| CVE |
Vendors |
Products |
Updated |
CVSS v3.1 |
| vLLM is an inference and serving engine for large language models (LLMs). Prior to 0.22.0, an assert-based security check in vLLM's activation function loading allows any unauthenticated attacker to achieve arbitrary code execution on the server by publishing a malicious HuggingFace model, when vLLM runs in Python optimized mode (python -O or PYTHONOPTIMIZE=1). This vulnerability is fixed in 0.22.0. |
| Termix is a web-based server management platform with SSH terminal, tunneling, and file editing capabilities. Prior to 2.3.2, the terminal WebSocket accepts a user-controlled hostConfig.id and src/backend/ssh/host-resolver.ts resolves that host without requiring ownership or explicit access. When no credential is shared with the requester, resolveHostById performs an owner credential fallback, and src/backend/ssh/terminal.ts combines that credential with attacker-controlled ip, port, and username values. An authenticated low-privileged user can therefore make Termix authenticate to an attacker-controlled SSH server and disclose another user's stored SSH password or private-key material while the victim user's data key is unlocked. This issue is fixed in version 2.3.2. |
| Monkeytype is a minimalistic and customizable typing test. In 26.26.0 and earlier, the backend rate-limit key generator in backend/src/middlewares/rate-limit.ts uses client-controlled cf-connecting-ip and x-forwarded-for headers before the trust-proxy-derived req.ip value. An unauthenticated attacker can rotate either header to create a new bucket for each request, bypassing rootRateLimiter, badAuthRateLimiter, getKey(), and the getKeyWithUid() fallback used by public endpoints. This permits repeated POST /users/forgotPasswordEmail and verificationEmail requests, mail bombing registered users, consuming Firebase or SMTP quota, evading brute-force protection, and enabling resource exhaustion. Exploitability of cf-connecting-ip depends on deployment topology, but x-forwarded-for and direct-to-origin paths remain affected when those values are not overwritten by a trusted proxy. No fixed version is available as of this review. |
| The report.list procedure in packages/trpc/src/routers/report.ts accepted a projectId and a dashboardId and returned getReportsByDashboardId(dashboardId). The enforceAccess middleware in packages/trpc/src/trpc.ts verified membership for the supplied projectId, but nothing verified that the supplied dashboardId belonged to that project, and getReportsByDashboardId in packages/db/src/services/reports.service.ts selects reports by dashboardId alone with no project scoping. An authenticated user could therefore pair a projectId from their own organization, which satisfies the middleware, with a dashboardId belonging to another organization and receive every report in that dashboard. A correctly scoped helper, listReportsCore, already existed in the same service file and resolves the dashboard through getDashboardById(dashboardId, projectId) before returning reports, but the router did not use it. |
| openssl_encrypt versions before 1.4.0 contain a missing ownership verification vulnerability in the revoke_key method that allows authenticated clients to revoke any other client's key. Attackers can revoke arbitrary keys by providing a valid ML-DSA signature, bypassing the intended ownership restriction. |
| openssl_encrypt versions before 1.4.0 contain a sandbox bypass vulnerability where the plugin sandbox fails to restrict alternative file access methods like pathlib.Path and io.open. Attackers can import pathlib or io modules to read and write arbitrary files, completely bypassing the restricted_open file access controls. |
| openssl_encrypt versions before 1.4.0 fail to apply sandbox restrictions in the default process isolation mode for plugin execution. Attackers can execute malicious plugins with unrestricted access to the filesystem, network, subprocess execution, and all Python modules. |
| openssl_encrypt versions before 1.4.0 contain a sandbox escape vulnerability in the DangerousPatternVisitor AST analyzer that fails to detect dunder attribute traversal techniques. Attackers can use __class__, __bases__, __subclasses__(), and __globals__ chains to access restricted functions and execute arbitrary system commands from plugin code. |
| Crawlab fails to verify user ownership or administrative role on the password-change endpoint, allowing any authenticated user to reset any account's password. Attackers can enumerate user accounts through the user listing endpoint and change administrator credentials to achieve full account takeover and arbitrary code execution. |
| MemOS 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. |
| Onyx is an open-source AI platform. Prior to 4.3.0, Onyx Enterprise Edition's PATCH /manage/admin/user-group/{user_group_id} and POST /manage/admin/user-group/{user_group_id}/add-users endpoints in ee/onyx/server/user_group/api.py call update_user_group and add_users_to_user_group in ee/onyx/db/user_group.py without enforcing _validate_curator_can_modify_group, allowing a curator to add accounts to arbitrary groups and obtain document access through get_acl_for_user and the OpenSearch access_control_list filter. This issue is fixed in version 4.3.0. |
| Authorization bypass through User-Controlled key vulnerability in Netiket Information Technologies EdoWEB allows Accessing Functionality Not Properly Constrained by ACLs.
This issue affects EdoWEB: before 780-g7. |
| BetterDesk is a remote desktop management solution. BetterDesk versions through 2.3.0 improperly invalidate deleted device identities, allowing an unauthenticated client to replay or spoof a device ID and bypass registration controls. Version 3.0.0-alpha contains a patch. No known workarounds are available. |
| Subscriber Insecure Direct Object References (IDOR) in Modal Survey <= 2.0.2.2.3 versions. |
| Dell Watchdog Timer Driver versions prior to 2.0.0.1 contain an Exposed IOCTL with Insufficient Access Control vulnerability. A low privileged attacker with local access could potentially exploit this vulnerability, leading to Privilege Escalation. |
| Kurrier is a modern, self-hosted workspace for email, calendar, contacts, and storage. Prior to 1.2.4, Kurrier API endpoints for listing and retrieving webhook and identity resources did not enforce ownership checks for authenticated API requests. An attacker with a valid API key could use another account's identifiers to read and enumerate webhook and identity resources belonging to that account through apps/worker/server/routes/api/kurrier/webhooks/[id].get.ts, apps/worker/server/routes/api/kurrier/webhooks/index.get.ts, apps/worker/server/routes/api/kurrier/identities/[id].get.ts, and apps/worker/server/routes/api/kurrier/identities/index.get.ts. Anonymous requests and invalid API keys were rejected, and cross-user modification operations were blocked, but affected GET and list operations could expose another user's resource metadata. This issue is fixed in version 1.2.4. |
| Vulnerability in the Siebel Apps - Self Service product of Oracle Siebel CRM (component: Helpdesk/Training). Supported versions that are affected are 17.0-26.6. Easily exploitable vulnerability allows unauthenticated attacker with network access via HTTP to compromise Siebel Apps - Self Service. Successful attacks require human interaction from a person other than the attacker and while the vulnerability is in Siebel Apps - Self Service, attacks may significantly impact additional products (scope change). Successful attacks of this vulnerability can result in unauthorized creation, deletion or modification access to critical data or all Siebel Apps - Self Service accessible data as well as unauthorized access to critical data or complete access to all Siebel Apps - Self Service accessible data. CVSS 3.1 Base Score 9.3 (Confidentiality and Integrity impacts). CVSS Vector: (CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:C/C:H/I:H/A:N). |
| Authorization Bypass Through User-Controlled Key in the transaction API in Roskus Prospero
Flow CRM 5.0.0 through 5.3.5 allows an authenticated user to read the transactions of other
companies on the same instance via an incremented identifier in GET /api/transaction/{id},
which is resolved without company scoping and without any permission check. |
| n8n before 1.123.69, 2.x before 2.33.4, and 2.x before 2.34.1 contain an allowed-domains bypass in the GraphQL node. When the node's Authentication parameter is set to expression mode, every authentication-gated credential selector is treated as active; if two credentials of different types are attached, the node enforces the allowed-domains policy of only the first credential while still attaching material from both. An authenticated user with workflow-authoring rights can thereby send a domain-restricted credential to an attacker-controlled endpoint, exfiltrating it with the leaked credential's permissions. |
| stigmem-node before 0.9.0a12 contains a broken object level authorization (cross-tenant BOLA) vulnerability in the quarantine review endpoints. On multi-tenant deployments running the opt-in stigmem-plugin-multi-tenant, the list/count queries and _get_quarantined_fact in routes/quarantine.py lacked a tenant_id predicate and the garden lookup was not tenant-scoped, allowing a tenant administrator with only a plain tenant write capability to list, read, and admit or reject quarantined facts belonging to other tenants via the /v1/quarantine endpoints. Default single-tenant deployments are not affected. |
