pgEdge Postgres MCP Server write confirmation skipped for writes that begin with a reading keyword

pgEdge Postgres MCP Server enables SQL queries against PostgreSQL databases through Model Context Protocol compatible clients. Prior to 1.1.0, the bundled CLI and web clients classified query_database statements only by their first keyword before deciding whether to display a write confirmation. SELECT INTO, data-modifying common table expressions beginning with WITH, and EXPLAIN ANALYZE of a write were treated as reads and could execute without confirmation on an allow_writes connection. Read-only connections remain protected by PostgreSQL transaction access mode, and EXPLAIN without ANALYZE and SELECT FOR UPDATE remain classified as reads. This issue is fixed in 1.1.0.

Memory exhaustion denial of service in Elasticsearch data source index patterns

Grafana is an open-source platform for monitoring and observability. When an Elasticsearch data source is configured with a time-based index pattern, Grafana generated one index name for every interval in the requested query time range without any upper limit. A user with permission to query the data source, which includes users holding only the Viewer role in a default configuration, could submit a single small query with an extreme time range and cause the server to allocate memory until it was terminated. The allocation occurred before any request was sent to Elasticsearch, so the condition applied whether or not the Elasticsearch cluster was reachable. Data sources configured with a static index name are not affected.

Unintended API access through the Google Cloud Monitoring data source resource proxy

A user with query access to a Google Cloud Monitoring data source can use its resource proxy to send requests to other Cloud Monitoring API paths with the data source's credentials, and read data from any project those credentials can reach. When the data source forwards the user's OAuth identity or uses Workload Identity Federation, the impact depends on that credential's permissions. Fixed in Google Cloud Monitoring plugin 12.6.7.

Elasticsearch data source resource proxy allows arbitrary methods and paths with the data source's credentials

The Elasticsearch data source resource proxy forwards the request method and path to Elasticsearch using the data source's own credentials, without restricting them to the operations the plugin needs. A user who can query the data source can alter index mappings or reach other indices on the cluster. [Improper authorization / path validation]

Bootgen Improper Authentication Allows Fixed Salt RSA Padding

Improper authentication in Bootgen could result in the use of a fixed salt value for RSA padding instead of a randomly generated salt, potentially impacting data integrity.

WordPress RestroPress plugin <= 3.4.11 - Cross Site Request Forgery (CSRF) vulnerability

Cross-Site Request Forgery (CSRF) vulnerability in Magnigenie RestroPress restropress allows Cross Site Request Forgery.This issue affects RestroPress: from n/a through 3.4.11.

BlueprintUE Self-Hosted Edition 1.0.0 before 4.3.5 Storage Exhaustion via Anonymous Blueprints

BlueprintUE Self-Hosted Edition 1.0.0 before 4.3.5 contains a resource exhaustion vulnerability that allows unauthenticated attackers to fill server storage by repeatedly creating anonymous blueprints. Attackers can send large requests to HomeController::doProcessCreateBlueprint() with a 'never' expiration, bypassing purges to fill the disk and deny service.

When parsing request variables such as query-string and form parameters, HHVM did not check that appending a nested array succeeded before descending into it. A remote, unauthenticated attacker could send a GET or POST request which would crash the HHVM server process before any application code runs, leading to denial of service. This issue affected HHVM prior to 23c061f4e804a51badd5f85227e65185651253b5.

Heym: SSRF via POST /api/mcp/fetch-tools: any authenticated user reaches local/private URLs through the SSE and streamable-HTTP transports

Heym builds and runs agentic systems and business-process automations on self-hosted infrastructure. Prior to version 0.0.70, POST /api/mcp/fetch-tools is reachable by any authenticated user (lowest role; no admin/workflow association) and fetches an attacker-supplied URL through the sse and streamable_http transports with no scheme / private-IP / allowlist / redirect guard. The transport error is reflected in the HTTP 422 response, giving a semi-blind SSRF oracle, and the attacker also controls the outbound request headers. This issue has been patched in version 0.0.70.

Heym: Server-Side Request Forgery (SSRF) in HTTP Workflow Node

Heym builds and runs agentic systems and business-process automations on self-hosted infrastructure. Prior to version 0.0.73, the HTTP workflow node in Heym makes outbound HTTP requests to user-supplied URLs without validation of the target host or IP address. If a workflow author has already configured an HTTP node that targets internal addresses such as cloud metadata endpoints or private network hosts, and exposed the workflow with anonymous execution, an unauthenticated attacker can trigger the preconfigured SSRF sink. The same project has comprehensive SSRF protection in its MCP integration, which is absent from the HTTP node. This issue has been patched in version 0.0.73.

PostgreSQL type confusion after ALTER ... SET DATA TYPE executes arbitrary code

Type confusion in PostgreSQL allows an object creator to execute arbitrary code as the operating system user running the database, via ALTER ... SET DATA TYPE primarily. Other DDL attack routes exist, such as "ALTER TABLE ... ADD COLUMN; ROLLBACK; ALTER TABLE ... ADD COLUMN". The system had inadequate tracking of typed values, and values were not self-identifying. Stale values could appear in data dictionary entries, PL/pgSQL variables, and many other parts of the system. PostgreSQL Versions before PostgreSQL 18.7, 17.12, 16.16, 15.20, and 14.25 are affected.

A potential improper permissions vulnerability was reported in Lenovo XiangBangbang that could allow a local application to perform unauthorized touch gestures.

proxygen's HTTP/3 codec buffered an incomplete non-DATA frame, such as a control-stream frame, a frame of an unknown or reserved type, or a HEADERS frame, until its entire declared length had arrived, without limiting that length, and re-scanned all buffered data each time more bytes arrived. A remote, unauthenticated HTTP/3 peer could declare a very large frame and deliver it slowly, causing unbounded memory growth and CPU usage that grows quadratically with the number of fragments, leading to denial of service. This issue affects proxygen versions prior to v2026.11.09.00.

WordPress SureForms plugin <= 2.13.0 - Broken Access Control vulnerability

Missing Authorization vulnerability in Brainstorm Force SureForms sureforms allows Exploiting Incorrectly Configured Access Control Security Levels.This issue affects SureForms: from n/a through 2.13.0.

WordPress Element Pack Elementor Addons plugin <= 10.0.4 - Cross Site Scripting (XSS) vulnerability

Improper Neutralization of Input During Web Page Generation ('Cross-site Scripting') vulnerability in BdThemes Element Pack Elementor Addons bdthemes-element-pack-lite allows Stored XSS.This issue affects Element Pack Elementor Addons: from n/a through 10.0.4.

Apache XML Graphics Batik: Cross-Site Scripting via Unencoded Font Metadata in Apache Batik SVGFont

A Cross-Site Scripting via Unencoded Font Metadata in Apache Batik SVGFont. This issue affects Apache XML Graphics Batik: through 1.19. Users are recommended to upgrade to version 1.20, which fixes the issue.

Authorization bypass in the Git Sync provisioning jobs API

A user with the Editor role could delete, move, push or pull Git Sync managed resources in folders where they only had view access, by creating a provisioning job through the jobs resource API, which skipped the folder permission checks (incorrect authorization).

Team preferences RBAC bypass through the set-home-dash endpoint

Any authenticated user, including a Viewer, could change the preferences of any team, such as its home dashboard, theme and timezone, through an undocumented endpoint with no access control, and so redirect every member of that team to a dashboard of their choice (missing authorization).

proxygen's QPACK decoder enforced its size limit on a string literal in an encoder-stream instruction only after the entire literal had been received. A remote, unauthenticated HTTP/3 peer could send an insert instruction on the QPACK encoder stream that declares an arbitrarily large literal and then keep sending data, causing the decoder to buffer it without bound for the lifetime of the connection and exhaust memory, leading to denial of service. The issue occurs regardless of the advertised QPACK dynamic table capacity, including when the dynamic table is disabled. This issue affects proxygen versions prior to v2026.11.09.00.

In proxygen's HTTP/3 session, when a HEADERS block referenced QPACK dynamic table entries that had not yet been received, header decoding for that stream was paused, but the session continued to read the stream from the QUIC transport. The unread data bypassed the session's ingress buffer limit, returned flow-control credit to the peer, and was re-scanned in full on every read. With QPACK blocked streams enabled (the default), a remote, unauthenticated HTTP/3 peer could send a HEADERS block that references a dynamic table insert it never sends, followed by a continuous stream of data, causing memory growth until the request timed out and CPU usage that grows quadratically, leading to denial of service. This issue affects proxygen versions prior to v2026.11.09.00.