Search vulnerabilities

Results 34,404

CVE-2026-68771

ComfyUI 0.23.0 Unauthenticated RCE via LoadTrainingDataset Pickle Deserialization

ComfyUI v0.23.0 contains an unsafe deserialization vulnerability in the LoadTrainingDataset node that allows unauthenticated remote attackers to execute arbitrary Python code by uploading a crafted pickle file and triggering its deserialization. Attackers can upload a malicious shard_*.pkl file via the unauthenticated POST /upload/image endpoint and then queue a workflow graph via POST /prompt referencing the uploaded file, causing torch.load to deserialize the attacker-controlled pickle payload using __reduce__ and execute arbitrary commands as the ComfyUI process user.

CVSS 4.0 · 9.3 Critical EPSS 1.1% KEV · Not listed
Vendors
Products
CVE-2026-68770

sentence-transformers Arbitrary Code Execution on Local Model Load Despite trust_remote_code=False

sentence-transformers contains a security control bypass vulnerability that allows attackers to achieve arbitrary code execution by exploiting a logic flaw in the import_module_class helper within sentence_transformers/util/misc.py, where the guard condition includes an 'or os.path.exists(model_name_or_path)' clause that satisfies the trust gate whenever the supplied path exists on the local filesystem, regardless of the trust_remote_code=False argument. Attackers who can control or influence the contents of a model directory on disk can place malicious Python files such as modeling_*.py referenced via modules.json, causing the code to execute at import time when an application loads the model with SentenceTransformer(path, trust_remote_code=False), bypassing the documented security contract and achieving code execution within the loading process.

CVSS 4.0 · 9.3 Critical EPSS < 1% KEV · Not listed
Vendors
Products
CVE-2026-68587

SiYuan before v3.7.3 Information Disclosure via getHeading*Transaction

SiYuan versions before v3.7.3 contain an information disclosure vulnerability in the getHeadingDeleteTransaction, getHeadingLevelTransaction, and getHeadingInsertTransaction endpoints that return rendered block DOM without publish-access checks. Anonymous readers or publish RoleReader tokens can supply a heading block ID to read full rendered content of publish-disabled documents that should be restricted.

CVSS 4.0 · 9.2 Critical EPSS < 1% KEV · Not listed
Vendors
Products
CVE-2026-68586

SiYuan before v3.7.3 Content Disclosure via getBacklinkDoc

SiYuan before v3.7.3 fails to apply publish-access filters to the getBacklinkDoc and getBackmentionDoc content endpoints (/api/ref/getBacklinkDoc and /api/ref/getBackmentionDoc). While the corresponding backlink list endpoints filter publish-forbidden documents, the content endpoints (gated only by CheckAuth) do not. A publish-mode reader — including an anonymous reader when publish Basic Auth is disabled — can call these endpoints directly with a publish-forbidden document's ID to retrieve its rendered DOM content and to determine whether the document references a given block (a reference-existence oracle).

CVSS 4.0 · 9.2 Critical EPSS < 1% KEV · Not listed
Vendors
Products
CVE-2026-68584

SiYuan before v3.7.3 Authentication Bypass via Content Endpoints

SiYuan versions before v3.7.3 contain an authentication bypass vulnerability in publish mode where content-returning endpoints getHeadingChildrenDOM, getHeading*Transaction, and getBacklinkDoc perform no password check despite protecting the primary getDoc endpoint. Anonymous attackers can retrieve full content of password-protected documents by obtaining internal block IDs from reader-accessible endpoints and calling unprotected content endpoints to bypass the password gate.

CVSS 4.0 · 9.2 Critical EPSS < 1% KEV · Not listed
Vendors
Products
CVE-2026-67622

Flowise 3.1.4 IDOR in OpenAI Assistants Integration

Flowise through 3.1.4 contains an insecure direct object reference vulnerability in the OpenAI Assistants integration that allows authenticated attackers to access credentials belonging to other workspaces by supplying an arbitrary credential UUID to Assistants endpoints without workspace ownership verification. Attackers can enumerate cross-workspace assistant metadata, retrieve file and vector store listings, and upload files into victim workspaces by exploiting the missing workspace-scoped authorization check in the credential lookup logic.

CVSS 4.0 · 8.5 High EPSS < 1% KEV · Not listed
Vendors
Products
CVE-2026-67614

CyberPanel < 3.0.0 Hard-coded JWT Secret Authentication Bypass via WebTerminal

CyberPanel before 3.0.0 contains a hard-coded JWT secret vulnerability in the WebTerminal FastAPI SSH service that allows unauthenticated remote attackers to forge valid authentication tokens and obtain an interactive root shell via WebSocket on port 8888. Attackers can craft a forged JWT signed with the hardcoded secret value, specifying ssh_user=root, to authenticate to the terminal service without any valid credentials and receive a root shell.

CVSS 4.0 · 9.3 Critical EPSS < 1% KEV · Not listed
CVE-2026-67598

Emlog Pro 2.6.23 TLS Certificate Validation Disabled in ai.php

Emlog Pro through 2.6.23 contains a disabled TLS certificate validation vulnerability in include/service/ai.php that allows network-adjacent attackers to intercept outbound HTTPS requests to configured LLM providers by presenting arbitrary TLS certificates, as CURLOPT_SSL_VERIFYPEER and CURLOPT_SSL_VERIFYHOST are unconditionally disabled across sendStream(), sendImageRequest(), send(), and fetchSearchHtml() with no option to re-enable verification. Attackers can perform man-in-the-middle interception to extract Authorization Bearer API keys from every AI request and inject crafted AI responses that may be acted upon by the tool-call execution pipeline, including the query_database and update_config tool handlers.

CVSS 4.0 · 9.1 Critical EPSS < 1% KEV · Not listed
Vendors
Products
CVE-2026-67595

VaahCMS 2.0.0 - 2.3.4 Malicious JavaScript Supply Chain via security-otp.blade.php

VaahCMS versions 2.0.0 through 2.3.4 contain a malicious obfuscated JavaScript payload embedded in the Blade template responsible for rendering security OTP emails, allowing remote attackers to execute unauthorized code in any browser that renders the affected email template with JavaScript enabled. The payload establishes a WebSocket connection to a hardcoded command-and-control endpoint, installs a password-field keylogger using MutationObserver to capture dynamically added inputs, scrapes WhatsApp Web DOM content, and accepts remote commands to redirect or overwrite the rendered page.

CVSS 4.0 · 9.2 Critical EPSS < 1% KEV · Not listed
Vendors
Products
CVE-2026-67192

Xlight FTP Server < 3.9.5 Pre-Auth Stack Buffer Overflow via SSH GCM Cipher

Xlight FTP Server before 3.9.5 contains a pre-authentication stack buffer overflow vulnerability that allows unauthenticated attackers to corrupt stack memory by sending malformed SSH packets when a GCM cipher is negotiated. Attackers can craft packets with an unvalidated length field passed directly to the GCM decrypt function, overwriting the stack cookie and return address to potentially achieve remote code execution before any authentication occurs.

CVSS 4.0 · 9.2 Critical EPSS 1.0% KEV · Not listed
Vendors
CVE-2026-67191

Xlight FTP Server < 3.9.5 Pre-Auth Heap Buffer Overflow via SSH Parser

Xlight FTP Server before 3.9.5 contains a pre-authentication heap buffer overflow vulnerability that allows remote unauthenticated attackers to write past the end of a heap buffer by sending a malformed SSH client identification string. A logic error in the recv loop's termination condition uses an incorrect OR operator where an AND operator is required, enabling exploitation on any SSH or SFTP connection before authentication occurs.

CVSS 4.0 · 9.3 Critical EPSS < 1% KEV · Not listed
Vendors
CVE-2026-65688

Bold Reports Standalone Report Designer < 14.1.12 Arbitrary File Read via Font Processing

Bold Reports Standalone Report Designer before 14.1.12 contains a missing filepath validation vulnerability in its font processing feature that allows unauthenticated attackers to read arbitrary files from the server filesystem by supplying a crafted request. Attackers can exploit this path traversal weakness to disclose sensitive server files, including authentication credentials, enabling full unauthorized access to the application. The vulnerability is specific to the DataHub module, which was introduced in Bold Reports 6.3. Therefore, versions prior to 6.3 are not affected.

CVSS 4.0 · 9.3 Critical EPSS < 1% KEV · Not listed
CVE-2026-65687

Bold Reports Standalone Report Designer < 14.1.12 Arbitrary File Read via SVG Processing

Bold Reports Standalone Report Designer before 14.1.12 contains a missing filepath validation vulnerability in its SVG processing feature that allows unauthenticated attackers to read arbitrary files from the server filesystem by supplying a crafted request. Attackers can exploit this path traversal weakness to disclose sensitive server files, including authentication credentials, enabling full unauthorized access to the application. The vulnerability is specific to the DataHub module, which was introduced in Bold Reports 6.3. Therefore, versions prior to 6.3 are not affected.

CVSS 4.0 · 9.3 Critical EPSS < 1% KEV · Not listed
CVE-2026-64622

Network-AI 5.12.2 through 5.13.3 Missing Authorization via ApprovalInbox

Network-AI (npm: network-ai) versions 5.12.2 through 5.13.3 fail to apply the configured authorization check (checkAuth/secret) to the ApprovalInbox GET read routes, so even when an operator configures a secret, unauthenticated actors can access sensitive approval request details. The GET /approvals/?status=all, GET /approvals/:id, GET /approvals/stats, and GET /approvals/sse routes disclose full ApprovalEntry content including action/target shell-command strings, file paths, justifications, and risk levels. All responses also carry a hardcoded Access-Control-Allow-Origin: * header, enabling cross-origin disclosure from any website the operator visits. This is an incomplete fix for GHSA-mxjx-28vx-xjjj.

CVSS 4.0 · 9.3 Critical EPSS < 1% KEV · Not listed
CVE-2026-59702

repomix - Server-Side Request Forgery via Unvalidated Repository URLs in POST /api/pack

repomix contains a server-side request forgery vulnerability in the POST /api/pack endpoint that allows unauthenticated attackers to make arbitrary outbound requests. The endpoint fails to properly validate http://, https://, and file:// URLs before passing them to git clone, enabling attackers to access private network addresses, GCP metadata services, or local filesystem paths.

CVSS 4.0 · 9.2 Critical EPSS < 1% KEV · Not listed
Vendors
Products
CVE-2026-5857

Contiki-NG MQTT Client Out-of-Bounds Write in PUBLISH Topic Parser via Persistent State Between TCP Segments

Contiki-NG's MQTT client parse_publish_vhdr() in os/net/app-layer/mqtt/mqtt.c sets topic_len_received=1 before checking topic_len against the 64-byte limit, so an over-length topic returns early but leaves the flag set. On the next TCP segment, tcp_input() re-invokes the parser with topic_received==0, and the persisted topic_len_received==1 skips the length-reading block containing the guard, falling through directly to a memcpy() that uses the unvalidated 16-bit topic_len as the copy length. The 65-byte topic[] destination overruns into adjacent struct fields including the payload_chunk pointer, which subsequent MQTT code dereferences, giving a compromised or attacker-controlled broker an arbitrary-pointer-write primitive. Contiki-NG's MQTT implementation has no TLS support so the connection is plaintext. Impact ranges from information disclosure and denial of service to remote code execution on embedded targets without memory protection.

CVSS 4.0 · 9.2 Critical EPSS < 1% KEV · Not listed
Vendors
Products
CVE-2026-54363

CentreStack < 17.5 Hardcoded Key Token Forgery RCE

CentreStack before 17.5 contains a hardcoded cryptographic key vulnerability that allows unauthenticated attackers to forge arbitrary encrypted tokens by exploiting a static SysNumber value used as entropy for AccessTicket.Encrypt() and AccessTicket.Decrypt() across all installations. Attackers can use the hardcoded key to craft valid x-glad-auth headers and call privileged API endpoints such as acquiretenantbackuptoken to obtain a domain administrator IdentityTicket, enabling a complete unauthenticated remote code execution chain.

CVSS 4.0 · 9.3 Critical EPSS < 1% KEV · Not listed
Vendors
Products
CVE-2026-53793

rsync < 3.5.0 Path Confinement Bypass via /./ Boundary Marker in Chroot Mode

rsync before 3.5.0 contains a path confinement bypass vulnerability that allows remote clients to escape the intended inner-module root confinement by constructing paths that resolve outside the chroot boundary when the module root contains a /./ boundary marker. Attackers can exploit improper handling of the /./ notation or forge delta-basis transfers referencing xname paths that cross the /./ boundary to gain unauthorized read or write access to files outside the module's subtree.

CVSS 4.0 · 9.1 Critical EPSS < 1% KEV · Not listed
Products
CVE-2026-53790

rsync < 3.5.0 Command Injection via Multiple Code Paths

rsync before 3.5.0 contains multiple command and argument injection vulnerabilities that allow attackers to execute arbitrary commands by supplying malicious input through several code paths, including the RSYNC_CONNECT_PROG environment variable, daemon hooks, the rsync-ssl wrapper, and remote-shell command newline injection. Attackers can inject shell metacharacters or newline characters into unsanitized user-supplied values such as hostnames and hostspecs to execute arbitrary commands under the privileges of the rsync process or the invoking user.

CVSS 4.0 · 9.2 Critical EPSS < 1% KEV · Not listed
Products
CVE-2026-43632

llama.cpp b7492–b9060 Use-After-Free in Tokenization Endpoints

llama.cpp builds b7492 through the latest b9060 contains a use-after-free vulnerability in llama-server affecting six tokenization endpoints (/tokenize, /detokenize, /infill, /apply-template, /rerank, and /anthropic/count_tokens) that bypass the task queue and access ctx_server.vocab directly on HTTP worker threads. Attackers can exploit a time-of-check-time-of-use race condition where the main thread destroys and frees vocab after the synchronization lock is released but before the handler finishes using it, causing a crash or potential code execution when --sleep-idle-seconds is configured.

CVSS 4.0 · 9.2 Critical EPSS < 1% KEV · Not listed
Vendors
Products