Initialization of a resource with an insecure default in GitHub Copilot and Visual Studio Code allows an unauthorized attacker to disclose information over a network.

Execution with unnecessary privileges in Azure Synapse allows an authorized attacker to elevate privileges over a network.

Improper neutralization of special elements used in a command ('command injection') in Microsoft Copilot allows an unauthorized attacker to perform tampering over a network.

Improper authentication in Azure Active Directory allows an unauthorized attacker to elevate privileges over a network.

Improper neutralization of input during web page generation ('cross-site scripting') in Microsoft Entra ID allows an authorized attacker to perform spoofing over a network.

Quarkus is a Java framework for building cloud-native applications. Prior to versions 3.37.0, 3.36.3, 3.33.2.1, 3.33.3, 3.27.4.1, 3.27.5, and 3.20.6.2, Quarkus HTTP path-based authorization policies can be bypassed using encoded semicolons (%3B) to smuggle matrix parameters past the security layer, and using encoded slashes (%2F) or backslashes (%5C) to access protected static resources. This is a distinct issue from CVE-2026-39852, which addressed only literal semicolon stripping. Versions 3.37.0, 3.36.3, 3.33.2.1, 3.33.3, 3.27.4.1, 3.27.5, and 3.20.6.2 contain a patch.

Authelia is an open-source authentication and authorization server providing two-factor authentication and single sign-on (SSO) for applications via a web portal. In versions 4.36.0 through 4.39.19, due to lack of canonicalization of domains in very specific edge cases, an access control rule may be skipped when it should match a request. The specific conditions that could lead to a security issue for vulnerability are: 1. The specific target resource of the attack must be using the forwarded authorization integration; 2. The requested domain must have two additional segments compared to a session domain i.e. `a.b.example.com` is requested, but the session domain is `example.com`; 3. There access control rules must specify two separate rules which both contain inexact domain matches such as `*.b.example.com` and `*.example.com` i.e. wildcards, username matches, group matches; 4. The rules must be in order of most specific domain to least specific domain; 5. The second rule must be more permissive than the first rule; 6. The attacker must specifically request a URL for the more specific domain, with the second part containing one or more capitalized letters i.e. `https://a.B.example.com` and no other segment with capitalized letters; 7. The integration used must not be the Envoy ExtAuthz integration; and 8. The proxy must not canonicalize the requested host name in the relevant header before sending it to the relevant authorization endpoint. The kind of configuration used to produce this issue and result in a `bypass` rule being matched has long been highly discouraged. Essentially hosts which should be bypassed entirely should not be secured by having the proxy check them with the authorization handlers. Upgrade to 4.39.20 to receive a patch.

Authelia is an open-source authentication and authorization server providing two-factor authentication and single sign-on (SSO) for applications via a web portal. In versions 4.38.0 through 4.39.19, when a user authenticates via Basic Auth (i.e via the `Authorization` header with the `Basic` scheme) on the authz verification endpoint, Authelia takes the username directly from the `Authorization` header and passes it as is to the regulation system for ban checking and attempt recording. LDAP treats usernames case insensitively : `john`, `John`, and `JOHN` all bind as the same user. But the regulation SQL queries treat the lookup of these values in certain scenarios as case sensitive. This allows each variation of a usernames case to have its own ban bucket. Upgrade to 4.39.20 to receive a patch. As a workaround, explicitly disable the basic auth mechanism.

Kestra is an open-source, event-driven orchestration platform. Prior to versions 1.3.19, 1.2.19, 1.1.19, and 1.0.43, Kestra task `inputFiles` writes rendered file names directly under the task working directory. When a flow forwards untrusted execution or webhook data into an `inputFiles` file name, a caller can use `../` path segments to create or overwrite files outside that task working directory on the worker filesystem. Versions 1.3.19, 1.2.19, 1.1.19, and 1.0.43 patch the issue.

libde265 is an open source implementation of the h.265 video codec. Prior to version 1.1.0, a crafted H.265 bitstream with large SPS dimensions and 16-bit bit depth causes a signed integer overflow in `de265_image_get_buffer()` (`libde265/image.cc:128`). The overflow wraps the plane allocation size to a small value (~1 KB), but the subsequent `fill_image()` call computes the real size using `size_t`, writing ~4 GB into the undersized heap buffer. Version 1.1.0 patches the issue.

libde265 is an open source implementation of the h.265 video codec. Prior to version 1.0.20, a crafted H.265 bitstream can cause an out-of-bounds array write in `decoder_context::process_reference_picture_set()` (`libde265/decctx.cc:1376`). The root cause is a missing aggregate bound check on predicted short-term reference picture set entries. Individual list sizes are validated, but the combined count after predicted RPS construction can exceed the 16-entry `PocStFoll` array, writing at index 16. Version 1.0.20 patches the issue.

libde265 is an open source implementation of the h.265 video codec. Prior to version 1.0.20, a crafted sequence of H.265 NAL units causes `decoder_context::read_slice_NAL()` (`libde265/decctx.cc:481`) to attach slice headers to a finished picture object that has no active image unit, resulting in attacker-controlled unbounded heap growth. The retained headers are never freed until the picture is released, which may not happen during continuous streaming. Version 1.0.20 patches the issue.

gin-vue-admin is an AI-assisted basic development platform. In version 2.9.1, an authenticated attacker with access to the code-generation feature and MCP management interface can exploit this vulnerability by injecting attacker-controlled Go source code through POST /autoCode/addFunc, and then invoking POST /autoCode/mcpStart to trigger a rebuild and restart of the standalone MCP service. This allows arbitrary operating system commands to be executed on the server with the privileges of the application process. Successful exploitation may lead to remote code execution (RCE), modification of backend source code or runtime logic, deployment of persistent backdoors, access to or manipulation of application data and configuration, and further impact on local resources running under the same service account or privilege context. The risk is highest in deployments that retain the source tree, allow writes to source files, and support local build or startup of standalone MCP components. In environments using binary-only releases, read-only filesystems, or with local build capabilities removed, the exploitability of the full attack chain is significantly reduced. However, once the online code-generation capability and MCP-hosted startup workflow are enabled, the overall security impact may reach high to critical severity. As of time of publication, it is unknown if a patched version is available. As a workaround, enforce strict allowlist validation on path- and identifier-related fields such as `humpPackageName`, `packageName`, `FuncName`, and `Router`, and only permit safe identifier formats.

DevGuard provides vulnerability management for the full software supply chain. Prior to 1.4.2, on a DevGuard API instance with one or more public assets, any authenticated user — including users from a different organization with no membership or role in the affected org/project — can create, update, reapply, and delete VEX rules on those public assets. The same flaw affects the other vulnerability-triage write endpoints exposed under a public asset, including VEX rule create / update / reapply / delete; dependency-vuln event creation (accept / reject / mitigate decisions), batch event creation, vuln sync, and mitigation; license risk creation; external reference writes; and/or artifact creation and license refresh. The attacker needs a valid account on the instance, but no membership in the victim organization, project, or asset is required. Version `v1.4.2`contains a patch. As a workaround, make affected assets non-public. In the asset settings, switch visibility from public to private. This removes the public-read exemption in the access-control middleware and restores correct authorization on all write endpoints for that asset. Downstream consumers that previously relied on the public `vex.json` / `sbom.json` endpoints will need to be granted explicit access or must receive an exported file version until the patched release is deployed.

ProxySQL is a proxy for MySQL and its forks, as well as PostgreSQL. In versions 3.0.0 through 3.0.8, ProxySQL's GenAI/MCP `run_sql_readonly` tool violates its documented read-only contract for MySQL targets. The tool validates only the full input string with a substring blacklist and first-keyword allowlist, but then executes the entire SQL string on a backend connection created with `CLIENT_MULTI_STATEMENTS`. As a result, a caller can submit a read-only first statement followed by a side-effecting second statement, such as `SELECT 1; RENAME TABLE ...`. The validator accepts the payload because it starts with `SELECT` and because side-effecting MySQL statements such as `RENAME TABLE`, `SET`, `RESET`, `LOCK TABLES`, and `KILL` are not rejected by the blacklist. In a live MCP runtime test, the `/mcp/query` endpoint accepted a `run_sql_readonly` request. The MCP response reported success for the first `SELECT`, and direct backend verification showed that the table had actually been renamed. This violates the endpoint's read-only security contract and lets an MCP caller perform backend writes or administrative SQL, limited by the configured MCP target account's database privileges. Version 3.0.9 contains a fix. Other operator mitigations include: keeping MCP disabled unless required; setting a non-empty `mcp-query_endpoint_auth` token before exposing `/mcp/query`; restricting MCP listener network exposure; configuring MCP backend target credentials as database-level read-only users; and adding temporary MCP query rules to block obvious multi-statement patterns.

ProxySQL is a proxy for MySQL and its forks, as well as PostgreSQL. In versions 2.0.0 through 3.0.8, the ProxySQL MySQL frontend accepts the `PROXY UNKNOWN <addr> <addr> <port> <port>\r\n` PP1 frame as a well-formed PROXY protocol header. The HAProxy PROXY protocol v1 specification says that when the protocol token is `UNKNOWN`, the receiver MUST ignore any address fields that follow it, because the proxy has declared it cannot determine the client identity. ProxySQL parses those address fields anyway via `sscanf` and writes the spoofed source address into the session's `addr.addr` field. From there it flows directly into the query-rule matcher, where the `client_addr` predicate decides routing and ACL. When `mysql-proxy_protocol_networks = '*'` (the default), any TCP peer can send a PP1 frame and choose any source IP claim. With that, any `mysql_query_rules` row pinned to a `client_addr` value is forgeable: the attacker writes the address they want to match into the PP1 line, and ProxySQL routes their query as if it came from that address. In practice this is a routing and ACL bypass. Real deployments use `client_addr` for read-write splitting (internal apps go to the primary, public traffic to read replicas), per-app schema pinning, and query-filter rules (DDL allowed only from admin CIDR, public queries blocked from dangerous patterns). An attacker that can reach the frontend port can forge their way into any of those routes. Version 3.0.9 patches this issue.

ProxySQL is a proxy for MySQL and its forks, as well as PostgreSQL. Versions 2.0.18 through 3.0.8 have a pre-authentication heap memory corruption vulnerability in the MySQL and PostgreSQL protocol first-read paths. A remote unauthenticated client can declare an oversized first packet length, and ProxySQL passes that attacker-controlled length directly to `recv()` while writing into a fixed 32 KB input queue. Version 3.0.9 patches the issue.

Mercator is an open source web application that enables mapping of the information system. Prior to version 2025.05.19, a Server-Side Request Forgery (SSRF) vulnerability exists in Mercator's CVE configuration panel (`/admin/config/parameters`). The `testProvider()` method in `ConfigurationController` passes user-supplied input directly to `curl_init()` without validating the scheme, hostname, or destination IP address. An authenticated user with the `configure` permission can force the Mercator server to issue arbitrary outbound network requests. The suffix `/api/dbInfo` appended to the URL can be bypassed by injecting a `#` fragment character (e.g. `http://TARGET/PATH#`), allowing full control over the target URL. No scheme whitelist, host whitelist, or private/loopback IP block is applied. The `telnet://` scheme can be used for internal port scanning; the `gopher://` scheme enables interaction with unauthenticated internal services (Redis, Memcached), potentially leading to Remote Code Execution under specific deployment conditions. Version 2025.05.19 patches the issue.

Mercator is an open source web application that enables mapping of the information system. Prior to version 2025.05.19, Mercator's Query Engine (`/admin/queries/execute`) accepts a JSON DSL (`from` / `select` / `filters` / `traverse` / `output`), translates it into an Eloquent query, and returns results as JSON. The controller method `QueryController::execute()` does not enforce an authorization gate, unlike `store()` and `massDestroy()` in the same controller which are correctly protected. As a result, any authenticated account — including the read-only Auditor role — can query models beyond its intended scope, including the `User` model. Additionally, the `password` column, although declared `$hidden`, is not excluded from filter predicates, which allows it to be used in `LIKE` conditions. The `schema()` and `schemaModel()` endpoints of the same controller are similarly unguarded. The Query Engine is read-only; integrity and availability are not affected. Version 2025.05.19 patches the issue.

radvd is a router advertisement daemon for IPv6. Prior to version 2.21, the `radvdump` utility shipped with radvd contains a stack buffer overflow in the Route Information option parser. When processing a crafted ICMPv6 Router Advertisement, `print_ff()` copies up to 2032 bytes from attacker-controlled packet data into a 16-byte `struct in6_addr` on the stack, overflowing by up to 2016 bytes. Note that the main `radvd` daemon is not affected by the vulnerability. Version 2.21 patches the issue.

YARD is a documentation generation tool for the Ruby programming language. Prior to version 0.9.44, YARD's static cache lookup reads a request path before the router's path cleanup runs. When a server is configured with a document root, a traversal path such as `/../yard-cache-secret.html` is joined against that root and can return a readable sibling `.html` file outside the intended static tree. Version 0.9.44 patches the issue.

gonic is a music streaming server / free-software subsonic server API implementation. Prior to version 0.21.0, a logic error in `ServeCreateOrUpdatePlaylist` allows any authenticated Subsonic user (including non-admin) to write playlist M3U content to an attacker-controlled absolute filesystem path on the gonic host, and to create intermediate directories with `0o777` permissions. The bug is independent of CVE-2026-49338 and CVE-2026-49339. It is an unreachable guard clause combined with no path containment in `Store.Write`. Version 0.21.0 patches the issue.

gonic is a music streaming server / free-software subsonic server API implementation. Prior to version 0.21.0, the Subsonic API endpoints `/rest/deletePlaylist.view` and `/rest/getPlaylist.view` perform no per-resource authorization. Once authenticated as any user (admin or not), an attacker can delete any playlist owned by any other user (including admin) by passing its `id` and read the full contents (name, comment, song list) of any other user's **private** (non-public) playlist by passing its `id`. The Subsonic playlist `id` is `base64url("<userID>/<filename>.m3u")`. Because filenames are user-supplied or time-derived and the `userID` is a small integer, IDs are guessable and frequently exposed (e.g. a previously-public playlist that was later made private still has the same ID). This breaks the multi-user trust boundary of gonic: a low-privileged user can wipe an administrator's curated playlists, and a user can exfiltrate any private playlist they obtain an ID for. The issue was fixed in commit `6dd71e6a3c966867ef8c900d359a7df75789f410`, which is part of version 0.21.0.

A TraceQL query in Grafana Tempo with a large exemplars hint value can cause the Tempo instance to allocate an excessive amount of memory, resulting in an out-of-memory crash. This could allow an authenticated user to trigger a denial of service against the Tempo service.

A flaw was found in the AWX GitHub webhook integration. When processing GitHub pull_request webhooks, the controller stores the pull_request.statuses_url value from the webhook payload without validating that it points to a trusted GitHub API endpoint. If a job template is configured with a GitHub Personal Access Token as its webhook credential, the controller later POSTs that token to the stored callback URL when posting job status updates. An attacker who can submit a correctly signed forged webhook using the job template's webhook_key can redirect the callback to an attacker-controlled URL and exfiltrate the configured GitHub PAT.