Edgeless Systems Contrast versions 1.21.0 up to but not including 1.25.0 have a flaw in the imagepuller component where wildcard registry configurations may be incorrectly preferred over exact domain matches. This can cause credentials or metadata intended for wildcard domains to be applied to specific registries, potentially leaking sensitive information. The vulnerability has a low severity score and requires local attacker privileges with high attack complexity to exploit.
phpMyFAQ versions before 4.1.10 have an object-level authorization bypass vulnerability. This flaw allows unauthenticated attackers to modify or delete visitor questions by providing arbitrary openQuestionID values to the POST /api/faq/create endpoint. Attackers who solve the CAPTCHA can rewrite question categories, forge answered entries in the faqquestion_history, or delete questions permanently if the records.enableDeleteQuestion setting is enabled.
phpMyFAQ versions before 4.1.10 have a server-side request forgery (SSRF) vulnerability in the PDF export feature. This flaw allows attackers who can control FAQ content to embed external image URLs that cause the server to make unintended requests, including to internal or loopback services. The server then includes the responses from these requests in generated PDFs.
vLLM versions from 0.20.1 up to but not including 0.29.0 have an improper input validation vulnerability. This flaw allows remote attackers to crash the engine by providing out-of-vocabulary stop_token_ids in OpenAI chat, completion, generate, or batch requests. The vulnerability can cause engine termination or memory corruption by writing outside the V2 runner's logits buffer.
vLLM versions from 0.8.5 up to but not including 0.31.0 contain a denial of service vulnerability due to the use of uninitialized resources in the logprobsTensors.empty_cpu() function and the V1 model runner. This flaw can cause the EngineCore to crash and drop all co-tenant requests when a client sends specially crafted prompt_logprobs or echo with logprobs against a cached prefix.
vLLM versions from 0.18.0 up to but not including 0.31.0 have a resource exhaustion vulnerability. This flaw allows API clients to bypass the frame-count limit for video/jpeg processing by manipulating the num_frames parameter. By setting num_frames to -1 with a large inline video, the VideoMediaIO.load_base64 function decodes every frame, leading to memory exhaustion and crashing the API frontend.
vLLM versions from 0.20.0 up to but not including 0.30.0 contain a denial of service vulnerability. This flaw allows remote attackers to crash the inference engine by sending malformed multimodal metadata to the /inference/v1/generate endpoint. The issue arises from improper input validation, specifically when replaying valid multimodal items with repeated or overlapping placeholders or mismatched data lists, leading to a negative text span in Qwen2.5-VL position construction that terminates the engine.
vLLM versions from 0.10.2 up to but not including 0.25.0 have an uncontrolled resource consumption vulnerability. This flaw allows unauthenticated attackers to cause memory exhaustion by submitting specially crafted decompression-bomb images to the /v1/chat/completions endpoint. The vulnerability arises because the Nemotron-VL processors globally set Image.MAX_IMAGE_PIXELS to None, disabling limits on image size and enabling denial of service via out-of-memory conditions.
vLLM versions from 0.9.2 up to but not including 0.24.0 contain a server-side request forgery (SSRF) vulnerability. This flaw allows API callers to embed a decoded address in the X-Request-Id header, forcing the server to make outbound connections to attacker-controlled hosts. Exploitation can lead to leakage of NCCL unique_id and KV-cache tensors or cause denial of service by blocking worker processes.
vLLM versions from 0.19.0 up to but not including 0.24.0 contain a decompression bomb vulnerability in the AudioMediaIO multimodal path. Authenticated attackers can send small compressed audio inputs that decompress to very large PCM data, exhausting server memory and potentially causing denial of service.
vLLM versions from 0.22.0 up to but not including 0.24.0 have an improper input validation vulnerability in the Rust gRPC Generate and GenerateStream services. This flaw allows attackers to submit out-of-vocabulary prompt token IDs, causing a fatal embedding lookup failure that terminates the EngineCore process and leaves the server in an unhealthy state.
vLLM versions from 0.22.0 up to but not including 0.24.0 contain a denial of service vulnerability. Remote attackers can exploit this by sending specially crafted Generate or GenerateStream requests with output_logprobs enabled and more than 128 output_candidates.token_ids. This triggers a fatal sampler exception in the Rust gRPC frontend, causing the EngineCore to crash and resulting in denial of service.
vLLM versions from 0.22.0 up to but not including 0.24.0 have an improper input validation vulnerability in the Rust gRPC converter. This flaw allows remote attackers to bypass max_logprobs limits by sending oversized or integer-wrapping top_n values in CandidateTokens.top_n. Exploiting this can cause torch.topk failures that terminate the EngineCore process.
SiYuan 3.5.4 before 3.8.7 on Windows contains a sensitive-path denylist bypass in IsSensitivePath because the Windows extended-length prefix (\\?\) survives filepath.Clean and matches no forbidden prefix. Authenticated API callers can prepend the prefix to localPath in /api/import/importStdMd to copy workspace conf files, including the API token and TLS private keys, into assets served to unauthenticated clients.
SiYuan before v3.8.7 contains an sql injection vulnerability that allows admin-token holders to read history.db by injecting unescaped op and type=3 query values. Attackers can send crafted UNION SELECT payloads to searchHistory or getHistoryItems endpoints to exfiltrate deleted document content, titles, and paths.
SiYuan from 2.10.0 before 3.8.7 contains a stored cross-site scripting vulnerability that allows users with edit permission to inject script by placing javascript: URLs in math block content rendered with KaTeX trust enabled. Attackers can store \href, \url, or \includegraphics payloads that mathRender assigns via innerHTML, executing script in the application origin when a reader clicks the rendered link.
SiYuan before 3.8.7 contains a path traversal vulnerability in the MCP asset tool because validateAssetUploadPaths relies on an incomplete sensitive-path blacklist that omits .env and other secret files. Authenticated attackers driving the MCP asset upload action can copy such local files into the workspace and read them back via /assets/ over HTTP.
SiYuan versions from 2.9.0 up to but not including 3.8.7 have a cross-site scripting (XSS) vulnerability in the unauthenticated /check-auth page. This vulnerability arises because the to parameter is assigned to window.location.href without validating the URI scheme. An attacker can craft a link with a javascript: URI that executes after the victim logs in, potentially allowing access to kernel APIs to read or modify workspace content and settings.
SiYuan versions 3.6.0 up to but not including 3.8.7 have a vulnerability where the networkServeTLS setting is accepted but not enforced. This causes the kernel port to serve the full API over plain HTTP alongside HTTPS, allowing network-adjacent attackers to sniff cleartext login credentials and session cookies. The session cookie lacks the Secure attribute and responses omit HSTS, increasing the risk of session hijacking.
Parse Server versions before 8.6.101 and versions from 9.0.0 up to but excluding 9.10.4-alpha.4 contain a missing authorization vulnerability affecting MongoDB deployments. This flaw allows unauthenticated attackers to send REST API delete requests to the internal _SCHEMA collection, enabling deletion of stored schema elements such as field definitions, class-level permissions, and protected fields. This unauthorized deletion can expose data to unauthorized reads and modifications.