CVE Feed

A security vulnerability has been detected in RooCodeInc Roo-Code up to 3.51.1. This affects the function fetch_instructions of the file malicious_mcp_server.py of the component MCP Integration Trust Model. The manipulation leads to code injection. The attack is possible to be carried out remotely. The exploit has been disclosed publicly and may be used. Multiple isses were reported to the vendor beforehand. They explain, that "they all apply to Roo Code, a project we no longer support - the repository was archived a while ago, and we don't encourage anyone to use it." This vulnerability only affects products that are no longer supported by the maintainer.

Ceph is an open-source distributed storage platform providing object, block, and file storage. In versions prior to 20.2.4 and 19.2.6, the CephX authentication protocol encrypts tickets with AES-128-CBC in an unauthenticated mode that uses a hard-coded initialization vector and no message authentication, allowing an attacker to forge credentials and gain cluster-wide access. Because the ciphertext is malleable and the monitor will encrypt attacker-chosen entity names, an attacker holding one low-privilege key and able to observe CephX traffic can use the monitor as an encryption oracle and splice ciphertext blocks into valid tickets for privileged entities such as Manager, MDS, and OSD. The same lack of authentication also lets an attacker with CephX permissions escalate privileges by flipping a single bit in a service ticket to set its allow_all field to true. This issue is fixed in versions 20.2.4 and 19.2.6.

PayRange API is missing proper authorization on management endpoints, which allows verbose details of every device on the PayRange network to be publicly accessible, with or without an account.

An authenticated OS command injection vulnerability exists in ZoneMinder's event export functionality. The exportFile HTTP request parameter is passed unsanitized into a shell command executed via PHP's exec(), allowing any authenticated user with View Events permission to execute arbitrary operating system commands on the server.

FastGPT is an open-source LLM platform for building AI applications on a knowledge base. In versions prior to 4.15.2, the WeChat (iLink) share-channel endpoints authorize requests using only the public shareId, with no authenticated identity or team-ownership check. As a result, an unauthenticated attacker who knows a victim team's shareId can take that team's WeChat bot offline or hijack the channel to their own bot: the logout endpoint is gated only by an existence check yet wipes the outLink's stored WeChat token, and the QR-code status endpoint performs no authorization at all and writes attacker-supplied bot credentials into the outLink identified by shareId. By generating a QR for a victim shareId, scanning it with their own WeChat, and calling the status endpoint, an attacker binds the victim team's app to the attacker's bot, exposing the app's private responses, displacing the legitimate binding, and consuming the victim's resources. The shareId is exposed in every shared chat URL, iframe, and embed, so it is not a secret. This issue is fixed in version 4.15.2.

ASE2000 2.35 through 2.37 is vulnerable to an improper certificate validation vulnerability, which may allow an attacker to impersonate the trusted peer, complete the TLS handshake, and read or modify protected communications.

Frappe 15.11.0 through 16.32.0 stores and renders the workspace card description without XSS filtering. The description field of the Workspace Link doctype is declared with "ignore_xss_filter": 1 in frappe/desk/doctype/workspace_link/workspace_link.json, and _sanitize_content() in frappe/model/base_document.py skips any field carrying that flag, so the value is stored exactly as submitted. frappe.desk.desktop.get_desktop_page returns it unchanged, and LinksWidget.set_body() in frappe/public/js/frappe/widgets/links_widget.js interpolates it into a Bootstrap popover created with html: true, by way of the __() translation helper, which performs no HTML escaping. A user holding the Workspace Manager role can therefore place arbitrary markup in a public workspace's card description and have it execute in the session of any desk user who opens that workspace and hovers the description, including higher-privileged users, allowing session token theft and authenticated requests as the victim. The flag is removed on the development branch but remains present in the shipped 15.x and 16.x release lines, so no released version carries the fix.

Dolibarr 9.0.0 through 23.0.4 saves inbound email attachments under the name supplied in the message's MIME headers without reducing it to a safe basename. The global saveAttachment() in htdocs/emailcollector/lib/emailcollector.lib.php builds $filepath = $path . $filename . '.' . $ext and hands it to file_put_contents(), and the private saveAttachment() in htdocs/emailcollector/class/emailcollector.class.php writes to $destdir.'/'.$filename; the name reaches both from the attachment's own getName() or getFilename() value by way of the record-join, create-ticket and create-project operations. A traversal sequence in the filename therefore survives intact, so any sender who can email a mailbox that an EmailCollector monitors, which is the module's ordinary use for a support or ticket inbox, can place attacker-controlled content outside the per-object attachment directory without holding a Dolibarr account. Under the hardened layout Dolibarr's SECURITY.md requires, with htdocs read-only, the write is confined to the documents tree and corrupts or forges other objects' documents; where htdocs is writable the same primitive reaches a web-executable path. Version 24.0.0 applies dol_sanitizePathName() and dol_sanitizeFileName() before the write.

Dolibarr before 23.0.4 authorizes REST API document deletion against the wrong permission. Documents::delete() in htdocs/api/class/api_documents.class.php calls dol_check_secure_access_document() with the mode argument 'read' when handling DELETE /api/index.php/documents, while the sibling builddoc() path passes 'write', the correct mode for an operation that modifies stored data. An authenticated API user who holds only a read permission for a document-bearing module, for example societe:lire or facture:lire, and no create, write, delete or admin permission, therefore passes the check and can permanently delete that module's documents: third-party files, invoices, orders, proposals, project files and generated PDFs, with no recovery path. The call site is htdocs/api/class/api_documents.class.php:1276 in 23.0.3 and passes 'write' from 23.0.4 onward.

Dolibarr before 24.0.0 contains a SQL injection in its CSV and XLSX import wizard. The wizard reads its update keys with GETPOST('updatekeys', 'array') in htdocs/imports/import.php, which applies only the generic alphanohtml filter: that strips HTML but leaves SQL keywords, comment markers, parentheses, spaces and quotes intact. import_insert() in htdocs/core/modules/import/import_csv.modules.php then iterates the submitted values and builds a filter with $where[] = $key.' = '.$data[$key], having first applied preg_replace('/^.*\./i', '', $key), an alias strip that does nothing to a value containing no dot. The assembled string is executed through $this->db->query(). The injected SELECT resolves the row id that the import then assigns to $lastinsertid, which becomes the WHERE target of a subsequent UPDATE, so a UNION SELECT returning an attacker-chosen integer both exfiltrates arbitrary table content and redirects which row the import overwrites; for category link tables the raw filter array is spliced into that UPDATE directly. The interface offers a fixed list of legitimate column codes but the server never checks the submitted values against it. A user holding the import permission can exploit this. Release 23.0.4 does not carry the fix; the allow-list test was added in 24.0.0.

A relative path traversal issue in the zip extraction functionality in AWS diagram-as-code (awsdac) in versions 0.10 through 0.23 can allow a third party to write arbitrary files to the local filesystem via crafted zip entry names containing path traversal sequences. This could allow the third party to perform inappropriate actions in the diagram bundle. To remediate this issue, users should upgrade to the version 0.24 or later.

A weakness has been identified in RooCodeInc Roo-Code up to 3.51.1. Affected by this issue is the function ExecaTerminalProcess of the component README File Handler. Executing a manipulation can lead to code injection. The attack can be executed remotely. The exploit has been made available to the public and could be used for attacks. Multiple isses were reported to the vendor beforehand. They explain, that "they all apply to Roo Code, a project we no longer support - the repository was archived a while ago, and we don't encourage anyone to use it." This vulnerability only affects products that are no longer supported by the maintainer.

Trilium is an open-source hierarchical note-taking application. In versions up to and including 0.103.0, the public share-search endpoint does not enforce the per-note shareCredentials and shareHiddenFromTree controls, allowing an unauthenticated visitor to read the titles, tree paths, and content of protected shared notes. The endpoint authorizes only the ancestor note supplied in the request and then runs a full-text search across the entire published subtree, returning each matching note's title, share identifier, and hierarchical path without re-checking whether that individual note requires a share password or is hidden from the navigation tree. Because the search matches note content, an attacker can enumerate protected notes and use the endpoint as a boolean oracle that confirms arbitrary substrings, recovering the full contents of notes that should be gated behind a password. This issue is fixed in version 0.104.0.

Trilium is an open-source hierarchical note-taking application. In versions prior to 0.104.0, the automatic image-download feature accepts file:// URLs in a note's img tags and reads the referenced local file with no path validation, allowing any authenticated user to disclose arbitrary files readable by the Trilium process. When a text note is saved, Trilium scans its HTML for image sources and downloads each external one; because the HTML sanitizer keeps file as an allowed scheme, a source such as file:///etc/passwd is passed straight to a filesystem read and its contents are stored as a note attachment the user can then retrieve. Pointing the same primitive at an unbounded source such as /dev/zero causes uncontrolled memory allocation that crashes the server process. The feature is enabled by default and is reachable through the web UI, the ETAPI, the web clipper, and note imports, requiring only an authenticated session or an ETAPI token. This issue is fixed in version 0.104.0

A security flaw has been discovered in RooCodeInc Roo-Code up to 3.51.1. Affected by this vulnerability is the function optimizeQuery of the file src/utils/helpers.ts of the component CodeIndexManager. Performing a manipulation results in code injection. Remote exploitation of the attack is possible. The exploit has been released to the public and may be used for attacks. Multiple isses were reported to the vendor beforehand. They explain, that "they all apply to Roo Code, a project we no longer support - the repository was archived a while ago, and we don't encourage anyone to use it." This vulnerability only affects products that are no longer supported by the maintainer.

Redis contains a use-after-free vulnerability in the 'tlsProcessPendingData()' function, which handles the TLS pending-data list if Redis is configured with TLS support. A remote, unauthenticated attacker may be able to execute arbitrary commands with the privileges of the Redis server.

SpeechBrain before 1.1.1 contains an arbitrary code execution vulnerability that allows attackers to execute arbitrary code by supplying a crafted CKPT.yaml checkpoint metadata file parsed with PyYAML's unsafe loader during candidate enumeration in Checkpointer.recover_if_possible(). Attackers can embed malicious Python object construction tags such as !!python/object/apply in any CKPT.yaml file within the configured checkpoint path to trigger code execution during candidate discovery, even if the malicious checkpoint is never selected for recovery.

Unrestricted Upload of File with Dangerous Type in the product photo upload in Roskus Prospero Flow CRM before 5.16.0 allows an authenticated user holding the create product permission (routine Seller role) to execute arbitrary JavaScript in the application origin. The photo validation rule classifies the file only by its content (magic bytes) and rejects only a fixed list of PHP extensions, while ProductSaveController::save() names the stored file using the client-supplied extension and copies it into the public web root. A file that begins with an image header and carries an HTML extension passes validation, is stored under public/asset/upload/product/, and is served with a text/html content type, turning the upload into first-party stored script execution.

Trilium is an open-source hierarchical note-taking application. In versions up to and including 0.103.0, the default-on "Safe import" filter sanitizes HTML only for text notes and excludes the book note type, whose content is stored without sanitization and later rendered as HTML, allowing an attacker-supplied import archive to embed a payload that executes as script. A book note's content is routed through the same rendering path as text notes and injected into the page with jQuery's html method when the note is shown as a grid-view preview card, so a malicious note survives Safe import and runs as soon as the victim opens the containing note. On the desktop client the Electron renderer runs with Node integration enabled, so the injected JavaScript escalates from cross-site scripting to full remote code execution on the victim's machine. This issue is fixed in version 0.104.0.

Trilium is an open-source hierarchical note-taking application. In versions up to and including 0.103.0, the default-on "Safe import" filter sanitizes HTML only for text notes and excludes the mindMap note type, whose JSON content is stored without sanitization, allowing an attacker-supplied import archive to embed a payload that renders as arbitrary HTML. A mind map node can carry a dangerouslySetInnerHTML property that the Mind Elixir library assigns directly to a node's innerHTML, so a malicious note survives Safe import and executes script as soon as the victim opens the imported mind map. On the desktop client the Electron renderer runs with Node integration enabled, so the injected JavaScript escalates from cross-site scripting to full remote code execution on the victim's machine. This issue is fixed in version 0.104.0.

Passing an effectively empty string to the `,ccs=` syntax extension of the mode argument in the `fopen` function in the GNU C Library version 2.45 or earlier may result in a heap buffer overflow when the mode string input to the function is attacker controlled. This usage pattern is not seen in applications in common GNU/Linux distributions and applications that process user-supplied values for `ccs` should not pass them through without validation.

ServiceNow has remediated a SQL injection vulnerability that was identified in in the ServiceNow AI platform. This vulnerability could enable an unauthenticated user, in certain circumstances, to execute arbitrary SQL statements against the instance's underlying database and gain access to, or modify, instance data beyond what was intended.  ServiceNow deployed a security update to hosted instances and ServiceNow provided the update to our partners and self-hosted customers. We are not currently aware of malicious exploitation against ServiceNow instances.  We recommend customers promptly apply appropriate updates or upgrade to a patched release if they have not already done so.

ServiceNow has remediated an improper access control vulnerability that was identified in the ServiceNow AI platform. This vulnerability could enable an unauthenticated user, in certain circumstances, to create or modify instance data beyond what was intended, resulting in privilege escalation.  ServiceNow deployed a security update to hosted instances and ServiceNow provided the update to our partners and self-hosted customers. We are not currently aware of exploitation against ServiceNow instances.  We recommend customers promptly apply appropriate updates or upgrade to a patched release if they have not already done so.

ServiceNow has remediated a code injection vulnerability that was identified in the ServiceNow AI platform. This vulnerability could enable an unauthenticated user, in certain circumstances, to execute arbitrary code in the ServiceNow platform and gain access to, or modify, instance data beyond what was intended.  ServiceNow deployed a security update to hosted instances and ServiceNow provided the update to our partners and self-hosted customers. We are not currently aware of malicious exploitation against ServiceNow instances.  We recommend customers promptly apply appropriate updates or upgrade to a patched release if they have not already done so.

ServiceNow has remediated a sandbox escape security issue that was identified in the ServiceNow AI Platform. This security issue could allow an unauthenticated user to execute arbitrary code within the ServiceNow AI Platform, potentially leading to more access to the ServiceNow AI Platform than intended.  ServiceNow deployed a security update to hosted instances and ServiceNow provided the update to our partners and self-hosted customers. We are not currently aware of malicious exploitation against ServiceNow instances.  We recommend customers promptly apply appropriate updates or upgrade to a patched release if they have not already done so.

DNG SDK versions 1.7.1 2502 and earlier are affected by an out-of-bounds read vulnerability that could lead to memory exposure. An attacker could leverage this vulnerability to disclose sensitive information from memory. Exploitation of this issue requires user interaction in that a victim must open a malicious file.

DNG SDK versions 1.7.1 2502 and earlier are affected by an out-of-bounds write vulnerability that could lead to application denial-of-service. An attacker could leverage this vulnerability to corrupt memory, causing the application to crash or become unresponsive. Exploitation of this issue requires user interaction in that a victim must open a malicious file.

A weakness in the client-side encryption configuration surface of the MongoDB C# Driver causes sensitive key-management credential material supplied by the application to be reproduced verbatim in the driver's human-readable diagnostic representation of its client settings, instead of being masked as other secret fields are. A party able to read the application's logs, diagnostic output, or a process memory dump may thereby recover the plaintext credentials and use them to decrypt protected field data.

Improper neutralization of delimiters in connection-URL construction allows connection-option injection in the MongoDB C# Driver. When an application passes untrusted text into the driver's connection-URL builder and round-trips the builder back into a client configuration, the untrusted text is serialized without neutralizing the URL/option delimiters and is then re-parsed as authoritative connection options. A low-privileged user of such an application can thereby introduce or suppress security-relevant connection settings.

A MongoDB C# driver document-replacement code path omits the element-name/shape validation that the equivalent write paths apply, so a value supplied as a replacement is forwarded to the server without neutralization of query-language special elements. An application that passes untrusted, loosely-typed input as a replacement value therefore allows that input to be interpreted by the database as update logic rather than as data, executing under the application's own database credentials. Applications using strongly-typed document mappings are not affected.

A NoSQL/expression injection weakness exists in the LINQ-to-aggregation query translation layer of the MongoDB C# Driver, in both aggregation expression and query filter translation. When application-supplied values are embedded in certain query constructs, special elements contained within those values are not properly escaped before the resulting query is transmitted to the database, so portions of the value may be interpreted by the database as query logic rather than as data. A user able to supply values that an application incorporates into an affected query may thereby cause unintended data to be returned or query results to be altered.

The MongoDB Rust Driver does not neutralize special characters in a caller-supplied target identifier before embedding it in the request it sends to the server. An actor able to influence that identifier in an application using the driver may cause write operations to be applied to an unintended target within the same deployment using the application's own credentials. This may result in unauthorized modification of data belonging to another logical boundary enforced by the application.

The MongoDB client library for PHP does not sufficiently sanitize special elements in application-supplied namespace identifiers before using them to construct the target namespace for database operations. An application that incorporates untrusted text into these identifiers may have operations silently directed at a different storage location than the one the application intended.

A weakness in the MongoDB C Driver allows special elements in caller-supplied database and collection name components to pass without sanitization when the driver composes the target namespace for an operation. An application that incorporates untrusted input into these name components can have operations directed at a resource other than the one intended.

A missing input-validation issue in MongoDB libmongocrypt's automatic-encryption context setup allows a caller-supplied database identifier to be accepted without sanitization. The resulting impact is limited to incorrect schema selection, which may lead to limited disclosure or modification of information handled by the application.

A weakness in the MongoDB C++ Driver's handling of caller-supplied namespace identifiers allows special characters embedded in those identifiers. An application that builds a namespace identifier from untrusted input without validating it may therefore have its operation directed at a different target than intended. This can result in limited unauthorized read and write access to data belonging to another logical tenant of the affected application.

The MongoDB Go Driver's client-level bulk write operation may accept a caller-supplied database name containing a reserved separator character without escaping it before the name is used to build the target namespace for the operation. An application that passes untrusted input as a database name could therefore have the write directed at a database and collection other than the ones it intended. Only the Client.BulkWrite API is affected.

When an IntegrationFlow uses .fluxTransform() with an asynchronous/reordering fluxFunction that emits raw payloads, concurrent requests on the same FluxMessageChannel subscription have their reply headers (replyChannel, errorChannel, correlationId, any propagated security/tenant headers) copied from whichever message was most recently consumed upstream. Spring Integration 7.1.0 Spring Integration 7.0.0 - 7.0.5 Spring Integration 6.5.0 - 6.5.10 Spring Integration 6.4.0 - 6.4.12 Spring Integration 5.5.21 and earlier

The EmbeddedHeadersJsonMessageMapper defaults to an overly permissive header parsing posture in its constructor. When decodeNativeFormat processes raw byte payloads, it deserializes embedded JSON headers into a plain Map and constructs a GenericMessage with MutableMessageHeaders without sanitizing or filtering untrusted header names by default. Spring Integration 7.1.0 Spring Integration 7.0.0 - 7.0.5 Spring Integration 6.5.0 - 6.5.10 Spring Integration 6.4.0 - 6.4.12 Spring Integration 5.5.21 and earlier

A single ScriptEngine instance is reused for every message on a script-backed channel. For JSR-223 engines that report THREADING=null (not thread-safe, e.g. the Kotlin kts engine), concurrent message processing can corrupt engine-internal state, potentially leaking one message's payload/headers bindings into another message's script evaluation or throwing spurious exceptions. Spring Integration 7.1.0 Spring Integration 7.0.0 - 7.0.5 Spring Integration 6.5.0 - 6.5.10 Spring Integration 6.4.0 - 6.4.12 Spring Integration 5.5.21 and earlier

When a container-level ErrorHandler is configured (the mitigation for finding 221000), each delivery whose processing throws still permanently consumes one link credit. After initialCredits (default 100) failing messages the receiver's credit reaches zero and the broker stops delivering, leaving the listener silently stalled while isRunning() remains true. Spring AMQP 4.1.0

RedisChatMemoryRepository.findByMetadata() builds RediSearch tag and text queries from caller-supplied metadata values without applying RediSearchUtil.escape(), unlike get(), clear(), and findByTimeRange() in the same class which do escape their inputs. An application that passes user-controlled values to findByMetadata() on a tag-typed metadata field allows an attacker to inject RediSearch syntax (e.g. x} | *) that breaks out of the tag clause and matches all indexed chat messages across every conversation in the index. Spring AI 2.0.0

DeadLetterPublishingRecovererFactory reads the retry_topic-original-timestamp header from an inbound ConsumerRecord and passes its raw bytes directly to new BigInteger(header.value()) with no length or format validation. Spring for Apache Kafka 4.1.0 Spring for Apache Kafka 4.0.0 - 4.0.6 Spring for Apache Kafka 3.0.0 - 3.3.16 Spring for Apache Kafka 2.9.0 - 2.9.14 Spring for Apache Kafka 2.8.12 and earlier

Spring Authorization Server's default consent page renders user-controlled values without HTML entity encoding. When using the DefaultConsentPage, an attacker can craft an OAuth2 authorization request containing a malicious value that is stored server-side and later rendered unencoded in the default consent page presented to the end user. Spring Authorization Server 1.5.0 - 1.5.8 Spring Authorization Server 1.4.0 - 1.4.11

The Spring Cloud Config Monitor is susceptible to Denial of Service attacks via malicious payloads. Spring Cloud Config 5.0.0 - 5.0.4 Spring Cloud Config 4.3.0 - 4.3.4 Spring Cloud Config 4.0.0 - 4.2.8 Spring Cloud Config 3.1.14 and earlier

Applications that build a Content-Disposition header value from untrusted input may be vulnerable to HTTP response splitting when the input is a malicious file name. Spring Framework 7.0.0 - 7.0.8 Spring Framework 6.2.0 - 6.2.19 Spring Framework 6.1.0 - 6.1.28 Spring Framework 6.0.0 - 6.0.30 Spring Framework 5.3.0 - 5.3.49 Spring Framework 5.2.25.RELEASE and earlier

A local unprivileged user on the same host can redirect all Zip/UnZip transformer output into a directory of their choosing by pre-creating /tmp/ziptransformer as a symlink before the application starts. Spring Integration 7.1.0 Spring Integration 7.0.0 - 7.0.5 Spring Integration 6.5.0 - 6.5.10 Spring Integration 6.4.0 - 6.4.12

Potential for deserialization of untrusted types in Spring Cloud Stream. Spring Cloud Stream 5.0.0 - 5.0.2 Spring Cloud Stream 4.3.0 - 4.3.3 Spring Cloud Stream 4.2.0 - 4.2.6

Spring MVC applications using the functional web framework are vulnerable to stream corruption when using Server-Sent Events (SSE). Spring Framework 7.0.0 - 7.0.8 Spring Framework 6.2.0 - 6.2.19 Spring Framework 6.1.0 - 6.1.28 Spring Framework 6.0.0 - 6.0.30 Spring Framework 5.3.0 - 5.3.49

An operator who calls JdbcMessageStore.addAllowedPatterns(...) to restrict deserialization receives no protection at all when the store is a Spring-managed bean. Spring Integration 7.1.0 Spring Integration 7.0.0 - 7.0.5 Spring Integration 6.5.0 - 6.5.10 Spring Integration 6.4.0 - 6.4.12