Overview of incoming advisories.
2179 results found
CVE-2026-61686 affects SolidInvoice, an open-source invoicing platform, prior to version 3.0.1. The vulnerability exists in the DataGrid LiveComponent, which deserializes a context prop value using PHP's unsafe unserialize() function after receiving it from the client. Because the prop is marked as writable: true, an authenticated attacker can supply an arbitrary PHP serialized payload, potentially leading to remote code execution or other serious impacts. The vulnerability has been patched in version 3.0.1. Users are advised to upgrade immediately to mitigate the risk of exploitation by authenticated attackers.
View original advisory →OGX (formerly Llama Stack), at commit fbe8e0f, contains an unauthenticated Server-Side Request Forgery (SSRF) vulnerability in its OpenAI-compatible POST /v1/responses endpoint. MCP tool definitions accept a server_url parameter that is fetched server-side without any destination validation, as the existing validate_url_not_private() guard is not applied to this parameter. In the default starter configuration, which runs without authentication, a remote unauthenticated attacker can cause the server to connect to arbitrary internal addresses, including cloud metadata endpoints such as http://169.254.169.254/. The attacker can also forward attacker-supplied headers and bearer tokens to those internal destinations, enabling credential theft and internal network reconnaissance. This vulnerability poses a significant risk in cloud-hosted deployments where metadata endpoints expose sensitive credentials. A fix would require applying the existing URL validation guard to the server_url parameter as well.
View original advisory →SiYuan versions before v3.8.2 contain a denial of service vulnerability in the unauthenticated /api/system/uiproc endpoint. The endpoint accepts and retains attacker-controlled process identifiers without enforcing size limits or requiring authentication. Attackers can repeatedly send requests with unique identifiers to exhaust process memory. This leads to degraded service availability and potential full denial of service. The vulnerability requires no authentication, lowering the barrier for exploitation. A fix is available in SiYuan v3.8.2 and later. The issue is tracked as CVE-2026-85581 and has been published via NVD, GitHub Security Advisories, and VulnCheck.
View original advisory →CVE-2026-77393 affects Inductive Automation Ignition versions 8.1.53 and earlier, where the Gateway 'Create Project Role(s)' setting shipped with a blank default value. This misconfiguration allowed any authenticated user to create projects, provided they had the ability to execute gateway scripts. The vulnerability represents an improper access control issue in the project creation workflow. Ignition version 8.1.54 addresses the issue by restricting project creation exclusively to Designer sessions and removing reliance on the flawed setting. The Ignition 8.3 series is not affected by this vulnerability. CISA has issued an ICS advisory (ICSA-26-246-06) regarding this issue, indicating relevance to operational technology and industrial control system environments. A fix has been made available and users are advised to upgrade to version 8.1.54 or later.
View original advisory →A critical vulnerability in undici versions 8.10.0 through 8.10.2 omits the destination origin from cache and request-deduplication keys when the cache or deduplicate interceptor is composed directly onto a Client or Pool. The internal cache key falls back to an empty origin string, causing cached responses from one upstream origin to be returned for requests to a different trusted origin when method, path, and headers match. This enables cross-origin information disclosure and persistent cache poisoning. A demonstrated exploit showed full authentication bypass where a JWT signed with an attacker-controlled key was accepted as belonging to a trusted issuer without ever contacting the trusted origin. The vulnerability is a regression introduced in version 8.10.0 and does not affect applications using an Agent. Users are advised to upgrade to undici 8.10.3 or later to remediate the issue.
View original advisory →sift.js version 17.1.3 contains a critical vulnerability where the library enumerates query keys using for...in, which traverses the object prototype chain and dispatches matched operator keys including $where. The $where operator compiles string values into executable functions via new Function(), which is enabled by default (CSP_ENABLED is not set by default). This creates two attack vectors: first, if a prototype-pollution primitive elsewhere sets Object.prototype.$where to a malicious string, even benign sift({}) calls execute arbitrary JavaScript; second, passing an untrusted query object with a string $where directly to sift results in immediate code execution. The vulnerability effectively chains prototype pollution with arbitrary code execution in Node.js environments using this widely-used MongoDB-style query filtering library. Mitigation requires enabling CSP_ENABLED or upgrading to a patched version. The issue has been reported via GitHub and tracked by VulnCheck.
View original advisory →A security vulnerability exists in Fastify versions before 5.12.2 related to incomplete HTTP header name normalization. The framework lowercases header names in route schemas but fails to apply this transformation to trigger and dependent names inside JSON Schema Draft 7 'dependencies' keywords. Because Node.js stores request header names in lowercase, conditional header dependencies never match, causing presence assertions to be silently skipped. This allows an unauthenticated remote client to send a privilege-activating header while omitting the required authentication header, effectively bypassing conditional security checks. The vulnerability can lead to unauthorized access to privileged functionality. No authentication is required to exploit this flaw remotely. Users are advised to upgrade to Fastify 5.12.2 or later to remediate the issue.
View original advisory →A missing authorization vulnerability has been discovered in SourceCodester Class and Exam Timetabling System version 1.0. The flaw exists in the file /admin/session.php, where manipulation of the argument ID bypasses authorization checks. The vulnerability can be exploited remotely without requiring local access. A public exploit has already been released, increasing the risk of active exploitation. The affected product is a web-based academic scheduling application. The vulnerability has been assigned CVE-2026-85512 and documented on NVD and VulDB. No patch or mitigation details are currently mentioned in the article. The public availability of the exploit raises the criticality level significantly.
View original advisory →excel-mcp-server version 0.1.8 contains a path confinement enforcement failure in stdio mode when the EXCEL_FILES_PATH environment variable is unset. This vulnerability allows attackers to supply unchecked file paths to read and write tools, enabling access to any file accessible to the process. The flaw effectively bypasses intended file access restrictions, resulting in arbitrary file read and write capabilities. The issue is tracked as CVE-2026-85661 and was reported via GitHub issue #149. A fix or mitigation requires proper path validation and confinement enforcement regardless of environment variable configuration. The vulnerability affects the server.py and validation.py components of the package. It poses a significant risk in environments where the MCP server is exposed to untrusted input.
View original advisory →A missing authorization vulnerability has been identified in SourceCodester Class and Exam Timetabling System version 1.0. The flaw resides in the file /admin/session.php, where manipulation of the 'ID' argument bypasses authorization checks. The vulnerability can be exploited remotely without requiring physical access to the system. A public exploit has already been released, increasing the risk of active exploitation. The issue is tracked as CVE-2026-85512 and is listed on NVD and VulDB. Affected systems running this version of the timetabling software are at risk of unauthorized access to administrative sessions. Users are advised to apply patches or mitigations as soon as they become available.
View original advisory →A SQL injection vulnerability has been identified in code-projects Hospital Information System version 1.0. The flaw resides in the getSinglePresp function within the file includes/presp/PrespController.php, where manipulation of the 'ID' argument leads to SQL injection. The vulnerability can be exploited remotely without requiring physical access to the system. A public exploit has already been released, increasing the risk of active exploitation. The affected product is a hospital management web application, making the impact potentially critical due to sensitive patient data exposure. The vulnerability is tracked as CVE-2026-85399 and has been documented across multiple security databases including NVD and VulDB.
View original advisory →A vulnerability exists in undici's BalancedPool constructor where the options object is deep-cloned via JSON serialization, causing function-valued TLS options such as checkServerIdentity callbacks and custom connectors to be silently discarded. This results in TLS certificate validation bypass, where peers with certificates that should be rejected by custom logic are instead accepted if they pass Node's default hostname and chain checks. The flaw affects only BalancedPool; Client, Pool, and Agent dispatchers are not impacted. Affected versions include undici 7.24.1 up to 7.29.1 and 8.0.0 up to 8.10.2. The vulnerability is only triggered when function-valued connect or tls options are supplied to BalancedPool. Users are advised to upgrade to undici 7.29.1 or 8.10.2 to remediate the issue.
View original advisory →A server-side request forgery (SSRF) vulnerability exists in jina-ai/reader where URL validation is only applied to the initial request and not to subsequent HTTP redirect hops. Attackers can exploit this by crafting a publicly accessible URL that redirects to internal network addresses or cloud metadata endpoints such as AWS IMDSv1. The server fetches and returns the response body of the redirected target to the attacker, potentially exposing sensitive internal data. The vulnerable code paths are identified in curl.ts and puppeteer.ts within the reader service. This bypass allows attackers to enumerate and access internal infrastructure that would otherwise be protected by the initial URL validation logic. The issue has been reported and tracked in the project's GitHub issue tracker. An advisory has also been published by VulnCheck detailing the attack vector and impact. This vulnerability poses a high risk especially in cloud-hosted deployments where metadata endpoints are accessible.
View original advisory →TEN Framework version 0.11.71 is affected by critical unauthenticated arbitrary file read and write vulnerabilities in the TMAN Designer file-content API endpoints. Attackers can exploit the /api/designer/v1/file-content endpoints via POST and PUT HTTP requests without any authentication. This allows reading of arbitrary files from the system or writing malicious content to sensitive system paths. Exploitation can lead to remote code execution through manipulation of authorized_keys files, cron jobs, or executable graph files. No credentials or prior access are required to exploit these vulnerabilities. The vulnerabilities are documented in the TEN Framework GitHub repository and tracked under CVE-2026-85688. The issue has been reported via a GitHub issue and a VulnCheck advisory. Organizations running TEN Framework 0.11.71 should treat this as a high-priority remediation.
View original advisory →CVE-2026-85525 describes a critical flaw in Snowflake's Python, Go, JDBC, and Node.js drivers where improper OCSP response validation allowed revoked TLS certificates to be accepted as valid. The vulnerability stemmed from OCSP responses not being reliably bound to the certificate under validation, and definitive verification failures being treated as transient errors. A man-in-the-middle attacker possessing a revoked certificate and its private key for a Snowflake or stage hostname could intercept TLS sessions. This would allow the attacker to read and modify data transmitted within the intercepted connection. Successful exploitation requires an on-path network position and the corresponding private key for a revoked certificate. Impact is scoped to data carried within the intercepted TLS connection. Patches are available across affected driver versions and users must manually upgrade to remediate the issue.
View original advisory →OpenPanel versions before 2.3.0 are affected by an unauthenticated server-side request forgery (SSRF) vulnerability. The vulnerability exists in the /misc/favicon and /misc/og endpoints, which accept an attacker-supplied url parameter without sufficient validation. Attackers can exploit this flaw to force the API to fetch arbitrary internal hosts and cloud metadata endpoints. Because small responses are returned verbatim to the attacker, this enables credential theft and internal service enumeration. No authentication is required to exploit this vulnerability, significantly raising its risk. Cloud environments are particularly at risk due to exposure of metadata services such as AWS IMDSv1. The issue is resolved in OpenPanel version 2.3.0. Users are strongly advised to upgrade immediately. References include the NVD entry, a GitHub security advisory, and a VulnCheck advisory.
View original advisory →Chroma version 1.5.9 contains a vulnerability where the HNSW index parameters max_neighbors, ef_construction, and ef_search lack maximum bounds validation in collection-create requests. Unauthenticated attackers can exploit this by supplying arbitrarily large parameter values, leading to server memory exhaustion. The denial of service condition is triggered during index compaction operations. No authentication is required to exploit this vulnerability, significantly increasing its risk. The vulnerability is tracked as CVE-2026-85664 and affects the Rust-based frontend and types components of the Chroma vector database. A GitHub issue and VulnCheck advisory have been published referencing this flaw. Organizations running Chroma 1.5.9 should assess exposure and apply mitigations or patches promptly.
View original advisory →A path traversal vulnerability (CVE-2026-85684) exists in the marker application through version 2.0.0. The vulnerability resides in the FastAPI /marker/upload handler, which fails to sanitize the file.filename parameter. Unauthenticated attackers can exploit this flaw by supplying filenames containing directory traversal sequences (e.g., '../'). Successful exploitation allows attackers to write arbitrary files to any location on the system or delete existing files. The vulnerability requires no authentication, significantly lowering the barrier for exploitation. The affected code is located in marker/scripts/server.py. Users are advised to upgrade beyond version 2.0.0 or apply mitigations to sanitize uploaded filenames. The issue has been tracked and discussed in the marker GitHub repository issue #1047.
View original advisory →A privilege escalation vulnerability exists in DirectIo64.sys, a kernel driver used by PassMark PerformanceTest, BurnInTest, and OSForensics. The flaw stems from missing validation of the physical address parameter in an exposed IOCTL handler. Local attackers can obtain a device handle and supply an arbitrary 64-bit physical address with a bit index to invoke MmMapIoSpace, allowing them to clear arbitrary bits at any physical memory address. This can be leveraged to corrupt kernel code pages or page table entries, enabling local privilege escalation or full system compromise. Affected versions include PerformanceTest before 11.1 build 1012, BurnInTest before 11.1 build 1000, and OSForensics before 11.1 build 1016. Fixed builds have been released by PassMark for all three products. The vulnerability is particularly concerning as it affects widely used benchmarking and forensics tools that install kernel-level drivers. No active exploitation in the wild has been reported at the time of disclosure.
View original advisory →cli-mcp-server version 0.2.5 contains a command allowlist bypass vulnerability in the _validate_command_with_operators function. The vulnerability is triggered when ALLOW_SHELL_OPERATORS is enabled, allowing attackers to use shell command substitution syntax such as $(...) or backticks to execute commands not included in the ALLOWED_COMMANDS allowlist. This effectively bypasses the intended security validation check designed to restrict which commands can be executed. The flaw resides in the server.py source file of the cli-mcp-server package. It has been documented in a GitHub issue and reported via VulnCheck advisories. Exploitation could allow unauthorized command execution in environments relying on the allowlist as a security boundary. Users are advised to review configurations and apply available patches or mitigations.
View original advisory →