🚨 CVE-2026-73490
Loofah is a general library for manipulating and transforming HTML/XML documents and fragments, built on top of Nokogiri. Prior to 2.25.2, Loofah's HTML5 sanitizer applies its local-reference restriction only to the xlink:href attribute on SVG use and feImage elements, while browsers also accept the plain href attribute. A crafted sanitized SVG can therefore reference an arbitrary same-origin external document; use may render external SVG content containing scripts or other dangerous content, and feImage may load external images for tracking. Applications that sanitize user-supplied SVG with Loofah's default allowlist are affected. This issue is fixed in version 2.25.2.

🎖@cveNotify

🚨 CVE-2026-73491
Loofah is a general library for manipulating and transforming HTML/XML documents and fragments, built on top of Nokogiri. From 2.25.0 until 2.25.2, Loofah::HTML5::Scrub.allowed_uri? does not reject javascript: URIs whose scheme is split or prefixed with the HTML5 named whitespace character references 	 or 
. CGI.unescapeHTML leaves those references intact, so allowed_uri? reports the URL safe even though a browser decodes and strips the tab or line feed and executes the resulting javascript: URL. This issue affects only callers that pass HTML-encoded strings directly to allowed_uri?; Loofah's default sanitize() path is not affected. This issue is fixed in version 2.25.2.

🎖@cveNotify

🚨 CVE-2026-73492
Loofah is a general library for manipulating and transforming HTML/XML documents and fragments, built on top of Nokogiri. From 2.25.0 until 2.25.2, Loofah::HTML5::Scrub.allowed_uri? does not reject javascript: or vbscript: URIs whose scheme is split by semicolon-less numeric character references such as &#58, &#9, &#10, or &#13. CGI.unescapeHTML leaves these references encoded, so allowed_uri? reports the URL safe even though a browser decodes an encoded colon or strips encoded whitespace and executes the resulting URI scheme. This issue affects only callers that pass HTML-encoded strings directly to allowed_uri?; Loofah's default sanitize() path is not affected. This issue is fixed in version 2.25.2.

🎖@cveNotify

🚨 CVE-2026-73495
blaze is a Scala library for building asynchronous pipelines, with a focus on network IO. Prior to 0.23.18 and 1.0.0-M42, blaze-server can merge HTTP/1.1 chunked-body trailer fields into Request.headers. Because trailer fields are attacker-controlled, an unauthenticated remote client can inject arbitrary header names and values, including X-Forwarded-For and internal authorization headers, that a fronting proxy sanitized from the request-header section, bypassing header-based trust decisions in the application. Any http4s application using BlazeServerBuilder over HTTP/1.1 whose routes or middleware trust proxy-set headers, including X-Forwarded-For, X-Real-IP, and X-Forwarded-Host, is affected. If a fronting proxy strips or normalizes those headers but forwards chunked bodies with trailers intact, an attacker can spoof client IP for allow-lists, rate limits, or auditing, forge the https scheme, or inject internal authorization headers. A promoted Connection: close trailer is also honored, allowing attacker-controlled termination of pooled backend connections. This issue is fixed in versions 0.23.18 and 1.0.0-M42.

🎖@cveNotify

🚨 CVE-2026-46382
The Meeting Room Booking System (MRBS) is a PHP-based application for booking meeting rooms. Prior to version 1.12.2, a user-supplied private/local URI can be made to be fetched without checks. Version 1.12.2 contains a fix. No known workarounds are available.

🎖@cveNotify

🚨 CVE-2026-46688
The Meeting Room Booking System (MRBS) is a PHP-based application for booking meeting rooms. Prior to version 1.12.2, an unauthenticated request can be made to redirect the user to a query-specified location. This allows an attacker to create a specially-crafted URL to an MRBS installation that will cause the user who clicks it to be redirected to the attacker-specified redirect URL, which could be a spoofed MRBS login page, for example. Version 1.12.2 contains a fix. No known workarounds are available.

🎖@cveNotify

🚨 CVE-2026-48791
sigstore-java is a sigstore java client for interacting with sigstore infrastructure. Version 2.0.0 erroneously removed verification of the integrated (Rekor entry) time) against the Fulcio certificate. Version 2.1.0 re-added this verification with enhancements that adhere to the Sigstore verification spec. The old sigstore-conformance test for this check was built incorrectly. This vulnerability impacts only users verifying bundles with `dev.sigstore:sigstore-java:2.0.0`. Older versions are not affected; it is fixed in `dev.sigstore:sigstore-java:2.1.0` A malicious actor may exploit this if they were able to access a users system and exfiltrate the temporary private key used during signing and then reuse an old fulcio certificate later without requiring direct access to the user's credentials. Users may protect themselves by re-verifying their artifacts using the newest sigstore-java or another current sigstore client. Transparency logs may also be audited for unauthorized signatures for a suspected reused identity.

🎖@cveNotify

🚨 CVE-2026-49473
@cedar-policy/authorization-for-expressjs is an open-source Express.js middleware that integrates Cedar authorization into Express applications by mapping HTTP requests to Cedar actions and evaluating authorization policies before allowing requests to proceed. Versions prior to 0.3.0 have an issue where, under certain circumstances, the middleware matches incoming requests against Cedar action mappings using req.originalUrl, which includes the query string, while Express routes requests using only the path component. The middleware uses req.originalUrl to match incoming requests against Cedar action mappings. In Express, req.originalUrl includes the query string, while route matching uses only the path. This creates a divergence between what Cedar authorizes and what Express executes. When an application defines separate actions for overlapping path prefixes with different authorization requirements (for example, GET /users for listing all users with admin-only access, and GET /users/{id} for retrieving a single user with any authenticated user access), an actor can append a query string to bypass the more restrictive policy. Sending GET /users/?x=1 causes the middleware to match against /users/{id} (with id parameter set to ?x=1) and evaluate the less restrictive action, while Express routes the request to the /users list handler. This allows inappropriate access to the more restrictive endpoint. This issue has been addressed in version 0.30. Some workarounds are available. Validate and sanitize incoming request paths before they reach the authorization middleware. Ensure that applications do not rely solely on the middleware for authorization when defining multiple actions on overlapping path prefixes with different permission levels.

🎖@cveNotify

🚨 CVE-2026-49819
UpSnap is a wake on lan web app. Versions 4.4.1 through 5.3.5 are vulnerable to a missing-authentication / privilege-escalation chain in `pb.HandlerInitSuperuser` (`backend/pb/handlers.go:249`), reachable as `POST /api/upsnap/init-superuser`. The vulnerable code lacks any authentication, setup token, IP allow-list, or rate limit and is gated only by a `totalSuperusers > 0` count check — a condition that is false on every fresh install — allowing an unauthenticated network-adjacent attacker to register the initial superuser account, receive a long-lived JWT, and pivot to root remote code execution at `backend/networking/wake.go:43` (`exec.CommandContext(ctx, "/bin/sh", "-c", wake_cmd)`). Version 5.4.0 fixes the issue.

🎖@cveNotify

🚨 CVE-2026-50544
NortheBridge/luminalshine is a Sunshine-compatible game stream host for Moonlight. Prior to version 26.05.0-rc4, a latent gap exists on a default install, the file at `src/platform/windows/misc.cpp` lives at `C:\ProgramData\LuminalShine\config\apps.json` and is created by the `SYSTEM` service. Under Windows' default `C:\ProgramData` inheritance, that gives `BUILTIN\Users` only Read+Execute — not writable — so the canonical EoP doesn't actually trigger on a vanilla install. Version 26.05.0-rc4 contains a patch for the issue. As a workaround, use default condition DACLs for `ProgramData`.

🎖@cveNotify

🚨 CVE-2026-49478
Fulcio is a certificate authority for issuing code signing certificates for an OpenID Connect (OIDC) identity. Versions through 1.8.5 improperly follow cross-host redirects and attach Kubernetes ServiceAccount tokens during OIDC discovery, allowing a malicious or compromised issuer to perform blind SSRF, substitute and cache malicious JWKS keys, or disclose ServiceAccount tokens to external hosts. Version 1.8.6 blocks cross-host redirects, restricts token injection, and restricts local token loading. No known workarounds are available.

🎖@cveNotify

🚨 CVE-2026-73505
Oh My Posh is the most customisable and low-latency cross platform/shell prompt renderer. Prior to 29.35.1, the setStyle() function in src/segments/path.go passed pt.Path, which includes raw folder names, to template.Render, whose function map exposes cmd, so an attacker-controlled directory name containing a Go template expression could execute arbitrary operating system commands as the current user whenever the prompt rendered inside that directory or a descendant. This issue is fixed in version 29.35.1.

🎖@cveNotify

🚨 CVE-2026-73506
Oh My Posh is the most customisable and low-latency cross platform/shell prompt renderer. Prior to 29.35.1, write(s rune) in src/terminal/writer.go emitted attacker-controlled current directory names and Git metadata, including Commit.Subject, Commit.Author.Name, Commit.Author.Email, and RawUpstreamURL, without removing C0/C1 terminal control characters such as ESC, BEL, CSI, and OSC, allowing terminal escape sequence injection during prompt rendering that could overwrite the clipboard, spoof the prompt or screen, manipulate the window title, or disrupt the terminal. This issue is fixed in version 29.35.1.

🎖@cveNotify

🚨 CVE-2026-73507
Netty is an asynchronous, event-driven network application framework. Prior to 4.1.136.Final and 4.2.16.Final, io.netty.handler.codec.xml.XmlFrameDecoder.decode() failed to preserve closing-tag parser state across invocations, so an unauthenticated remote attacker could trickle-feed repeated </ sequences that repeatedly rescanned the accumulated buffer and exhausted an EventLoop thread's CPU, causing denial of service with a maxFrameLength of 1 MB. This issue is fixed in versions 4.1.136.Final and 4.2.16.Final.

🎖@cveNotify

🚨 CVE-2026-73508
Netty is an asynchronous, event-driven network application framework. Prior to 4.1.136.Final and 4.2.16.Final, io.netty.handler.codec.dns.AbstractDnsRecord, io.netty.handler.codec.dns.DefaultDnsRecordDecoder.decodeRecord(), and io.netty.handler.codec.dns.DnsCodecUtil.decompressDomainName() failed to release retained or newly allocated ByteBuf objects when IDN.toASCII() or encodeDomainName() rejected a malformed domain name, allowing unauthenticated remote DNS packets to leak direct memory incrementally until denial of service. This issue is fixed in versions 4.1.136.Final and 4.2.16.Final.

🎖@cveNotify

🚨 CVE-2026-73555
vLLM is an inference and serving engine for large language models. Prior to 0.26.0, the validation_exception_handler in vllm/entrypoints/openai/server_utils.py converts FastAPI RequestValidationError objects with str(exc), and sanitize_message in vllm/entrypoints/utils.py does not remove traceback-style file paths, allowing unauthenticated malformed JSON requests to /v1/chat/completions, /v1/completions, /tokenize, and /detokenize to disclose the OS username, home and virtual-environment paths, Python version, internal package structure, line numbers, and endpoint handler names. This issue is fixed in version 0.26.0.

🎖@cveNotify

🚨 CVE-2026-73556
vLLM is an inference and serving engine for large language models. Prior to 0.26.0, the structured_outputs.regex parameter in vllm/v1/structured_output/backend_lm_format_enforcer.py is passed to lmformatenforcer.RegexParser without compile_regex_with_timeout or validation in validate_structured_output_request_lm_format_enforcer, allowing an unauthenticated /v1/completions request against the lm-format-enforcer backend to consume a CPU core and stall the structured-output engine path with a catastrophic regular expression. This issue is fixed in version 0.26.0.

🎖@cveNotify

🚨 CVE-2026-73557
vLLM is an inference and serving engine for large language models. From 0.20.2rc0 until 0.26.0, safe_load_prompt_embeds in vllm/renderers/embed_utils.py uses torch.sparse.check_sparse_tensor_invariants, whose process-global save, enable, and restore state can be raced by concurrent prompt_embeds parts submitted to POST /v1/chat/completions through AsyncMultiModalItemTracker.resolve_items, asyncio.gather, and the default executor, allowing an invalid sparse tensor to reach tensor.to_dense despite the CVE-2025-62164 guard when enable_prompt_embeds is enabled. This issue is fixed in version 0.26.0.

🎖@cveNotify

🚨 CVE-2026-73558
vLLM is an inference and serving engine for large language models. Prior to 0.27.0, an integer overflow in blockIdx.x * 2 * d in activation_kernels.cu can cause act_and_mul_kernel to consume another batched user's input, allowing a request processed in the same inference batch to receive a partial or complete copy of another user's inference result. This issue is fixed in version 0.27.0.

🎖@cveNotify

🚨 CVE-2026-73559
vLLM is an inference and serving engine for large language models. From 0.19.0 until 0.26.0, the /v1/completions CompletionRequest.prompt field in vllm/entrypoints/openai/completion/protocol.py accepts an unbounded list[str] or list[list[int]], prompt_to_seq() in vllm/renderers/inputs/preprocess.py and OnlineRenderer.preprocess_completion() in vllm/renderers/online_renderer.py expand every element, and vllm/entrypoints/openai/completion/serving.py creates one engine generator and response slot per prompt, allowing an authenticated API client to exhaust CPU, memory, async scheduling capacity, engine request slots, and response buffering with one request. This issue is fixed in version 0.26.0.

🎖@cveNotify

🚨 CVE-2026-73653
Vitest is a testing framework powered by Vite. Prior to versions 3.2.7, 4.1.10, and 5.0.0-beta.6, Browser Mode provider commands including upload, takeScreenshot, screenshotMatcher, stopChunkTrace, deleteTracing, and annotateTraces accept browser-supplied file paths without enforcing the allowWrite permission gate or confining paths to the project root. A client that can reach the Browser Mode API can read arbitrary local files, create or overwrite image and trace files, or delete files accessible to the Vitest process even when allowWrite is false. This issue is fixed in versions 3.2.7, 4.1.10, and 5.0.0-beta.6.

🎖@cveNotify