A vulnerability was determined in iFlytek astron-agent up to 1.1.2. This affects an unknown function of the file core/memory/database/api/v1/exec_ddl.py of the component DDL Whitelist Handler. This manipulation causes incomplete blacklist. The attack may be initiated remotely. The exploit has been publicly disclosed and may be utilized. The vendor was contacted early about this disclosure but did not respond in any way.
A flaw in OpenAI's context compaction process allowed researchers to transfer reasoning traces between AI models, enabling the cheaper distillation of rival AI models. This technique, called compaction smuggling, bypassed initial mitigations and was fixed by OpenAI in October 2026. The vulnerability enabled extraction of proprietary reasoning data from advanced models and its reuse in weaker models, facilitating large-scale copying of AI capabilities. This has been exploited by Chinese entities to develop competitive AI systems, raising national security concerns.
Mattermost versions 11.7.x <= 11.7.10 fail to apply outbound internal-connection filtering to the OpenID discovery endpoint request which allows a System Administrator to determine whether internal addresses and ports are reachable from the server via an unauthenticated login request that triggers the discovery fetch.. Mattermost Advisory ID: MMSA-2026-00774
Mattermost versions 11.9.x <= 11.9.1, 11.8.x <= 11.8.5, 11.7.x <= 11.7.10, 11.10.x <= 11.10.1 fail to bind the volume of text extracted from an uploaded PDF which allows an authenticated user to exhaust server memory and cause a denial of service via a crafted PDF whose compressed content stream expands into a very large amount of text.. Mattermost Advisory ID: MMSA-2026-00769
Mattermost versions 11.9.x <= 11.9.1, 11.8.x <= 11.8.5, 11.7.x <= 11.7.10, 11.10.x <= 11.10.1 Mattermost fail to apply allocation limits to the byte-string decode paths in its WebSocket msgpack decoder, which allows an authenticated remote attacker to cause excessive server memory allocation via a crafted binary WebSocket frame declaring a Bin32 payload length of up to 4,294,967,295 bytes.. Mattermost Advisory ID: MMSA-2026-00768
Path Traversal and Missing Authorization in cloud-storage source in Google Cloud MCP Toolbox for Databases (googleapis/mcp-toolbox) versions 1.2.0 through 1.4.0 allows an unauthenticated remote attacker to exfiltrate host files to external Google Cloud Storage buckets and overwrite arbitrary local files via crafted tool invocation parameters.
Vulnerability in the SQL storage module of Apache SIS. When sorting a stream of features, the clause is copied after ORDER BY without quoting the identifier character. A crafter sort clause could inject SQL in the query.
This issue affects the sql-storage module of Apache SIS from version 1.0 through 1.6 inclusive. Other modules such as sis-referencing are not impacted. Users are recommended to upgrade to version 1.7, which fixes the issue.
CVE-2026-93847 is an authorization vulnerability in Mattermost versions 11.7.0 through 11.7.10. It occurs because these versions fail to properly validate reserved system post types during scheduled post creation and publication. This flaw allows an authenticated channel member to publish messages that appear as trusted system events via the scheduled post endpoint.
The root-privileged `wsserver` terminal-bridge daemon builds a shell command from client-supplied WebSocket fields and passes it to `system()` and `popen()` without quoting, escaping, or character-set validation. The daemon's public-key session path performs no appliance-session check.
Security researchers identified a sophisticated supply chain attack campaign involving 42 compromised GitHub repositories across 11 legitimate owners connected to the testmetest npm operation. The campaign evolved to use compiled native addons that decrypt hidden payloads at runtime, evading detection in CI environments through extensive anti-analysis checks. Attackers deployed polymorphic wrappers ranging from 12.9 to 13.7 MB that decode to identical publisher payloads, embedding execution hooks for AI coding assistants and IDEs. The operation included credential theft, repository hijacking, forged commit timelines, and automated malicious package publishing through GitHub Actions using npm trusted publishing. The infrastructure also links to shellpanel exploitation activities targeting cPanel systems. The campaign demonstrates advanced evasion techniques including environment detection for CI systems, containers, sandboxes, debuggers, and security tools before executing malicious code on developer workst...
Adversaries are exploiting Microsoft Teams' default external chat functionality to conduct social engineering attacks and deliver malicious payloads. Analysis of over 400 incidents reveals threat actors consistently impersonate IT support personnel using throwaway Microsoft 365 tenants with IT-themed display names marked as "(Internal)" to bypass external-user warnings. Attacks frequently begin with email bombing followed by Teams contact offering assistance. Adversaries leverage legitimate remote monitoring tools like Quick Assist, AnyDesk, and ScreenConnect to gain hands-on-keyboard access, progressing from initial contact to payload execution in under 15 minutes. Infrastructure patterns show a shift from default onmicrosoft domains toward paid top-level domains, particularly .top domains comprising 50 percent of observed cases. The technique exploits legitimate collaboration features rather than software vulnerabilities, requiring organizations to implement allow-list policies and user training to disru...
CVE-2026-115930 is a high-severity vulnerability in MongoDB Shell (mongosh) versions prior to 2.14.0. It involves improper certificate validation due to the inclusion of recoverable eMudhra intermediate CA certificates in the Node.js TLS root certificate store used by mongosh. This can be mitigated by updating mongosh to version 2.14.0 or later, which bundles a fixed Node.js runtime version 24.22.1 or newer. If mongosh is installed via npm, the Node.js runtime on the host must be updated separately.
Kameleo versions 5.3.0 and 5.3.1 contain an information disclosure vulnerability in the Local API component, specifically in the /general/user-info endpoint. This flaw allows remote attackers to obtain sensitive information without authentication. The vendor has not responded to disclosure attempts and no patch or fix is currently available.
Mattermost versions <=11.9 11.1.9 11.5.8 11.10.7.0 fail to validate Zoom chat line length before slicing which allows an authenticated attacker to crash the Agents plugin via summarizing a Zoom chat transcription containing short continuation lines. Mattermost Advisory ID: MMSA-2026-00723
The sudo packages contain the sudo utility which allows system administrators to provide certain users with the permission to execute privileged commands, which are used for system management purposes, without having to log in as root.
Security Fix(es):
* sudo: sudo: TZ environment variable allows bypass of NOTBEFORE/NOTAFTER time-based authorization (CVE-2026-96512)
For more details about the security issue(s), including the impact, a CVSS score, acknowledgments, and other related information, refer to the CVE page(s) listed in the References section.
Allocation of Resources Without Limits or Throttling vulnerability in hackney_conn in benoitc hackney allows a malicious HTTP server to exhaust client memory via an oversized or highly compressed response body. hackney:request/5 and hackney:body/1 return the whole response body, and nothing bounds how much the connection process buffers. Any server the client requests can trigger it.
hackney_conn:read_full_body/2 and the HTTP/2 and HTTP/3 accumulators append every chunk until the body ends. With {auto_decompress, true} the buffered body is then inflated in one step with zlib:gunzip/1 or zlib:inflate/2, which return the whole output at once.
This issue affects hackney: from 2.0.0-beta.1 before 4.9.0.
CVE-2026-115693 is an improper certificate validation vulnerability in the hackney_ssl module of benoitc hackney. It allows an on-path attacker to impersonate a server pinned to a private certificate authority by exploiting the acceptance of certificate chains ending in a publicly trusted root. The hostname check remains enforced, requiring the attacker to have a publicly trusted certificate for the pinned hostname. This affects hackney versions from 1.24.0 up to but not including 4.9.0.
CVE-2026-113440 is a vulnerability in the benoitc hackney HTTP client library versions from 2.0.0 up to but not including 4.9.0. It allows a redirecting server to capture basic authentication and cookie credentials when a redirect occurs on the same host but changes scheme (e.g., HTTPS to HTTP) or port. This happens because the library reissues requests with credentials after comparing only the hostname, exposing sensitive headers in cleartext or to unintended services.
CVE-2026-115700 is a Server Side Request Forgery (SSRF) vulnerability in the benoitc hackney library affecting versions from 4.0.1 up to but not including 4.9.0. The flaw arises from an interpretation conflict when handling percent-encoded bracketed IPv6 literals or IP literals with trailing dots, allowing bypass of allowlist checks and unintended connections to internal IP addresses. This issue was introduced as a fix attempt for a previous vulnerability (CVE-2026-47076) but still allows certain malformed IP literals to bypass validation. The vulnerability has a medium severity with a CVSS 4.0 score of 6.9 and is fixed in version 4.9.0.
Improper Neutralization of CRLF Sequences in HTTP Headers ('HTTP Request/Response Splitting') vulnerability in hackney_http_connect in benoitc hackney allows HTTP Request Splitting via the value a proxy_auth_fun callback returns. After a 407 response from an HTTP CONNECT proxy, hackney_http_connect:handle_proxy_auth/11 calls the caller's proxy_auth_fun with the status code and the response headers and writes the returned binary into the Proxy-Authorization line of the CONNECT retry without checking it. A CR, LF or NUL byte in that value adds header lines to the request sent to the proxy.
The value is attacker-influenced when the callback derives it from untrusted data, for example from the proxy's challenge headers. Other caller-supplied values on the CONNECT path, such as the target host and the request headers, were already checked for control bytes.
This issue affects hackney: from 2.0.0 before 4.9.0.