CVE Feed

FlexPrice through 2.1.38 contains an authentication bypass vulnerability in HandleQuickBooksWebhook that skips signature verification when the intuit-signature header is omitted. Unauthenticated attackers knowing a tenant and environment ID can replay forged payment notifications to repeatedly add payment amounts, marking invoices paid or overpaid.

AstronRPA through 1.1.6 contains an insecure direct object reference vulnerability in the resource-service /file/download endpoint that allows authenticated users to download other tenants' files by supplying their fileId. Attackers can enumerate predictable time-ordered Snowflake file IDs, since FileMapper.getFile performs no owner or tenant check, to retrieve S3-stored files uploaded by other tenants.

lamp-boot 4.16.0-java17 through 4.21.2-java17 contains an insecure direct object reference vulnerability in the oauth module's /anyone/findDeptByCompany endpoint that skips URI permission checks. Authenticated low-privilege users can iterate sequential employeeId values to read every employee's department membership within the tenant.

bbPress before 2.6.19 contains an information disclosure vulnerability that allows unauthenticated visitors to view restricted subforum posts because hide_sitewide ignores ancestor forum visibility. Topics and replies in public subforums nested under private or hidden parents expose title, permalink and body via the BuddyPress activity stream and REST endpoint.

ILIAS before 9.24, 10.12, and 11.5 contains a PHP object injection vulnerability that allows authenticated users with Test creation or import permission to deserialize arbitrary objects via crafted Test archives. Attackers can inject a malicious tst_sequence sequence attribute that is passed to unserialize(), abusing Guzzle or Monolog gadget chains to write PHP files and execute code.

DataEase is an open source data visualization and analysis tool. Prior to 2.10.26, authenticated users with folder creation permission can store HTML-bearing names through datasource/createFolder or datasetTree/save. The datasource and dataset Move to dialogs in core/core-frontend/src/views/visualized/data/datasource/form/CreatDsGroup.vue and core/core-frontend/src/views/visualized/data/dataset/form/CreatDsGroup.vue read the stored name from a node title attribute and assign search-highlighted content to innerHTML. When another user opens a Move to dialog and types in its search box, attacker-controlled JavaScript can execute in the victim's DataEase session. This issue is fixed in version 2.10.26.

A vulnerability was identified in cemtan sar2html 4.0.0. Affected is the function updateDb of the file sar2html.py of the component File Upload. The manipulation leads to path traversal. Remote exploitation of the attack is possible. The project was informed of the problem early through an issue report but has not responded yet.

DataEase is an open source data visualization and analysis tool. Prior to 2.10.26, authenticated users who can edit dataset fields and chart configuration can store an HTML-bearing dimension value in a bar-horizontal chart. When lengthLimit truncates the category axis label, configAxisLabelLengthLimit in core/core-frontend/src/views/chart/components/js/panel/common/common_antv.ts retrieves the full label and assigns it to the g2-tooltip-title element through innerHTML. A victim who opens the dashboard and hovers the truncated label can execute attacker-controlled JavaScript in the victim's DataEase session. This issue is fixed in version 2.10.26.

In Eclipse Vert.x versions [5.0.0,5.0.12] or [5.1.0,5.1.9] or [5.2.0, 5.2.0], the Vert.x HTTP/1.x decoder uses a canonicalization for well known HTTP headers, the canonicalization process interprets well crafted headers as headers of this list, allowing for instance Content-Xength to be interpeted as Content-Length. This can be exploited for request smuggling purpose. The impacted component is io.vertx:vertx-core.

DataEase is an open source data visualization and analysis tool. Prior to 2.10.26, authenticated users can store a crafted custom dateFormat for a dataset field. DatasetTableFieldDTO.dateFormat is inserted without validation into date conversion templates used by Dimension2SQLObj, CustomWhere2Str, Field2SQLObj, WhereTree2Str, and ExtWhere2Str. When a chart renders the field on a cross-datasource dataset, Calcite parsing is bypassed and the crafted format can directly alter the generated SQL. On a single-datasource dataset, exploitation additionally requires the payload to survive Calcite reserialization as a valid string literal. Successful exploitation can compromise the confidentiality, integrity, and availability of the backend database. This issue is fixed in version 2.10.26.

DataEase is an open source data visualization and analysis tool. Prior to 2.10.26, authenticated users can store crafted FieldGroupDTO.min, FieldGroupDTO.max, FieldGroupDTO.startTime, or FieldGroupDTO.endTime values in dataset group-field binning definitions. Utils.transGroupFieldToSql inserts these free-form values into generated CASE expressions without validation, and the resulting SQL can reach unparameterized backend query execution when a chart uses the field. This stored SQL injection can compromise the confidentiality, integrity, and availability of the configured backend database. This issue is fixed in version 2.10.26.

DataEase is an open source data visualization and analysis tool. Authenticated users with datasource access can supply crafted schema names, connection usernames, table names, or database names to datasource metadata operations, including getSchema, validate, save, and update flows. CalciteProvider inserts these identifiers into metadata queries that are executed through a plain JDBC Statement, allowing the generated SQL to be altered and run against the configured backend with application-held datasource credentials. The issue can compromise the confidentiality, integrity, and availability of the connected database without requiring a victim action.

Dell OS Recovery Tool, versions prior to 2.4.4, contains an Insecure Storage of Sensitive Information vulnerability. A low privileged attacker with physical access could potentially exploit this vulnerability, leading to Information Disclosure.

Radius is a cloud-native, portable application platform that makes app development easier for teams building cloud-native apps. Prior to 0.58.0, the Kubernetes controller deserializes the user-controlled radapp.io/status annotation on a Deployment and uses its container resource identifier for deletion without confirming ownership by that Deployment. An attacker with permission to modify Deployment annotations and knowledge of a target Radius resource ID can forge radapp.io/status and set radapp.io/enabled to false, causing reconciliation to enter the privileged deletion path for another tenant's Radius-managed container resource. In a multi-tenant installation this confused-deputy condition can delete another team's recoverable container resource and disrupt service; in a single-tenant installation the impact is limited to self denial of service. The multi-tenant Radius controller reads the annotations in pkg/controller/reconciler/annotations.go and reaches the privileged delete in pkg/controller/reconciler/deployment_reconciler.go. The issue does not disclose data, grant privileges, or provide persistence. This issue is fixed in version 0.58.0.

In Eclipse Vert.x versions [4.0.0, 4.5.34] or [5.0.0,5.0.12] or [5.1.0,5.1.9] or [5.2.0, 5.2.0], a scope bypass vulnerability exists in Eclipse Vert.x Web's ChainAuthHandler when used with scope-requiring authentication handlers such as JWTAuthHandler or OAuth2AuthHandler. When a user is already authenticated on the routing context by a preceding handler in the request pipeline, ChainAuthHandler skips all scope validation configured via withScope() or withScopes(). This affects both any() and all() chain modes. The root cause is that postAuthentication() relies on internal state that is only set during the authenticate() phase, which is skipped entirely for pre-authenticated users. In any() mode, the absence of this state causes the method to fall through without delegating scope checks to any handler. In all() mode, scope checks are unconditionally skipped. This allows a user with a valid token but insufficient scopes to access resources protected by scope requirements, provided they were authenticated by an earlier handler in the pipeline. The most common trigger is overlapping authentication handlers on nested routes, where a parent route authenticates broadly and a child route uses a chain to enforce specific scopes. The impacted component Maven coordinate is io.vertx:vertx-web

Insertion of sensitive information into log files in the git_utils[.]py component in Amazon SageMaker Python SDK V3 before 3.10.0 and V2 before 2.257.3 might allow an authenticated remote user to obtain plaintext repository credentials from CloudWatch Logs when a git clone operation fails, because the runclone_command() function does not redact credentials from subprocess error output. To remediate this issue, users should upgrade to Amazon SageMaker Python SDK V3 3.10.0 or V2 2.257.3.

Dgraph is an open source distributed GraphQL database. Prior to 25.4.0, Dgraph Zero registers administrative HTTP routes on port 6080 without authentication when the administrative interface is enabled. The /removeNode route accepts attacker-controlled id and group values and passes them to Server.RemoveNode, where group 0 can trigger ProposePeerRemoval. A remote unauthenticated attacker who can reach the Zero HTTP port can read topology through /state, allocate identifiers through /assign, access the exposed /moveTablet administrative route, or remove membership through /removeNode, disrupting the Zero control plane and destabilizing Alpha nodes that depend on Zero. The route registration occurs in dgraph/cmd/zero/run.go, and the demonstrated membership-removal path reaches Server.RemoveNode and ProposePeerRemoval. This issue is fixed in version 25.4.0.

Multiple TP-Link Wi-Fi range extenders use a custom, reversible XOR-and-dictionary transformation in the HTTP web-management login flow. An unauthenticated attacker on the same local network who can capture the network traffic associated with legitimate administrator login may be able to derive an alternative credential value that produces the same accepted authentication result and use it to authenticate to the device’s web-management interface with administrator privileges. Exploitation requires the management login to use HTTP and the attacker to observe the session key, dictionary, and login value associated with a legitimate administrator login. Successful exploitation may allow an unauthenticated adjacent attacker to gain administrator access to the affected device and modify exposed configuration settings.

Pipecat is an open-source Python framework for building real-time voice and multimodal conversational agents. From 0.0.98 until 1.6.0, the pipecat-ai package's built-in development runner exposes unauthenticated endpoints including /start, /daily, and /daily-dialin-webhook that can create bot sessions and Daily rooms. The runner binds to localhost by default and is intended only for local development, so exploitation requires an operator to expose it to untrusted traffic through --host 0.0.0.0, a published port, or a tunnel without access controls. An unauthenticated caller can consume local resources and third-party quotas and, depending on configuration, receive a room URL and join token. Version 1.6.0 adds a prominent development-runner banner and documentation that clarifies the unauthenticated boundary and directs production deployments to an authenticated dispatcher or managed runtime, while the endpoint behavior remains by design. This issue is fixed in version 1.6.0.

dbt enables data analysts and engineers to transform their data using the same practices that software engineers use to build applications. From 2.0.0-alpha.0 until 2.0.0-alpha.5, dbt docs serve exposes the unauthenticated POST /api/v1/query endpoint in crates/dbt-docs-server/src/handlers/query.rs, which passes caller-supplied SQL to DuckDbViewsBackend::query_arrow() in crates/dbt-index-core/src/backend.rs without authorization or an allowlist. A client that can reach the server can invoke DuckDB filesystem-reading functions to read files available to the dbt process, including files outside the docs index directory. The server binds to localhost by default, but operators can expose it to a network through the host option or DBT_DOCS_HOST setting. This issue is fixed in version 2.0.0-alpha.5.

Mongoose is an embedded web server and network library. Prior to 7.23, ppp_handle_ipv6cp() in src/l2_ppp.c trusts the 16-bit IPv6CP len field without checking it against the received payload size. On a PPP or PPPoE link using MG_ENABLE_TCPIP and MG_ENABLE_IPV6, an unauthenticated adjacent peer can complete the required LCP exchange and send a short Configure-Request with an oversized length, causing ppp_tx_frame() to read beyond recv_queue.buf and write beyond tx.buf. The same length misuse can also expose adjacent heap bytes in a reply or hang the event loop through find_opt(). This issue is fixed in version 7.23.

Mongoose is an embedded web server and network library. Prior to 7.23, the built-in TCP/IP stack enabled by MG_ENABLE_TCPIP in src/net_builtin.c does not reject a TCP option whose declared opts[1] length exceeds the remaining option bytes. In handle_opt(), a NOP followed by a truncated MSS option causes opts[3] to be read outside the declared options area and used as the low byte of dmss. An unauthenticated remote peer can corrupt the per-connection MSS value so later mg_send() calls transmit truncated or zero-length application payloads while the TCP connection otherwise appears healthy. This issue is fixed in version 7.23.

Mongoose is an embedded web server and network library. Prior to 7.23, the built-in TLS certificate parser selected by MG_TLS_BUILTIN fails to verify that a long-form DER element length fits within the remaining certificate buffer. In mg_der_to_tlv() and mg_der_next(), a peer-controlled length can advance the parser beyond the certificate and underflow the remaining uint32_t length, causing mg_tls_parse_cert_der() to continue walking an invalid pointer. An unauthenticated malicious TLS server, or a client presenting a certificate to a server configured for client certificates, can trigger an out-of-bounds read and denial of service during the handshake. This issue is fixed in version 7.23.

Mongoose is an embedded web server and network library. Prior to 7.23, the built-in TLS stack selected by MG_TLS_BUILTIN in src/tls_builtin.c accepts a TLS 1.3 record whose decrypted size is exactly the 16-byte authentication tag with no plaintext or ContentType byte. In mg_tls_recv_record(), the msgsz length check permits that record and the recv_len calculation underflows to SIZE_MAX, so a later handshake step treats stale buffer contents as a trusted message with an effectively unbounded length. An unauthenticated malicious TLS peer can trigger a heap out-of-bounds read before certificate or Finished validation, causing a remote denial of service. This issue is fixed in version 7.23.

LDAP Account Manager (LAM) is a webfrontend for managing entries (e.g. users, groups, DHCP settings) stored in an LDAP directory. Prior to 9.6, installations with WebAuthn two-factor authentication leave the challenge in the same PHP session after the authentication response is processed. A user who has already completed valid password authentication can therefore retain the same signed WebAuthn ceremony state instead of consuming it once. The practical risk is limited by the requirement for a valid password, the same PHP session, and a valid authenticator signature. This issue is fixed in version 9.6.

Stirling-PDF is a locally hosted web application that facilitates various operations on PDF files. Prior to 2.13.2, an authenticated administrator using the default H2 database and security mode can upload a crafted SQL file to POST /api/v1/database/import-database. DatabaseService.validateSqlContent() splits normalized input on semicolons and accepts a statement when it contains any allowed keyword, allowing dangerous H2 functions such as FILE_READ and CSVWRITE, plus CREATE ALIAS, to bypass the insufficient whitelist and execute side effects when RUNSCRIPT processes the file. This permits arbitrary file access. The demonstrated operating-system command execution chain writes attacker-controlled startup files to a user-site location and then triggers an application Python interpreter configured to load those startup files, resulting in command execution as the JVM service account. This issue is fixed in version 2.13.2.

cpp-httplib is a C++ header-only HTTP/HTTPS library. From 0.29.0 until 0.51.0, Client::open_stream in httplib.h accepts a caller-controlled path that reaches detail::write_request_line() without the detail::encode_path() protection used by the ordinary buffered request path. Raw carriage-return and line-feed bytes in the path can therefore split the outbound HTTP/1.1 request line and inject attacker-shaped request metadata. Exploitation requires an embedding application to pass remote user input into the open_stream path and a backend or intermediary that interprets the injected bytes; the established impact is modification of backend-visible request data, not direct confidentiality or availability impact. This issue is fixed in version 0.51.0.

cpp-httplib is a C++ header-only HTTP/HTTPS library. Prior to 0.51.0, create_redirect_client() removes Host, Proxy-Authorization, and Authorization headers when following a cross-origin 301, 302, 307, or 308 redirect but leaves Cookie headers attached. detail::redirect() copies the request for those redirect codes, so cookies supplied directly or through default_headers_ are sent to the redirect target. An attacker who controls a redirect response, or a network position for a non-HTTPS origin, can receive session cookies intended for the original origin. A 303 redirect is not affected because that path clears all request headers. This issue is fixed in version 0.51.0.

lichess.org is the forever free, adless and open source chess server. Prior to commit c54596c9e1569d378fe20f80d4c790c6fdee54c4, a user who can add a chapter to a study can place attacker-controlled text in the PGN Annotator tag, which becomes Comment.Author.External and is returned verbatim by authorText. When a study node has at least two comments, ui/analyse/src/treeView/inlineView.ts interpolates that author string into HTML passed to the innerHTML hook without escaping it. Opening the affected study can therefore render stored attacker-controlled HTML and CSS for other viewers, including navigation to an attacker-controlled site, content spoofing, and CSS-based data exposure, although the site's content security policy blocks injected JavaScript. This vulnerability is fixed by commit c54596c9e1569d378fe20f80d4c790c6fdee54c4.

A vulnerability was determined in D-Link DIR-868L 2.01b05. This impacts the function system of the file /htdocs/cgibin of the component HNAP CGI Handler. Executing a manipulation of the argument SOAPACTION can lead to os command injection. The attack may be launched remotely. The exploit has been publicly disclosed and may be utilized.

Flyto2 Core is an execution kernel for automation and AI-agent workflows. From 2.26.2 until 2.26.4, the HTTP MCP POST /mcp route accepts unauthenticated JSON-RPC tools/call requests and forwards execute_module arguments to the module registry without the require_auth dependency used by the REST execution routes. An unauthenticated caller can select sandbox.execute_shell and supply its command parameter, which reaches asyncio.create_subprocess_shell with the privileges of the flyto-core server process. The endpoint is reachable locally under the default loopback binding and becomes network reachable when an operator binds the server to a non-loopback interface, allowing arbitrary OS command execution and potential full host compromise. This issue is fixed in version 2.26.4.

Flyto2 Core is an execution kernel for automation and AI-agent workflows. Prior to 2.26.3, validate_url_ssrf and is_private_ip in src/core/utils.py compare resolved addresses only against native private IPv4 and IPv6 ranges and do not unwrap IPv4-mapped, IPv4-compatible, 6to4, or NAT64 transition addresses. An authenticated workflow author can provide a transition-form literal URL through the Execution API POST /v1/execute route, http.get, or sibling guarded outbound-request modules, causing the guard to accept an address whose embedded IPv4 points to loopback, RFC 1918, or cloud metadata services. On a host that routes the selected transition form, the outbound request reaches the internal destination and returns the response body to the workflow caller, enabling read SSRF and disclosure of internal-service data or cloud credentials. This issue is fixed in version 2.26.3.

e107 is a content management system (CMS). Prior to 2.3.7, install.php decodes the attacker-controlled previous_steps POST parameter and passes it to unserialize() at file scope and again in the e_install constructor without restricting allowed classes. The installer remains accessible after a normal installation, and a crafted object graph using the bundled GuzzleHttp\Psr7\PumpStream class can flow through paths.hash into site_path, where string coercion invokes an attacker-controlled callable. An unauthenticated attacker who can reach install.php can therefore execute arbitrary PHP functions with the web server process's privileges, leading to full server compromise. This issue is fixed in version 2.3.7.

e107 is a content management system (CMS). Prior to 2.3.7, the bundled alt_auth plugin's otherdb and e107db authentication backends interpolate the submitted login username directly into SQL in e107_plugins/alt_auth/otherdb_auth.php and e107_plugins/alt_auth/e107db_auth.php. The login() methods execute those strings with PDO::query() before effective sanitization, so an unauthenticated user can inject SQL when alt_auth is installed and either backend is configured with a live external database connection. The injection can expose or modify data available to the configured database account, and query failures can additionally disclose the generated query through the debug-message path. This issue is fixed in version 2.3.7.

In Eclipse Vert.x versions [5.0.0,5.0.12] or [5.1.0,5.1.9] or [5.2.0, 5.2.0], the Vert.x HTTP client relies on an SSL context for efficiency purpose, the key of this cache should reflect entirely the SSL semantics of the desired SSL properties. It is possible for a client to ask for an SSL context and obtain an SSL context with weaker properties, for instance a context with hostname validation and obtain a context ignoring hostname validation. The impacted component is io.vertx:vertx-core.

Appium is a cross-platform automation framework for all kinds of apps, built on top of the W3C WebDriver protocol. Prior to 7.2.6, @appium/support ZIP extraction accepts an archive entry that creates a symbolic link to a location outside the extraction directory and then accepts a later regular-file entry whose legitimate name resolves through that link. The lexical containment check examines entry names without validating the symlink target or the real destination, so an attacker-controlled archive can write files to arbitrary locations writable by the Appium process. This is distinct from the earlier parent-directory entry-name traversal fixed in 7.0.6 because the bypass uses no parent-directory or absolute entry name. Depending on the writable destination, the arbitrary file write can overwrite configuration, startup scripts, or executables and lead to code execution. This issue is fixed in version 7.2.6.

Missing Authorization vulnerability in Skydropx Skydropx skydropx allows Exploiting Incorrectly Configured Access Control Security Levels.This issue affects Skydropx: from n/a through 2.1.4.

Improper Neutralization of Input During Web Page Generation ('Cross-site Scripting') vulnerability in WPManageNinja Best Payments Plugin for WP wp-payment-form allows Stored XSS.This issue affects Best Payments Plugin for WP: from n/a through 4.6.27.

Improper Neutralization of Input During Web Page Generation ('Cross-site Scripting') vulnerability in hippooo Hippoo Mobile App for WooCommerce hippoo allows Reflected XSS.This issue affects Hippoo Mobile App for WooCommerce: from n/a through 1.12.0.

Improper Neutralization of Input During Web Page Generation ('Cross-site Scripting') vulnerability in SimplyRETS SimplyRETS Real Estate IDX simply-rets allows Reflected XSS.This issue affects SimplyRETS Real Estate IDX: from n/a through 3.2.9.

Improper Control of Generation of Code ('Code Injection') vulnerability in PluginUs.Net MDTF wp-meta-data-filter-and-taxonomy-filter allows Code Injection.This issue affects MDTF: from n/a through 1.3.9.

Improper Control of Generation of Code ('Code Injection') vulnerability in VillaTheme Salevo woo-boost-sales allows Code Injection.This issue affects Salevo: from n/a through 1.3.0.

Ludwig is a low-code framework for building custom LLMs, neural networks, and other AI models. Prior to 0.17.8, ludwig.datasets.archives.extract_archive() validated tar member names before calling TarFile.extractall() but did not validate symlink members or hard-link members. A crafted archive could create a link that escaped the extraction directory and then write a later member through that link, placing attacker-controlled content at an arbitrary filesystem path with the extracting process's privileges. The same containment check used os.path.commonprefix, so a sibling path sharing the destination's character prefix could also pass. The dataset-loading path can download and extract configured archive URLs, but exploitation requires a user to process an archive from a source controlled or compromised by the attacker on supported Python 3.12 and 3.13 runtimes. Python 3.14 is not affected because its default tar data filter rejects escaping links. This issue is fixed in version 0.17.8.

FOG is a free open-source cloning/imaging/rescue suite/inventory management system. Prior to 1.5.10.1870, the management adInfo() endpoint in packages/web/lib/fog/fogpage.class.php was reported to return the FOG_AD_DEFAULT_PASSWORD value in its domainpass JSON field and to be dispatchable through node=ipxe without an authenticated session. The response was also consumed by host and group Active Directory configuration interfaces. This could disclose the default domain-join credential to a remote unauthenticated requester. This vulnerability is fixed in 1.5.10.1870.

SeaweedFS is a distributed storage system for files and blobs. From 4.0.388 until 4.41, the SFTP Helm chart embedded one unencrypted Ed25519 private host key in k8s/charts/seaweedfs/templates/sftp/sftp-secret.yaml as stringData.seaweedfs_sftp_ssh_private_key. A standalone deployment presented that public key when sftp.enabled and sftp.enableAuth were true, sftp.existingConfigSecret was not supplied, the legacy sftp.sshPrivateKey path remained selected, a valid config-ssh Secret allowed startup, and hostKeysFolder did not provide a later Ed25519 signer. An attacker able to redirect or intercept client traffic could impersonate that SFTP endpoint to clients that trusted the published host identity, exposing password credentials or data sent to the rogue endpoint. Default all-in-one and non-authenticated standalone paths did not load the key. This issue is fixed in version 4.41.

Redash is a package for data visualization and sharing. From 26.0.0 until 26.9.0, the DuckDB query runner's _connect() path in redash/query_runner/duckdb.py created connections with enable_external_access left at its default value of true. A user with query access to an administrator-created DuckDB data source could use native file-reading functions to retrieve files readable by the Redash process or use COPY TO to write files with that process's privileges. Reading application secrets such as REDASH_COOKIE_SECRET or REDASH_SECRET_KEY could also enable forged sessions and authentication bypass. This issue is fixed in version 26.9.0.

ServiceNext contains an insecure direct object reference (IDOR) in the /api/Usuario/GetUsuarios endpoint, which accepts a user-controlled numeric ID in the request body without verifying that the authenticated token belongs to the requested account (CWE-639). Any authenticated user can iterate the ID parameter to retrieve the full profile record of every other user, including passwords stored in cleartext (CWE-256), enabling account takeover and credential-stuffing attacks.

Redash is a package for data visualization and sharing. From 10.0.0 until 26.9.0, the JSON data source query runner in redash/query_runner/json_ds.py used _get_all_results() to combine its configured base_url with a user-supplied query URL by using urljoin. An absolute query URL replaced the configured base URL instead of remaining under it, allowing a user with execute_query permission to direct a server-side request to an arbitrary external origin. When HTTP Basic Auth was configured for the data source, the request also forwarded those credentials to the selected origin. Private and internal address access remained constrained when REDASH_ENFORCE_PRIVATE_IP_BLOCK was enabled. This issue is fixed in version 26.9.0.

Redash is a package for data visualization and sharing. From versions after 8.0.0 until 26.9.0, the Query Results query runner in redash/query_runner/query_results.py accepts param_query_N_{key=value} parameters and passed attacker-controlled values to replace_query_parameters() for raw substitution into a referenced query's SQL. This path bypassed the ParameterizedQuery schema validation used by normal query execution and omitted the is_safe permission check. An authenticated user with at least view-only access to both the Query Results data source and the referenced data source could therefore inject SQL through text parameters and execute it with the referenced data source's privileges. This issue is fixed in version 26.9.0.

The Model Context Protocol Python SDK is the official Python SDK for Model Context Protocol servers and clients. In mcp versions before 1.29.1 and from 2.0.0a1 through 2.0.1, the Streamable HTTP session manager, SSE message endpoint, and OAuth authorization-server routes buffer complete HTTP request bodies without enforcing a maximum size. An unauthenticated remote client able to reach an affected endpoint can send a large declared or streamed request body that consumes multiple times the body size in process memory and can repeatedly exhaust memory or terminate the server. Stdio-only servers and deployments with an effective upstream request-size limit are not affected, and applications that use StreamableHTTPServerTransport directly remain responsible for enforcing their own limit. The standard SDK server applications and session managers are fixed in versions 1.29.1 and 2.1.0.