The md-to-slides package on PyPI contains malicious code hidden in a specific dependency (is-it-ai) in version 0.1.1. This malicious code fingerprints the environment and contacts a remote server to download and execute a remote executable if instructed. Other versions appear to be decoys to conceal the malicious version. This campaign is clearly malicious and targets users by embedding harmful actions within a library dependency.
The skills-analyzer package on PyPI contains malicious code in its dependency is-it-ai version 0.1.1. This malicious code fingerprints the environment and contacts a remote server, which may respond with a URL to download and execute further malicious payloads. Other versions of the package exist to obscure the presence of the malicious version. The campaign is identified as a targeted attack with clear malicious intent, including downloading and executing remote executables.
The is-it-ai package on PyPI contains malicious code in specific versions (0.1.0, 0.1.1, 0.1.3) that, when triggered, fingerprints the environment and contacts a remote server. If the server responds with a URL, the package downloads and executes a remote executable. This behavior is part of a targeted attack campaign involving intentional inclusion of malicious code in a dependency, designed to hide malicious actions behind legitimate package versions.
CVE-2026-93522 is a heap buffer overflow vulnerability in the X.Org Server Glamor CopyArea component affecting version 3.0. The vulnerability is identified as CWE-122. No CVSS score or detailed impact information is provided. There is no indication of known exploits in the wild or available patches.
pgAdmin 4 versions before 9.19 contain a critical vulnerability in the Import/Export Data tool. The tool wraps user queries in a psql \copy (...) TO command and attempts to validate balanced parentheses to prevent command injection. However, it incorrectly handles backslashes inside double-quoted identifiers, allowing an attacker to bypass this check and inject a TO PROGRAM clause. This clause executes arbitrary commands on the pgAdmin host with the operating system privileges of the pgAdmin process. The vulnerability arises because the server setting standard_conforming_strings can be turned off by ordinary database roles, affecting how backslashes are interpreted, which pgAdmin does not reliably account for. This issue is a bypass of the previous fix for CVE-2026-17566.
CVE-2026-119553 is a path traversal vulnerability in Moodle affecting versions up to 5.3.0. It exists in the Blackboard question import format, allowing teachers with import or lesson edit rights to upload crafted zip files with image sources containing '../' sequences. This can lead to unauthorized copying of server files such as config.php into their draft area.
An unauthenticated remote denial of service vulnerability exists in the Twisted HTTP/2 server (CVE-2023-44487). A remote client can send a sequence of HTTP/2 frames that cause the server to render requests fully before processing stream resets, allowing CPU exhaustion via rapid stream resets on a single connection. This bypasses the concurrent-stream limit and leads to unbounded request processing. The vulnerability affects Twisted web HTTP/2 server version 26.4.0 and later. There is no impact on confidentiality or integrity, and no code execution is possible.
The Model Context Protocol (MCP) Python SDK client versions 1.10.0 through 1.29.1 and 2.0.0 through 2.1.1 have a vulnerability where the client fetches server-chosen $ref URLs during tool output schema validation. This causes the client to synchronously fetch external HTTP, HTTPS, or local file references without timeout, potentially freezing the client session. The issue affects applications using the SDK as a client calling tools on untrusted servers. Fixed in versions 1.30.0 and 2.2.0.
The Model Context Protocol (MCP) Python SDK versions 1.10.0 through 1.29.1 and 2.0.0 through 2.1.1 have a vulnerability in server-side bearer-token authentication. The SDK does not verify that a token's audience matches the intended server, allowing tokens issued for other services by the same authorization server to be accepted. This can lead to unauthorized access if tokens for different services are accepted. Fixed in versions 1.30.0+ and 2.2.0+, but enabling the audience check requires configuration changes. If upgrading is not possible, manual audience verification is recommended.
LangGraph Python SDK v3 streaming client versions 0.4.0 through 0.4.5 improperly handle unencoded thread_id and assistant_id values in request paths, allowing relative path segments to redirect authenticated GET requests to other paths on the same server. This affects both synchronous and asynchronous call sites in the get_tree and _fetch_state functions. The LangGraph Platform itself is not affected. Applications passing attacker-controlled identifiers may have responses redirected to attacker-chosen readable paths, but no write operations are possible. Upgrading langgraph-sdk to 0.4.6 or later and langgraph to 1.2.14 or later mitigates the issue.
A vulnerability in Python's HTTP response and chunk parsing logic causes incorrect processing of non-hexadecimal or negative integers. This flaw can lead to misinterpretation of HTTP response content, potentially causing data integrity issues. Exploitation requires a client application to connect to a malicious or untrusted web server. The vulnerability does not allow code execution or privilege escalation and is rated moderate severity by Red Hat. The issue affects Red Hat Hardened Images RPMs with Python 3.11 packages. A fixed update is available in python3.11 version 3.11.18-1.1.hum1.
A regular expression inefficiency in OpenTelemetry Python's TraceState.from_header method causes quadratic CPU time when parsing specially crafted tracestate headers with long runs of spaces or tabs. This can lead to denial of service by blocking server workers for several seconds on a single unauthenticated request. The issue affects versions from 1.0.0 through 5.44.9 and is fixed in 5.45.0.