A vulnerability in pypdf prior to version 6.21.3 allows an attacker to craft a PDF file with cross-reference streams that causes large memory usage when parsed. This issue has been fixed in pypdf version 6.21.3. Users unable to upgrade immediately can apply the changes from the referenced pull request as a workaround.
pyasn1 versions up to 0.6.4 have a vulnerability in decoding constructed BIT STRING values where CPU consumption grows quadratically with input size. This occurs because the decoder repeatedly shifts and copies accumulated data for each fragment, leading to potential CPU exhaustion and denial of service when processing attacker-controlled ASN.1 data. The issue affects BER and CER decoders and certain schema-driven DER decoding paths. The vulnerability is fixed in version 0.6.5.
pyasn1 versions 0.6.4 and earlier contain a denial of service vulnerability due to quadratic time complexity in decoding binary-encoded ASN.1 REAL values with large mantissas. This issue affects BER, CER, and DER decoders that use the vulnerable RealPayloadDecoder. An attacker supplying crafted ASN.1 data with large REAL mantissas can cause excessive CPU consumption during decoding. The vulnerability is fixed in pyasn1 version 0.6.5.
pyasn1 versions 0.6.4 and earlier contain a denial of service vulnerability due to quadratic-time processing when decoding constructed OCTET STRINGs, constructed character strings, and indefinite-length ANY types. This occurs because the decoder repeatedly concatenates fragments in a way that causes CPU time to grow quadratically with the number of fragments. An attacker supplying crafted ASN.1 input to an affected decoder can cause CPU exhaustion. The issue affects BER and CER decoding paths and certain non-DER constructed strings accepted by the DER decoder with specific schema. No confidentiality or integrity impact is reported.
pyasn1 versions from 0.3.3 up to but not including 0.6.5 have a vulnerability causing quadratic CPU usage when decoding attacker-controlled BER, CER, or DER data without an ASN.1 schema. This occurs due to inefficient handling of tag sets during schemaless decoding, allowing crafted inputs with many distinct long tag IDs to exhaust CPU resources and degrade service availability. No confidentiality or integrity impacts are reported. The issue is fixed in version 0.6.5.
GraphQL-core versions from 3.3.0a3 up to but not including 3.3.1 contain a denial of service vulnerability caused by a memory leak during aborted validations. When a validation is aborted due to exceeding the max_errors limit, a shared ValidationAbortedError instance accumulates traceback frames, retaining memory indefinitely. This allows unauthenticated clients to cause memory exhaustion by sending documents with over 100 validation errors, potentially leading to out-of-memory kills. The issue is fixed in version 3.3.1 by creating a new error instance per aborted validation and clearing tracebacks.
OAuthlib versions prior to 4.0.0 contain a vulnerability in token grant handlers where an authenticated low-privilege client can supply a different client_id in the token request body. This causes grant type and scope validation to be performed under the attacker-supplied client_id rather than the authenticated client's identity. The flaw affects the client_credentials and resource owner password credentials grant flows. It allows a low-privilege client to obtain tokens with broader scopes or grant types assigned to another client. The issue is fixed in version 4.0.0.
OpenWISP Controller contains an authorization flaw in its authenticated REST API endpoint for downloading CA Certificate Revocation Lists (CRLs). The endpoint does not properly scope requests to the managed organization, allowing an organization manager to access CRLs belonging to other organizations by enumerating CA IDs. The exposure is limited to CRL data, which includes CA issuer information and revoked certificate serial numbers, but does not include private keys or allow modification. The flaw affects versions from 1.0.0 up to but not including 1.3.2 and has been fixed in version 1.3.2.
joserfc versions prior to 1.7.6 have a vulnerability in their compact JWS size guard. The library allocates attacker-controlled segments by splitting the input bytes before validating their size, causing an additional large memory allocation before rejecting oversized inputs. This can lead to increased memory usage and potential denial of service due to memory exhaustion. A fix is available in version 1.7.6 that adds size checks before splitting the input.
The joserfc library's decrypt_json() function for JWE General JSON Serialization accepts an unbounded recipients array, performing an RSA private-key decryption for each recipient without limit. This allows an unauthenticated attacker to cause CPU exhaustion (Denial of Service) by sending a single JWE with many recipients. The library caps other JWE structural elements but does not limit recipient count. The documented mitigation option verify_all_recipients=False does not reduce CPU usage, as all recipients are decrypted regardless. This vulnerability is fixed in joserfc version 1.7.5.
The stdid-nyxprobe package on PyPI contains malicious code that exfiltrates basic host information such as IP address and username upon installation or import. The package overrides the install command to execute this code, serving no legitimate purpose beyond data exfiltration. The risk is limited but present due to the unauthorized data collection.
The log-utils-sdk package version 0.4.1 on PyPI contains malicious code that executes upon import or use. This code downloads additional malicious scripts and runs a background process that periodically connects to a remote host to receive further commands. The package is a clone of a legitimate library and is part of a known malicious campaign involving obfuscation and remote command execution.
The twitter-request-diagnostics package on PyPI version 1.0.0 contains malicious code that acts as a backdoor. It exposes an endpoint that can exfiltrate environment variables and execute arbitrary commands sent to it, posing a significant security risk to applications using this package.
The PyPI package 'test-01-rafamarrara-learning-dc' version 1.0.0 contains malicious code that acts as a backdoor. It offers middleware that purportedly adds a diagnostic endpoint to web applications but secretly allows exfiltration of environment variables and execution of arbitrary commands sent by an attacker.
The bq-helix-build-hook-6ade983e package on PyPI version 0.1.0 contains malicious code that collects sensitive information from the environment during installation. This includes credential files, environment variables, and data from cloud metadata endpoints, which are then exfiltrated. The campaign associated with this malware is identified as 2026-10-bq-helix-recon and is characterized by clear malicious intent to steal information.
The bq-helix-recon-6ade983e package on PyPI contains malicious code that executes on every Python startup by placing PTH/sitecustomize files. This code collects sensitive information such as credential files, environment variables, and cloud metadata, then exfiltrates this data. The campaign associated with this package is clearly malicious, functioning as an infostealer targeting cloud tokens and other credentials.
An unauthenticated remote denial of service vulnerability exists in the Twisted HTTP/2 server due to improper handling of rapid stream resets. A remote client can send a request with END_STREAM followed immediately by a stream reset, causing the server to render the request fully before processing the reset. This allows repeated rapid resets on a single connection to exhaust CPU resources, bypassing the concurrent-stream limit. The issue affects Twisted versions prior to 26.10.0rc1 and has no impact on confidentiality or integrity.
PraisonAI versions prior to 4.6.78 contain a vulnerability in the JobWorkflowExecutor._exec_inline_python() function that allows remote code execution. The vulnerability arises because the sandbox intended to safely execute inline Python expressions is broken: the 'os' module is allowed to be imported, and dangerous functions like os.system() can be called. This enables an attacker to execute arbitrary OS commands with the privileges of the user running the workflow, particularly in multi-tenant or CI environments where untrusted YAML workflows may be executed.
The poc-revshell package version 0.1.0 on PyPI contains a reverse-shell backdoor that activates both during installation and import. It opens a TCP connection to localhost on port 4444, providing an interactive shell to any local listener. This backdoor runs with the installing user's privileges and is confined to local attackers on the same host.
The seapy-ops package on PyPI, versions 1.0.0 through 1.2.2, contains malicious functionality hidden behind undocumented commands. It masquerades as a local file analysis tool but includes backdoor features that exfiltrate files from the user's working directory to an attacker-controlled server and allow remote code execution. The package uploads files recursively, including via symlinks, and can write arbitrary files into the working directory based on server responses. Later versions implement a Python REPL that sends input to the server and executes server-supplied code in memory, disguising errors to conceal activity. These behaviors were confirmed through isolated testing with synthetic credentials and mocked servers, without contacting live endpoints.