Eclipse Zenoh's zenoh-pico implementation, versions 0.7.2-rc up to but excluding 1.10.2, does not validate the Zenoh identifier (ZID) length when deserializing timestamps. An attacker-controlled length greater than 16 causes an out-of-bounds write to a fixed 16-byte identifier buffer and causes the subsequent zero-fill operation to receive a negative length converted to a large unsigned size. Truncated timestamp inputs can also cause out-of-bounds reads or assertion failures. A remote attacker able to deliver crafted timestamp-bearing Zenoh messages through an accepted transport can cause memory corruption and denial of service, with potential remote code execution in builds lacking appropriate binary protections. This issue is fixed in version 1.10.2.
Depending on the release and message-decoding path, the unchecked timestamp identifier length causes a heap-based or stack-based buffer overflow. Ordinary-frame network message decoding moved from heap-backed storage to stack-local storage in version 1.5.0. Earlier affected releases already contained stack-based decoding paths, including reassembled fragmented messages.
SQLBot 1.10.0 through 1.10.3 fails to apply its PostgreSQL dangerous-function denylist because DS_SPECIFIC_DANGEROUS_FUNCTIONS uses 'postgresql' while lookups use the 'pg' type. Authenticated users can steer LLM-generated SQL past check_sql_read() using pg_read_file, pg_write_file, lo_import or lo_export to read or write files on the database host.
CVE-2026-21609 is a high severity improper authorization vulnerability affecting Jira Service Management Data Center versions 5.12.0, 10.3.0, and 11.3.0. It allows an authenticated attacker to gain unintended access, potentially exposing resources or functionality, which may lead to sensitive information disclosure or arbitrary code execution. Atlassian recommends upgrading to fixed versions 10.3.27 or later for the 10.3.x line, and 11.3.13 or later for the 11.3.x line.
The lettre library has a vulnerability where a mailbox display name containing carriage return (CR) or line feed (LF) characters causes a panic when formatted or included in email headers like From or To. This can lead to denial of service by crashing the thread or task building the message. The issue affects versions from 0.10.0-rc.5 up to but not including 0.11.24. Applications that only use trusted display names are not exposed. A fix is available in version 0.11.24.
A regression in FIPS enforcement for hash algorithms in Python's hashlib module affects only Python version 3.14.10 when used with OpenSSL in FIPS mode. This issue does not impact other Python versions. A fix is available in Python 3.14.11.
ProxySQL versions prior to 3.0.10, 3.1.10, and 4.0.10 contain a vulnerability in the MySQL compressed packet decompression function. An authenticated MySQL client negotiating compression can manipulate the payload length, causing ProxySQL to read uninitialized heap memory and forward stale data to the backend. This can lead to disclosure of prior SQL queries, result data, credentials, or session state, and may also cause excessive memory allocations. The issue is fixed in versions 3.0.10, 3.1.10, and 4.0.10.
Parse Server versions from 9.7.0 up to but not including 9.10.5 contain an authorization flaw involving installation update misdirection via the X-Parse-Installation-Id header. This flaw allows attackers who know another installation's object ID to trigger protected field processing on that installation while writing data elsewhere, when the protectedFieldsTriggerExempt feature is enabled. The issue is fixed in version 9.10.5.
Parse Server versions from 9.7.0 up to but not including 9.10.5 have an information disclosure vulnerability. This flaw allows clients with create or update permissions to read protected field values written by beforeSave triggers when the protectedFieldsSaveResponseExempt setting is false. The issue arises because clients can access these protected fields using dot notation keys in the same request. The vulnerability is fixed in version 9.10.5.
Parse Server versions before 8.6.108 and from 9.0.0 up to but not including 9.10.5 contain a prototype pollution vulnerability. This flaw allows clients with class write access to send specially crafted keys that corrupt built-in object functions, leading to denial of service by causing all requests to fail until the server is restarted.
Parse Server versions before 8.6.106 and versions 9.0.0 up to but not including 9.10.5 mishandle dot notation keys with null or undefined values during object saving. This causes writes to be stored but returns an error that skips afterSave triggers and LiveQuery events. Exploiting this flaw allows a client with write access to bypass audit logging, notifications, and derived data updates.
Parse Server versions before 8.6.107 and from 9.0.0 up to but not including 9.10.5 have an information disclosure vulnerability. This occurs because the server returns the stored value of a field reverted by a beforeSave trigger in object update responses, even to clients without read access to that object or field. An authenticated client able to update such an object can read protected values, including its own account's raw MFA authData when the protectedFieldsTriggerExempt feature is enabled.
Parse Server versions before 8.6.105 and from 9.0.0 up to but not including 9.10.5 contain an ACL bypass vulnerability in $relatedTo relation queries. The server does not properly validate the owning object reference, allowing a client with read access to some objects to infer linked objects it should not have access to. This can disclose relationships between readable and non-readable objects.
MLflow versions from 3.10.0 up to but not including 3.17.0 contain a vulnerability that allows unauthenticated remote attackers to read arbitrary files on the server. This occurs because the prompt branch in the _create_model_version() function accepts schemeless relative sources without proper path validation. The flaw enables attackers to specify server-local paths as model version artifact locations and retrieve files via the /api/2.0/mlflow/model-versions/get-artifact endpoint. Default unauthenticated servers are fully exposed, while basic-auth deployments are vulnerable to authenticated users with model creation privileges. The issue is fixed in version 3.17.0.
MariaDB server versions from 10.11.1 up to but not including 10.11.19, 11.4.1 up to 11.4.13, 11.8.1 up to 11.8.9, 12.3.1 up to 12.3.3, and 13.0.1 up to 13.0.2 contain a vulnerability in the ALTER TABLE ... CONVERT ... PARTITION command. This flaw allows an authenticated user to bypass authorization checks by specifying partition names that require filename escaping, leading to unauthorized access or modification of table files. The issue is fixed in versions 10.11.19, 11.4.13, 11.8.9, 12.3.3, and 13.0.2.
AIOHTTP versions from 3.4.2 up to but not including 3.10.0 have a vulnerability in the WebSocket reader where the receive queue limits can be bypassed by small or compressed messages. This occurs because the reader continues parsing frames already received after the receive queue has paused the transport, and empty messages are not counted towards the queue size. When permessage-deflate compression is used, a single read of compressed frames can add many megabytes of decompressed messages to the queue. This can lead to excessive memory consumption on the server or client side.
AIOHTTP versions from 3.4.2 up to but not including 3.10.0 have a vulnerability where decompressing concatenated compressed streams (gzip, deflate, or zstd) causes excessive CPU usage due to inefficient decompression logic. This can allow an attacker to block the event loop of a server or client by sending specially crafted compressed payloads.
A vulnerability in AIOHTTP's pure-Python HTTP parser causes the request buffer to grow excessively when an upgrade request is declined, leading to repeated processing of pipelined requests and excessive memory use. This affects versions from 3.4.2 up to but not including 3.10.0. The issue occurs only when using the pure-Python parser, such as when the default C parser is disabled. A fix is available in version 3.10.0.
AIOHTTP versions from 3.4.2 up to but not including 3.10.0 have a vulnerability where data received during an HTTP upgrade request (such as a WebSocket handshake) is buffered without limit until the upgrade handler completes. This can lead to unbounded memory consumption if an attacker streams large amounts of data behind the upgrade request.