/message` HTTP endpoints do not authenticate requests and the `mcp-g...
CVE-2026-119978: Flowise through 3.1.4 contains a sandbox escape vulnerability that allows authenticated users to execute OS commands by
Flowise through 3.1.4 contains a sandbox escape vulnerability that allows authenticated users to execute OS commands by abusing allowlisted langchain and mammoth modules when ALLOW_BUILTIN_DEP=true. A...
CVE-2026-61738: SOCFortress CoPilot focuses on providing a single pane of glass for all your security operations needs. Prior to 0.1.77,
SOCFortress CoPilot focuses on providing a single pane of glass for all your security operations needs. Prior to 0.1.77, `POST /api/scoutsuite/generate-gcp-report` in `backend/app/integrations/scoutsu...
CVE-2026-61730: SOCFortress CoPilot focuses on providing a single pane of glass for all your security operations needs. In affected depl
SOCFortress CoPilot focuses on providing a single pane of glass for all your security operations needs. In affected deployments, the unauthenticated `/scoutsuite-report` `StaticFiles` mount in `backen...
CVE-2026-61729: SOCFortress CoPilot focuses on providing a single pane of glass for all your security operations needs. Prior to 0.1.76,
SOCFortress CoPilot focuses on providing a single pane of glass for all your security operations needs. Prior to 0.1.76, `verify_graylog_header` in `backend/app/active_response/routes/graylog.py` and...
CVE-2026-61728: SOCFortress CoPilot focuses on providing a single pane of glass for all your security operations needs. In affected snap
SOCFortress CoPilot focuses on providing a single pane of glass for all your security operations needs. In affected snapshots preceding the respective 0.1.70 and 0.1.76 sink repairs, portal users, inc...
CVE-2026-61707: SOCFortress CoPilot focuses on providing a single pane of glass for all your security operations needs. Prior to 0.1.76,
SOCFortress CoPilot focuses on providing a single pane of glass for all your security operations needs. Prior to 0.1.76, in the reported deployment, case-report templates uploaded by `customer_user`,...
CVE-2026-93680: IBM Langflow OSS 1.0.0 through 1.12.8 could allow a remote attacker to bypass security restrictions due to improper auth
IBM Langflow OSS 1.0.0 through 1.12.8 could allow a remote attacker to bypass security restrictions due to improper authentication.
CVE-2026-110186: IBM Langflow OSS 1.0.0 through 1.12.8 could allow a remote attacker to execute arbitrary code due to improper control of
IBM Langflow OSS 1.0.0 through 1.12.8 could allow a remote attacker to execute arbitrary code due to improper control of code generation.
CVE-2026-107329: IBM Langflow OSS 1.0.0 through 1.12.4 could allow a remote authenticated attacker to execute arbitrary code due to impro
IBM Langflow OSS 1.0.0 through 1.12.4 could allow a remote authenticated attacker to execute arbitrary code due to improper control of code generation.
CVE-2026-107328: IBM Langflow OSS 1.0.0 through 1.12.4 could allow a remote authenticated attacker to execute arbitrary code due to impro
IBM Langflow OSS 1.0.0 through 1.12.4 could allow a remote authenticated attacker to execute arbitrary code due to improper neutralization of code in eval/exec.
CVE-2026-107327: IBM Langflow OSS 1.0.0 through 1.12.4 could allow a remote authenticated attacker to execute arbitrary code due to missi
IBM Langflow OSS 1.0.0 through 1.12.4 could allow a remote authenticated attacker to execute arbitrary code due to missing authorization.
CVE-2026-104339: IBM Langflow OSS 1.0.0 through 1.12.3 could allow a remote attacker to execute arbitrary code due to improper neutraliza
IBM Langflow OSS 1.0.0 through 1.12.3 could allow a remote attacker to execute arbitrary code due to improper neutralization of special elements used in an OS command.
CVE-2026-119050: Agno through 3.1.11 contains an unsafe reflection vulnerability in PandasTools that allows attackers to invoke arbitrary
Agno through 3.1.11 contains an unsafe reflection vulnerability in PandasTools that allows attackers to invoke arbitrary pandas functions by supplying model-controlled function names to getattr(). Att...
zip: Critical Parser Differential & Security Control Bypass via Dual-Header Inconsistencies (Zip Cloaking) Leading to Arbitrary File Overwrite & RCE
## Summary A critical architectural parser differential exists in the `zip` crate (` Arbitrary File Overwrite -> RCE & DoS) against `zip2`: ```rust use std::fs; use std::io::{Cursor, Write}; use temp...
CVE-2026-117574: A flaw was found in ai-gateway-payload-processing. An authenticated remote attacker can exploit this vulnerability by su
A flaw was found in ai-gateway-payload-processing. An authenticated remote attacker can exploit this vulnerability by submitting a request specifying another tenant's model identifier in the request b...
CVE-2026-118375: llama.cpp b8841 before b11865 contains an out-of-bounds write vulnerability in the ggml-rpc RDMA transport rdma_recv() f
llama.cpp b8841 before b11865 contains an out-of-bounds write vulnerability in the ggml-rpc RDMA transport rdma_recv() function that allows unauthenticated peers to overflow the rpc-server stack. Atta...
CVE-2026-116692: MLflow is an open source AI engineering platform for agents, large language models, and machine learning models. From 3.
MLflow is an open source AI engineering platform for agents, large language models, and machine learning models. From 3.12.0rc0 until 3.17.0, third-party scorer deserialization in mlflow/genai/scorers...
CVE-2026-115727: AgenticSeek before commit 409840c contains a remote code execution vulnerability because llm_server/app.py hardcodes Fla
AgenticSeek before commit 409840c contains a remote code execution vulnerability because llm_server/app.py hardcodes Flask debug mode while binding to all interfaces. Attackers who reach the server po...
CVE-2026-107685: Improper Certificate Validation vulnerability in the WebSocket streaming client in agentjido req_llm allows an on-path a
Improper Certificate Validation vulnerability in the WebSocket streaming client in agentjido req_llm allows an on-path attacker to read and modify OpenAI WebSocket sessions, including the authorizatio...