Export limit exceeded: 398183 CVEs match your query. Please refine your search to export 10,000 CVEs or fewer.
Export limit exceeded: 398183 CVEs match your query. Please refine your search to export 10,000 CVEs or fewer.
Export limit exceeded: 398183 CVEs match your query. Please refine your search to export 10,000 CVEs or fewer.
Export limit exceeded: 398183 CVEs match your query. Please refine your search to export 10,000 CVEs or fewer.
Search
Search Results (398183 CVEs found)
| CVE | Vendors | Products | Updated | CVSS v3.1 |
|---|---|---|---|---|
| CVE-2026-93592 | 1 Vllm | 1 Vllm | 2026-09-22 | 7.5 High |
| vLLM versions before 0.28.0 fail to validate the lower bound of token IDs in the /v1/embeddings and /pooling endpoints, allowing unauthenticated attackers to crash the engine by submitting negative token IDs. A single request with a negative token ID triggers a CUDA device-side assertion that poisons the GPU context, causing all subsequent requests to fail until the process restarts. | ||||
| CVE-2026-94623 | 1 Vllm | 1 Vllm | 2026-09-22 | 7.5 High |
| vLLM through 0.29.0 contains a denial of service vulnerability in the NIXL connector's prefix caching implementation that fails to properly validate block counts across multi-prompt completion requests in prefill/decode disaggregated deployments. Attackers can trigger an assertion failure in NixlBaseConnectorWorker._apply_prefix_caching by submitting completion requests with multiple prompts of varying lengths, causing the decode worker to terminate and become unavailable until restarted. | ||||
| CVE-2026-94622 | 1 Vllm | 1 Vllm | 2026-09-22 | 7.5 High |
| vLLM versions through 0.29.0 contain a denial of service vulnerability in the NIXL connector's metadata handling for prefill/decode disaggregated deployments. Attackers can send requests with incomplete kv_transfer_params dictionary entries to trigger an uncaught KeyError in EngineCore scheduling, causing the decode engine to terminate and making all routed requests fail until manual restart. | ||||
| CVE-2026-94624 | 2 Vllm, Vllm-project | 2 Vllm, Vllm | 2026-09-22 | 7.5 High |
| vLLM through 0.29.0 contains a denial of service vulnerability in P2P KV offloading when OffloadingConnector is configured with TieringOffloadingSpec and a peer-to-peer secondary tier. Attackers can supply arbitrary remote host and port values in kv_transfer_params to create unreachable peer sessions that retain ZeroMQ sockets until the context quota is exhausted, causing an uncaught ZMQError that crashes EngineCore and stops all inference. | ||||
| CVE-2026-94626 | 1 Vllm | 1 Vllm | 2026-09-22 | 7.5 High |
| vLLM through 0.29.0 fails to validate the tp_size parameter in kv_transfer_params on OpenAI-compatible completion endpoints, allowing attackers to allocate unbounded memory. Attackers can supply arbitrary tp_size values in prefill/decode disaggregated deployments to exhaust memory and trigger kernel OOM-kill of the decode worker process. | ||||
| CVE-2026-94627 | 1 Vllm | 1 Vllm | 2026-09-22 | 7.5 High |
| vLLM Mooncake connector through 0.29.0 fails to properly manage GPU KV cache block ownership when concurrent child requests share a single transfer ID in prefill/decode disaggregated deployments. Attackers can trigger GPU memory exhaustion by submitting completion requests with multiple prompts, causing orphaned KV cache blocks to accumulate until process restart and eventually preventing legitimate requests from executing. | ||||
| CVE-2026-93436 | 1 Vllm | 1 Vllm | 2026-09-22 | 7.5 High |
| vLLM through 0.29.0 fails to properly clean up decode-side metadata for rejected inference requests in prefill/decode disaggregated deployments. Remote attackers can submit requests with max_tokens=0 to exhaust decode-worker memory without bound until the worker restarts. | ||||
| CVE-2026-93600 | 1 Rustls | 1 Webpki | 2026-09-22 | 2.2 Low |
| rustls-webpki (rustls/webpki) versions 0.101.0 through 0.103.11 and 0.104.0-alpha releases before 0.104.0-alpha.6 ignore X.509 name constraints that apply to URI names, causing such constraints to be accepted rather than enforced. Because name constraints are restrictions on otherwise properly issued certificates, the flaw is only reachable after successful signature verification and requires a misissued certificate to exploit; the library also provides no API for asserting URI names, and URI name constraints are otherwise unimplemented. Versions 0.103.12 and 0.104.0-alpha.6 reject URI name constraints unconditionally. | ||||
| CVE-2026-93601 | 1 Rustls | 1 Webpki | 2026-09-22 | 2.2 Low |
| rustls-webpki (the Rust webpki fork used by rustls) versions >= 0.101.0 and prior to 0.103.12 and 0.104.0-alpha.6 incorrectly accepted permitted-subtree DNS name constraints for certificates asserting a wildcard name. For example, a name constraint of accept.example.com was treated as satisfied by a certificate for *.example.com, which could feasibly assert reject.example.com — a name outside the permitted subtree. Because name constraints are restrictions applied to otherwise properly issued certificates, the issue is only reachable after signature verification succeeds and requires a misissued wildcard certificate to exploit. | ||||
| CVE-2026-93690 | 1 Garycourt | 1 Uri-js | 2026-09-22 | 7.5 High |
| uri-js through 4.4.1 contains a denial of service vulnerability in the removeDotSegments function that loops infinitely when a path segment begins with Unicode line or paragraph separators. Attackers can trigger this by calling removeDotSegments directly or through normalize/resolve functions with IRI handling enabled, causing the Node.js event loop to block indefinitely until heap exhaustion. | ||||
| CVE-2026-93751 | 1 Garycourt | 1 Uri-js | 2026-09-22 | 6.5 Medium |
| uri-js through 4.4.1 contains an improper UTF-8 decoding vulnerability in pctDecChars() that decodes invalid and overlong percent-encoded sequences into ASCII metacharacters. Attackers can craft percent-encoded payloads to bypass platform decoder validation and inject path traversal or CRLF sequences that downstream consumers process without filtering. | ||||
| CVE-2026-93841 | 1 Vllm | 1 Vllm | 2026-09-22 | 3.7 Low |
| vLLM through 0.29.0 contains a memory corruption vulnerability in the Triton _bincount_kernel where prompt token IDs index the penalty prompt-presence bitset without bounds checking against vocabulary size. Attackers can submit multimodal audio requests with tokens equal to vocabulary size, causing out-of-bounds writes that corrupt concurrent requests' sampler state and alter repetition penalty behavior. | ||||
| CVE-2026-92599 | 2 Hapijs, Redhat | 2 Joi, Hummingbird | 2026-09-22 | 7.5 High |
| joi (npm package `joi`, hapi.js) versions >=17.2.0 <17.13.7 and >=18.0.0 <18.2.6 are vulnerable to regular expression denial of service in the `Joi.string().isoDate()` validation rule. One of the regular expressions the rule applies to the input is unanchored, so a valid ISO date followed by a long run of fractional-second digits causes the regex engine to restart its search from every position in the string, yielding time proportional to the square of the input length (about 1.4 s for 64 KB of digits and about 22 s for 256 KB). A remote attacker who can supply a string to an isoDate validation can stall the application with a single request. Fixed in 17.13.7 and 18.2.6; as a workaround, cap the length of the string before it reaches joi. | ||||
| CVE-2026-78624 | 1 Okta | 1 Access Gateway | 2026-09-22 | 4.9 Medium |
| The Okta Access Gateway backup restore function does not validate the filename embedded in an encrypted backup payload. This results in writing file contents to unintended locations on the appliance filesystem. | ||||
| CVE-2026-95624 | 1 Tauri | 1 Tauri-plugin-updater | 2026-09-22 | 6.8 Medium |
| The Tauri updater plugin's 'check' IPC command accepts an allowDowngrades boolean parameter directly from frontend JavaScript code. When set to true, it replaces the version comparator from "update must be newer" to "update must be different." Because the default permission set grants allow-check to the webview, any XSS in the app frontend can invoke this command and bypass the only anti-rollback protection the updater offers. Combined with another bug, this enables downgrade attacks without even needing to fake a higher version number. | ||||
| CVE-2026-93712 | 2026-09-22 | 7.5 High | ||
| Dancer2 versions from 2.1.0 before 2.2.0 for Perl serve files from outside public_dir via relative path segments in the File route handler. The handler joins the request path onto public_dir without collapsing relative segments, and checks only that the result is a readable regular file. A request for `/../outside.txt` escapes public_dir, and percent-encoding the dots reaches the same file. The handler is off by default. An application is affected once it names File in route_handlers and sets static_handler to 0, which otherwise refuses a dot segment before the route runs. Any file the worker process can read is served to an unauthenticated request, including the application's config.yml above public_dir. | ||||
| CVE-2026-93711 | 2026-09-22 | 6.5 Medium | ||
| Dancer2 versions before 2.2.0 for Perl do not strip CR and LF from response header names in headers_to_array. The routine removes CR and LF from each header value but not from the name. A name carrying them therefore reaches the PSGI server intact. A server that does not validate keys writes it to the wire, so the bytes after the CRLF arrive as their own header line. The application has to derive the header name from request data, not just the value. An attacker who controls that data adds their own headers and splits the response. | ||||
| CVE-2026-93710 | 2026-09-22 | 7.5 High | ||
| Dancer2 versions from 2.0.0 before 2.2.0 for Perl dispatch a route that a dying hook refused when the exception handler halts the response in compile_hooks. A hook that dies fires core.app.hook_exception, then calls cleanup unless the failing hook is the exception handler. A handler that halts does not stop that cleanup, which discards the request, response and session the dispatcher has yet to read, so the refused route runs. The handler has to halt the response object by calling its halt method or setting is_halted: the halt keyword unwinds through with_return before cleanup runs. A check in a before hook is not enforced: the caller gets the refusal, while the route body runs and its writes land. | ||||
| CVE-2026-93012 | 1 Rjbs | 1 Email::sender::transport:sendmail | 2026-09-22 | 9.8 Critical |
| Email::Sender::Transport::Sendmail versions before 2.602 for Perl allow arbitrary command execution on Windows sending a message whose envelope address reaches the shell in _sendmail_pipe. On MSWin32 the envelope sender and every recipient go into a single command string, which open() passes to a shell. Every other platform gets the list form, which runs sendmail directly. When the caller supplies no envelope, Email::Sender::Simple takes the recipients from the To and Cc headers and the sender from the From header. An attacker who controls one of those header addresses runs commands as the sending process. | ||||
| CVE-2026-77558 | 2026-09-22 | 7.5 High | ||
| A malicious actor with access to the network could exploit an Out-of-bounds Read vulnerability found in certain UniFi gateway devices to execute a Denial of Service (DoS) attack on the device. | ||||