CVE-2026-71424Onyx is an open-source AI platform. Prior to 3.1.10, 3.2.14, and 4.0.0, Onyx's GET /api/mcp/servers and GET /api/mcp/servers/persona/{persona_id} endpoints expose another user's OAuth Authorization header because OnyxTokenStorage.set_tokens and OnyxTokenStorage.set_client_info in backend/onyx/server/features/mcp/api.py copy per-user tokens into a shared admin MCPConnectionConfig row and _db_mcp_server_to_api_mcp_server returns that row through auth_template.headers to any BASIC_ACCESS user. This issue is fixed in versions 3.1.10, 3.2.14, and 4.0.0.
2026-08-17 · score 9.6
CVE-2026-64849MLflow is an open source AI engineering platform for agents, large language models, and machine learning models. Prior to 3.15.0, the unauthenticated POST /api/2.0/mlflow/webhooks/{id}/test endpoint calls _validate_webhook_url() in mlflow/utils/validation.py only for the original URL while mlflow/webhooks/delivery.py follows redirects and re-resolves the hostname without pinning the validated address, allowing attackers to reach internal or cloud metadata services and receive response_status and response_body. This issue is fixed in version 3.15.0.
2026-08-17 · score 9.3
CVE-2026-75110MemOS is a memory operating system for LLMs and AI agents. In deployments where authentication is enabled (AUTH_ENABLED=true) but the undocumented, defaultless INTERNAL_SERVICE_SECRET environment variable is unset, the is_internal_request() check in src/memos/api/middleware/auth.py fails open: os.getenv("INTERNAL_SERVICE_SECRET") returns None and a request omitting the X-Internal-Service header also yields None, so the comparison None == None evaluates true. The request is then treated as a trusted internal principal and granted scopes: ["all"]. As a result, an unauthenticated remote attacker can reach the admin API-key management endpoints to mint API keys for any user, enumerate keys, revoke keys, and generate a master key for persistent privileged access, as well as all data endpoints.
2026-08-17 · score 9.8
CVE-2026-75106OpnForm derives editable-submission secrets from sequential row identifiers using Hashids with an empty default salt, allowing unauthenticated attackers to compute hashes for any submission. Attackers can read other respondents' full submission data through the submission-fetch endpoint or overwrite submissions by supplying predicted hashes to the answer endpoint.
2026-08-17 · score 9.1
CVE-2026-66795A flaw was found in the managedcluster-import-controller. The Certificate Signing Request (CSR) auto-approval logic improperly validates incoming CSRs, specifically by not inspecting the signer name or decoding the PEM-encoded x509 CSR. This vulnerability allows a privileged service account on a spoke cluster to submit a malicious CSR. Successful exploitation can lead to privilege escalation, enabling the attacker to obtain administrative credentials on the hub cluster.
2026-08-17 · score 9.1
