Septembre 2026

agent-zero 1.7, 1.8, 1.9, and 1.10 is vulnerable to Directory Traversal in python/helpers/file_browser.py:FileBrowser.__init__. The FileBrowser class initializes with the host root directory as the workspace, allowing the agent to access any file on the system without restriction.

agent-zero 1.7, 1.8, 1.9, and 1.10 is vulnerable to Directory Traversal in python/helpers/file_browser.py:FileBrowser.save_file_b64. The save_file_b64 method accepts user-controlled file paths without normalization or validation, allowing path traversal attacks.

py-pdf pypdf · Pypdf project pypdf

pypdf is a free and open-source pure-python PDF library. Prior to 6.19.0, a crafted PDF can provide unusually large alphabetical page-label values that cause pypdf/_page_labels.py to generate strings beyond a reasonable page-label length when an application retrieves document page labels, consuming excessive memory and potentially making the application unavailable. This issue is fixed in version 6.19.0.

py-pdf pypdf · Pypdf project pypdf

pypdf is a free and open-source pure-python PDF library. Prior to 6.19.0, a crafted PDF containing many embedded files can cause the dictionary-based attachments API in pypdf/_doc_common.py to reparse the full attachment list for each content lookup, producing repeated work and long runtimes when an application accesses the embedded-file mapping. This issue is fixed in version 6.19.0.

py-pdf pypdf · Pypdf project pypdf

pypdf is a free and open-source pure-python PDF library. Prior to 6.19.0, a crafted PDF with form field values can cause pypdf/generic/_appearance_stream.py appearance-stream generation to repeat invariant selection-data work inside a loop when an application updates fields with flattening enabled, resulting in excessive runtimes and application unavailability. This issue is fixed in version 6.19.0.

py-pdf pypdf · Pypdf project pypdf

pypdf is a free and open-source pure-python PDF library. Prior to 6.18.1, a crafted PDF containing a partially malformed /FlateDecode stream with padded data can force pypdf/filters.py to use inefficient byte-by-byte decompression while the earlier recovery counter fails to advance for bytes that successfully decode, causing long runtimes and application unavailability. This is a residual issue after the malformed FlateDecode recovery fix. This issue is fixed in version 6.18.1.

py-pdf pypdf · Pypdf project pypdf

pypdf is a free and open-source pure-python PDF library. Prior to 6.18.1, a crafted PDF can provide a TrueType or Type1 simple font with an unusually large /Widths array, causing pypdf/_font.py Font._collect_tt_t1_character_widths to process entries beyond the 256 character codes meaningful for a simple font and consume excessive memory during operations such as text extraction. This issue is fixed in version 6.18.1.

py-pdf pypdf · Pypdf project pypdf

pypdf is a free and open-source pure-python PDF library. Prior to 6.18.1, a crafted PDF can place unusually large source-code or destination-string tokens in a font /ToUnicode mapping, causing pypdf/_cmap.py parse_bfchar to decode and retain oversized values during operations such as text extraction and consume excessive memory. This is a second follow-up to earlier /ToUnicode resource-consumption fixes and is limited to the remaining token-length path. This issue is fixed in version 6.18.1.

py-pdf pypdf · Pypdf project pypdf

pypdf is a free and open-source pure-python PDF library. Prior to 6.18.0, a crafted PDF containing indirect-object identifiers or generation-number tokens that continue for a long time without whitespace can cause pypdf/_reader.py and pypdf/generic/_base.py to scan excessive input through read_until_whitespace, resulting in long runtimes and application unavailability. This issue is fixed in version 6.18.0.

py-pdf pypdf · Pypdf project pypdf

pypdf is a free and open-source pure-python PDF library. Prior to 6.17.0, a crafted PDF can provide unusually large Roman page-label values that cause pypdf/_page_labels.py to generate excessively large numeral strings when an application retrieves document page labels, consuming large amounts of memory and potentially making the application unavailable. This issue is fixed in version 6.17.0.

Python Software Foundation CPython

ssl.SSLContext.wrap_bio() didn't require the server_hostname argument to not be None if ssl.SSLContext.check_hostname was set. Due to a missing parameter check in SSLObject, if the server_hostname argument isn't supplied then hostname verification would be silently skipped. This defect could lead to programs where certificate hostname verification *appeared* to be succeeding with SSLContext.check_hostname = True and no ValueError being raised due to misconfiguration. If the program passes a server_hostname value that isn't an empty string or None to any of these APIs then certificate hostname verification proceeds as expected and the program is not affected by this vulnerability. Mitigating this vulnerability doesn't require updating Python or applying the patch. To mitigate, pass a valid non-None and non-empty server_hostname value to SSLContext.wrap_bio(), asyncio.create_connection(), or asyncio.loop.start_tls() and certificate hostname verification will proceed as expected. Upgrading to the latest version of Python or applying the patch only changes the behavior from silently skipping hostname verification to raising a ValueError, similar to SSLContext.wrap_socket(), when server_hostname isn't supplied.

pypa virtualenv

virtualenv is a tool for creating isolated virtual python environments. Prior to 21.7.12, BatchActivator.quote() returns prompt text unchanged before activate.bat inserts it into a cmd.exe set "VAR=value" statement. An attacker who influences --prompt, VIRTUALENV_PROMPT, or the corresponding configuration value can include a double quote that closes the assignment and leaves following cmd.exe operators as executable syntax. When a user activates the generated Windows environment, the injected commands run with that user's privileges. This issue is fixed in version 21.7.12.

pypa virtualenv

virtualenv is a tool for creating isolated virtual python environments. Prior to 21.7.12, download_wheel() accepts pip and setuptools seed wheels fetched for periodic updates or the --download option without checking their bytes against an authoritative digest equivalent to the embedded wheels' BUNDLE_SHA256 verification. A compromised index, stale mirror, or intercepted TLS connection can substitute a different wheel under the requested distribution, version, and filename, after which virtualenv caches and seeds the attacker-controlled wheel into subsequently created environments. The verification applies to the default PyPI path and is intentionally skipped when PIP_INDEX_URL, PIP_EXTRA_INDEX_URL, or PIP_INDEX configures a custom index that may legitimately publish rebuilt wheels. This issue is fixed in version 21.7.12.

pypa virtualenv

virtualenv is a tool for creating isolated virtual python environments. Prior to 21.7.13, the generated activate (bash and zsh) and activate.fish scripts place values already escaped by shlex.quote inside an additional quoted context. In the bash and zsh script, a crafted virtual environment path reaches __VIRTUAL_ENV__ when a relocated environment's recorded directory is absent; in the fish script, crafted Tcl or Tk library paths reach __TCL_LIBRARY__ or __TK_LIBRARY__. The surplus quotes can terminate the data-only quoted run and leave shell metacharacters parsed as commands when a user sources the activation script, allowing code execution with that user's privileges. This issue is fixed in version 21.7.13.

urllib3 · Python urllib3

urllib3 is an HTTP client library for Python. From 1.10.3 until 2.8.0, the HTTPResponse.read_chunked and HTTPResponse.stream methods can allocate unbounded memory because the streaming chunk parser buffers the chunk-size field until newline or EOF without a length bound. The trigger is that a malicious server returns Transfer-Encoding: chunked followed by a very long run of bytes without a newline. The attack mechanism is that a malicious HTTP server sends a very long unterminated chunk-size line. The impact is that unbounded memory allocation can exhaust the client process. This issue is fixed in version 2.8.0.

urllib3 · Python urllib3

urllib3 is an HTTP client library for Python. From 1.26.0 until 2.8.0, the proxy_ssl_context, proxy_assert_hostname, proxy_assert_fingerprint, ssl_context, cert_reqs, verify_mode, use_forwarding_for_https=True, and CERT_NONE configuration paths fail to remain separated because target-server TLS settings are incorrectly applied to the HTTPS proxy connection. The trigger is that an application uses an HTTPS proxy and configures target-server TLS settings that must remain separate from the proxy TLS handshake, including HTTPS forwarding with target-specific identity or credentials. Applying cert_reqs=CERT_NONE can overwrite proxy_ssl_context.verify_mode in place, and the mutation persists so later connections reusing the same context may connect to the HTTPS proxy without certificate verification. The attack mechanism is that an attacker intercepts and impersonates the HTTPS proxy after the effective proxy policy accepts the attacker's certificate. The impact is that the attacker can observe or modify forwarded traffic or receive a target TLS client certificate, while CONNECT tunneling still preserves the separate end-to-end target TLS connection. This issue is fixed in version 2.8.0.

Open Source Robotics Foundation Robot Operating System 2 (ROS 2)

A code injection vulnerability has been discovered in the Robot Operating System 2 (ROS 2) 'ros2topic' command-line tool, affecting all ROS 2 distributions from Crystal Clemmys up to and including Lyrical Luth and Rolling Ridley. The vulnerability lies in the 'hz' verb, which reports the publishing rate of a topic and accepts a user-provided Python expression via the --filter option. This input is passed directly to the eval() function without sanitization, allowing a local user to craft and execute arbitrary code.

jpadilla pyjwt · Pyjwt project pyjwt

PyJWT is a Python implementation of JSON Web Token standards. From 2.1.0 until 2.15.0, PyJWT OKPAlgorithm.from_jwk in jwt/algorithms.py is affected because private-JWK import path does not compare the public key derived from d with x. This occurs when an OKP private JWK supplies non-corresponding x and d components. As a result, identity derived from x can differ from operations performed with d. Consequently, if an integration also accepts private key parameters from a proof header without rejecting them, an attacker may use a stolen sender-constrained token without the legitimate private key. This issue is fixed in version 2.15.0.

jpadilla pyjwt · Pyjwt project pyjwt

PyJWT is a Python implementation of JSON Web Token standards. From 2.9.0 until 2.14.0, PyJWKSet does not catch the plain ValueError raised for malformed RSA JWK components by RSAAlgorithm.from_jwk in jwt/api_jwk.py. This occurs when a JWK Set contains a malformed RSA key alongside otherwise usable keys. As a result, one malformed member aborts construction of the entire PyJWKSet. Consequently, applications can experience authentication failures or request-level denial of service. This issue is fixed in version 2.14.0.

jpadilla pyjwt · Pyjwt project pyjwt

PyJWT is a Python implementation of JSON Web Token standards. From 2.13.0 until 2.14.0, PyJWT HMACAlgorithm.prepare_key is affected because HMAC key guard only recognizes top-level public JWK forms and misses container representations. This occurs when an application allows HMAC and asymmetric algorithms and passes a public JWK container as the raw key. As a result, public asymmetric key material is accepted as the HMAC secret. Consequently, an attacker who knows the public key can forge a token with arbitrary authenticated claims. This issue is fixed in version 2.14.0.

jpadilla pyjwt · Pyjwt project pyjwt

PyJWT is a Python implementation of JSON Web Token standards. From 2.13.0 until 2.14.0, HMACAlgorithm.prepare_key in jwt/algorithms.py is affected because raw-JWK detector does not normalize accepted Unicode byte-order marks before checking for JSON. This occurs when a public JWK is prefixed with a UTF-8 BOM and used in a mixed-algorithm verification path. As a result, public JWK bypasses asymmetric-key detection and becomes the HMAC secret. Consequently, an attacker who knows the public key can forge authenticated tokens. This issue is fixed in version 2.14.0.

jpadilla pyjwt · Pyjwt project pyjwt

PyJWT is a Python implementation of JSON Web Token standards. From 2.4.0 until 2.14.0, PyJWT HMACAlgorithm.prepare_key is affected because asymmetric-key guard relies on textual markers that are absent from DER encoding. This occurs when an application mixes HMAC and asymmetric algorithms and supplies a DER public key as the shared verification key. As a result, PyJWT uses public DER bytes as an HMAC secret. Consequently, an attacker who knows the public key can forge authenticated HMAC tokens. This issue is fixed in version 2.14.0.

jpadilla pyjwt · Pyjwt project pyjwt

PyJWT is a Python implementation of JSON Web Token standards. From 2.0.0a1 until 2.15.0, PyJWT PyJWKClient.get_signing_key_from_jwt is affected because payload parser catches ValueError but not RecursionError. This occurs when an attacker-controlled recursively nested payload reaches json.loads. As a result, documented PyJWT exception handling does not contain the failure. Consequently, an unauthenticated request can raise an exception that may produce an HTTP 500 response. The advisory-defined affected implementation also includes jwt/api_jwt.py, verify_signature=False. This issue is fixed in version 2.15.0.

jpadilla pyjwt · Pyjwt project pyjwt

PyJWT is a Python implementation of JSON Web Token standards. Prior to 2.14.0, PyJWT get_signing_key_from_jwt is affected because unknown kid misses force refreshes without a negative cache or minimum refresh interval. This occurs when unauthenticated tokens repeatedly use the same unknown kid or varying kid values absent from the cached JWKS. As a result, each cache miss causes PyJWKClient to refresh the JWKS. Consequently, attacker traffic can amplify outbound requests to the configured JWKS endpoint. This issue is fixed in version 2.14.0.

unslothai unsloth-zoo · unslothai unsloth

Unsloth Zoo versions 2025.9.9 before 2026.8.14, as implemented in Unsloth 2025.9.9 through 2026.8.19, contains a code injection vulnerability in the model-loading compile path where the get_transformers_model_type() function in hf_utils.py collects model_type values from nested model configurations without enforcing a character allowlist, allowing newlines and arbitrary Python source to survive normalization. Attackers can embed a newline in a nested model_type value within a malicious model's config.json to terminate the generated import statement and execute arbitrary Python code via exec() in unsloth_compile_transformers(), achieving remote code execution as the loading user when the model is loaded for training or inference.

universal-tool-calling-protocol python-utcp

python-utcp versions before 1.1.4 contain a server-side request forgery vulnerability in HttpCommunicationProtocol.call_tool that validates the initial tool URL but follows HTTP redirects without re-validating the target. Attackers controlling a tool endpoint can return a 302 redirect to internal services, allowing the UTCP client to reach cloud metadata endpoints or internal HTTP services and return their response bodies to the caller.

universal-tool-calling-protocol python-utcp

utcp-http before 1.1.4 fails to validate the OAuth2 tokenUrl field from remote OpenAPI specifications, allowing attackers to redirect credential submission to arbitrary endpoints. When a victim registers an attacker-controlled OpenAPI spec and invokes a generated OAuth2-protected tool, the library POSTs the victim's client_id and client_secret to the attacker-supplied token endpoint without URL validation.

heymrun heym

Heym before 0.0.53 evaluates workflow condition expressions using Python's eval() with insufficient sandboxing in the workflow executor service. Authenticated users can edit workflow condition nodes or import malicious templates to execute arbitrary Python and OS commands as the backend process user.

heymrun heym

heym before 0.0.91 contains a sandbox escape vulnerability in the expression engine's DotList map/filter and fallback resolver that allows authenticated users to execute arbitrary Python code. Attackers can craft workflow expressions using dunder attribute access through item expressions or the fallback resolver to access os.system and execute commands as the backend process.

Project-MONAI MONAI

MONAI through 1.6.0 contains a remote code execution vulnerability in the bundle configuration engine that resolves _target_ values to arbitrary importable callables without an allow list and passes $ expressions to Python eval(). Attackers can publish a malicious bundle with crafted configuration containing arbitrary code that executes when a victim loads the bundle using monai.bundle.load() or monai.bundle.run().

OpenClaw

OpenClaw (npm package 'openclaw') versions >= 2026.3.28 and < 2026.8.1 allow an untrusted workspace .env file to set the CLOUDSDK_PYTHON_ARGS environment variable. When an operator starts OpenClaw in attacker-controlled workspace content and then runs the Gmail setup flow, that value is inherited when gcloud is launched, and the gcloud launcher passes it as arguments to the trusted Python interpreter. A crafted CLOUDSDK_PYTHON_ARGS value can therefore cause Python to execute attacker-supplied code with the OpenClaw host user's permissions, allowing credentials to be read, files to be modified, or other processes to be started. This issue is fixed in version 2026.8.1; as a workaround, run Gmail setup only from trusted workspaces and clear inherited CLOUDSDK_* variables beforehand.

horilla-hr

Horilla is an HR and CRM software. Prior to 2.0.0, HorillaListView.export_data in horilla_views/generic/cbv/views.py accepts an authenticated user's columns POST parameter, takes field_tuple[1], interpolates it into dynamic_fn_str as Python source, and passes the generated function definition to exec(). A crafted string that remains valid under ast.literal_eval can inject Python syntax into a default argument evaluated during function definition, allowing arbitrary operating-system commands to execute with the application process privileges, including root privileges in the shipped Docker image. This issue is fixed in version 2.0.0.

MongoDB Python Driver

An integer overflow in the BSON document encoding component of the MongoDB Python Driver's bundled native extension may occur when a single document is built from an unusually large amount of caller-supplied data. Size arithmetic is performed in a signed 32-bit type, and the guard meant to catch the overflow is written in a form whose behavior is not defined by the C language standard. A party with no privileges who can place a very large value into data that an application encodes may, depending on how the native extension was built, cause a write outside the bounds of an allocated buffer inside the application's own process.

python-social-auth social-core

Python Social Auth is a social authentication/registration mechanism. Prior to version 5.0.0, the `vk-app` backend accepted VK application callback data without verifying the callback signature when the `auth_key` parameter was omitted. Applications using this backend could treat unsigned attacker-controlled data as a verified VK identity. An attacker could choose callback fields such as `viewer_id`, `access_token`, `api_id`, and `api_result`, potentially allowing authentication as an arbitrary VK user ID. The issue affects only applications using the `vk-app` backend. The issue has been fixed in version 5.0.0 by requiring `auth_key` to be present and valid before callback data is trusted.

Red Hat Ansible Automation Platform 2.6 for RHEL 9 · Red Hat Ansible Automation Platform 2.6 · Red Hat Ansible Automation Platform 2.7 · Red Hat Ansible Automation Platform 2

A path traversal vulnerability was found in pulpcore. The content upload API accepts a 'file_url' parameter that allows users with file repository privileges to specify a local file URL for Pulp to download and store. A URL scheme validation check uses a string prefix comparison that only rejects URLs beginning with 'file://', but Python's URL parser recognizes the 'file:' scheme without double slashes, creating a mismatch between what is validated and what is dispatched to the file downloader. An authenticated user with low-privilege repository permissions can supply a specially crafted URL using relative path traversal sequences to read any file accessible to the Pulp server process. In deployments that include Pulp Container, successful exploitation allows an attacker to read the container registry token signing private key and forge bearer tokens, granting unauthorized access to all private container repositories in the affected registry.

Red Hat Ansible Automation Platform 2.5 for RHEL 8 · Red Hat Ansible Automation Platform 2.5 for RHEL 9 · Red Hat Ansible Automation Platform 2.6 for RHEL 9 · Red Hat Ansible Automation Platform 2.6

A flaw was found in Ansible Automation Platform's automation-controller. The custom Credential Type environment-variable injector validates variable names against a deny-list (an ANSIBLE_* prefix check plus a fixed ENV_BLOCKLIST) that omits process-hijacking loader variables such as BASH_ENV, ENV, LD_PRELOAD, LD_LIBRARY_PATH, PYTHONSTARTUP and GIT_SSH_COMMAND. Combined with the credential file injector, a privileged user can write an attacker-controlled script into the execution environment and point BASH_ENV at it, obtaining arbitrary code execution inside the execution-environment container for any job that attaches a credential of that type.

Red Hat Ansible Automation Platform 2.5 for RHEL 8 · Red Hat Ansible Automation Platform 2.5 for RHEL 9 · Red Hat Ansible Automation Platform 2.6 for RHEL 9 · Red Hat Ansible Automation Platform 2.6

A flaw was found in Red Hat Ansible Automation Platform's automation- controller. The setting that formats the log message emitted for API 4XX errors is an administrator-controlled Python format-string template that is rendered with a live user object as an argument. Because Python string formatting permits attribute and item traversal on its arguments, an administrator can craft a template that walks from the user object into the application settings and reads the Django secret key and the database password. The formatted message is written to a logger that can be forwarded to an external log aggregator, whose destination is also administrator-controlled, allowing the secrets to be sent off the host. An authenticated administrator can thereby obtain the master encryption key used to protect all stored credentials and the database service password, enabling offline decryption of every stored credential, forgery of user sessions, and direct access to the controller database.

OpenC3 cosmos

OpenC3 COSMOS provides the functionality needed to send commands to and receive data from one or more embedded systems. From 5.12.0 until 7.3.0, an authenticated actor can write the pypi_url setting through set_setting at POST /openc3-api/api, then cause OpenC3::PluginModel.install_phase2 in openc3/lib/openc3/models/plugin_model.rb to interpolate the value into a shell command while installing a plugin with Python dependency metadata. Shell metacharacters in the setting are interpreted by the command shell, allowing arbitrary operating-system commands to run as the openc3 service user with access to Redis and bucket credentials. Open-source deployments permit any authenticated user to reach the affected operations, while Enterprise deployments require an administrator. This issue is fixed in version 7.3.0.

webpy web.py 0.76 is vulnerable to server-side template injection (SSTI). The template engine can be tricked into executing attacker-controlled template code that built-in security checks are designed to reject. When an application precompiles templates from a directory the attacker can write to and later renders them through the precompiled template loader, the sandbox is bypassed and the attacker's code runs, resulting in arbitrary Python code execution and OS command execution on the server.

Pollen Robotics Reachy Mini

The Reachy Mini daemon exposes an HTTP API for managing the robot. Its app installation endpoint, POST /apps/install in src/reachy_mini/daemon/app/routers/apps.py, has no authentication. The handler's only dependency is Depends(get_app_manager), which just hands back the manager object from application state, so nothing in the chain ever checks a credential. The endpoint takes an AppInfo body naming a Hugging Face Space. The daemon downloads that Space and installs it as a Python package through install_package in src/reachy_mini/apps/sources/local_common_venv.py, using uv or pip. Installing a Python package runs the package's own build and setup code, so whoever chooses the Space chooses what code the robot runs. Anyone can publish a public Hugging Face Space, so this is not a meaningful restriction on the attacker. How far this reaches depends on the model. In _resolve_bind_host in src/reachy_mini/daemon/app/main.py the daemon binds 0.0.0.0 when it runs as the wireless version and 127.0.0.1 otherwise, with the vendor's own comment explaining that the robot has to be reachable on the LAN. On a wireless unit, then, any host on the same network can install and run code on the robot without credentials. One related change has already shipped but does not fix this. Version 1.8.2 replaced the wildcard CORS policy with an allow list of localhost and Tauri origins. That closes the browser drive-by route, where a web page the victim visits silently calls the endpoint in the background. It has no effect on this issue: CORS is enforced by browsers and governs whether script may read a response, while a direct HTTP request from another machine on the network involves no browser, no preflight and no CORS check at all.

psd-tools

psd-tools is a Python package for working with Adobe Photoshop PSD files. Prior to 1.17.4, PSDImage.composite() and PSDImage.numpy() allocated output buffers from attacker-controlled PSD header geometry, including width, height, channels, depth, and per-layer rectangles, before validating those values against the available file data. A tiny crafted PSD could therefore cause multi-gigabyte memory allocation, and PSDImage.composite() could return a black image with only a warning instead of raising an exception. Services that composite untrusted PSD files could be terminated by out-of-memory handling. This issue is fixed in version 1.17.4.

Crypt::SelfCertificate versions from 1.01 through 1.05 for Perl contains malware which executes Python code from an obfuscated URL. The generate_certificate runs a Python script saved as a certificate file. The pyhton script attempts to retrieve code from a hardcoded http URL that is obfuscated with base64 encoding and run the response body directly. The impact is that arbitrary code can be invoked as the user, without a dropped script being saved on the affected host. The releases have no test scripts nor build hooks. The intention may have been to trigger the payload after installation. For version 1.01, the dropper script is in lib/Crypt/SelfCertificate/sample/validate.p12. For version 1.05, the dropper script is in lib/Crypt/SelfCertificate/sample/cert7.pem. The SHA-256 digests of the files are fbff21f45ff748365062a5e36fb2d72558cad82a507a6f357f320b4fcdf07760 Crypt-SelfCertificate-1.01.tar.gz 27b2d2d3174ad771474fff2521f5084ec231e9218ea8c832515aef1cbd5897bc lib/Crypt/SelfCertificate/sample/validate.p12 9fdfa7d69b034b77d4510cda567e8da1e486ca81c7daaadc5732a45c41d71991 Crypt-SelfCertificate-1.05.tar.gz 27b2d2d3174ad771474fff2521f5084ec231e9218ea8c832515aef1cbd5897bc lib/Crypt/SelfCertificate/sample/cert7.pem

sooperset mcp-atlassian · Mcp-atlassian mcp atlassian

MCP Atlassian is a Model Context Protocol (MCP) server for Atlassian products (Confluence and Jira). Prior to 0.22.0, validate_safe_path defaults its base directory to os.getcwd(), and affected Confluence attachment call sites omit base_dir, allowing attacker-selected writes within the working directory. This Python module overwrite can provide code execution when the application later imports the modified module, bypassing the remediation tracked as CVE-2026-27825. This issue is fixed in version 0.22.0.

kap-sh zapros

Zapros, a Python HTTP client, prior to version 0.14.0 is vulnerable to denial of service via memory exhaustion. The issue affects all callers who streamed compressed responses relying on the chunk size — explicit (`iter_bytes(chunk_size=...)`) or the default — to bound memory. The decoder ignored that bound, so a chunk could be far larger than requested and a single compressed response could overflow memory. Version 0.14.0 contains a patch. Some workarounds are available. Read the still-compressed body with `Response.iter_raw()` / `Response.async_iter_raw()`, which bypass the built-in decoders, and decompress it yourself with an explicit output-size bound (e.g. `zlib`'s `max_length`), aborting once a configured limit is exceeded. Where feasible, send `Accept-Encoding: identity` to disable response compression so bodies are not decompressed client-side. Avoid decoding response bodies from untrusted servers.

Tautulli

Tautulli is a Python based monitoring and tracking tool for Plex Media Server. Prior to 2.17.2, the import_config handler and the database_file branch of import_database in plexpy/webserve.py join the attacker-controlled config_file.filename or database_file.filename directly to CACHE_DIR without basename reduction or a containment check. An administrator or caller with the instance API key can submit a multipart filename containing parent-directory segments, causing the upload to be created or overwritten outside CACHE_DIR before file-content validation runs. The write is limited to paths permitted to the Tautulli process, but it can enable configuration tampering, service disruption, or code execution. This issue is fixed in version 2.17.2.

conda

Conda is a system-level binary package and environment manager that runs on major operating systems and platforms. Prior to 26.5.2, parse_entry_point_def in conda/common/path/python.py accepted an unvalidated entry-point command from a noarch:python package's info/link.json metadata. CreatePythonEntryPointAction in conda/core/path_actions.py interpolated that command into target_short_path, and PrefixPathAction.target_full_path joined it to the installation prefix without verifying that the result remained under the intended bin or Scripts directory. create_python_entry_point in conda/gateways/disk/create.py then wrote an executable wrapper to the resulting path. A malicious package could use path separators, traversal segments, or an absolute command path to write outside the prefix or overwrite another in-prefix entry point during default install and environment transactions. Out-of-prefix writes require the target parent directory to exist, while an overwritten entry point can execute attacker-controlled Python when later invoked with the installing user's privileges. This issue is fixed in version 26.5.2.

SepineTam mcp-for-stata

MCP-for-Stata is a MCP server for integrating Stata into agent loops with a safety-first design. Prior to version 1.19.0, the ado_package_install MCP tool in stata-mcp concatenates user-controlled input directly into a Stata command string without any validation or sanitization. An attacker who can invoke the MCP tool or the equivalent Python API can embed newline characters in the package argument to inject arbitrary Stata commands. Because Stata supports a shell escape command, this leads to full OS-level arbitrary command execution (RCE) under the account running the Stata-MCP server. The tool is registered in the default all profile, so no non-default configuration is required. This issue has been patched in version 1.19.0.

Syslifters sysreptor

SysReptor is a fully customizable pentest reporting platform. Prior to 2026.61, authenticated users of SysReptor Professional can upload image files whose formats cause image processing to invoke Ghostscript, allowing embedded PostScript to operate in the shared temporary directory. An attacker can combine that behavior with a race involving GnuPG configuration files in temporary subdirectories to cause GnuPG to copy attacker-controlled Python code into the application code directory. The injected code executes with the privileges of the SysReptor application process after a worker restart. The Community edition is not affected. Version 2026.58 contains a partial mitigation, and this issue is fully fixed in version 2026.61.

InternLM lmdeploy

LMDeploy is a toolkit for compressing, deploying, and serving large language models. Versions 012.1 through 0.12.2 contain a code injection vulnerability in `lmdeploy/pytorch/config.py` line 620 that allows an attacker to execute arbitrary Python code by publishing a malicious HuggingFace model with a crafted `quantization_config.quant_dtype` value. When a user loads the model with lmdeploy, the `quant_dtype` is passed to `eval(f'torch.{quant_dtype}')` without any validation. Version 0.12.3 contains a patch.

dataease SQLBot

SQLBot is an intelligent Text-to-SQL system based on large language models and RAG. Prior to 1.9.0, the POST /api/v1/datasource/parseExcel endpoint in backend/apps/datasource/api/datasource.py uses attacker-controlled multipart filename data when selecting where uploaded content is stored, writes the content before spreadsheet parsing and validation finish, and can transform a double-extension filename into a Python source file. An attacker able to submit a crafted multipart upload can use these behaviors to place attacker-controlled content in /opt/sqlbot/app/alembic/versions/ even when a spreadsheet parsing failure after the file write causes the endpoint to return an error. The planted file remains on disk, and subsequent SQLBot startup or migration processing causes Alembic to import the module and execute its module-level statements in the SQLBot application runtime. This issue is fixed in version 1.9.0.

netbox-community devicetype-library

NetBox Device Type Library is a collection of community-sourced device type definitions for import into NetBox. The absence of tests/init.py and the lack of --import-mode=importlib cause pytest prepend import mode to place the tests directory at the front of sys.path during collection. An unauthenticated contributor can add a module such as tests/git.py that shadows GitPython when tests/definitions_test.py executes from git import Git, Repo, or add tests/conftest.py for automatic collection-time execution. Python imports and runs the pull-request module before any test function, allowing arbitrary code execution on the GitHub Actions runner, test-result tampering, and access to tokens or network resources exposed to the workflow. This module-shadowing path is independent of the earlier pickle deserialization flaw and the separately tracked NETBOX_DT_LIBRARY_URL issue. This vulnerability is fixed by commit b0d9a3dadd0a0a9d3c93b0b2777559fd4bad1037.

AWSIoTPythonSDK

Improper validation of certificate with host mismatch in the MQTT client TLS connection layer in AWS IoT Device SDK for Python 1.5.3 through 1.6.0 on Python 3.7 and later might allow an adversary-in-the-middle actor to impersonate the AWS IoT Core endpoint, read device telemetry, and inject arbitrary MQTT messages that the device processes as authentic, via a certificate issued for an unrelated hostname by a certificate authority present in the device trust store. To remediate this issue, users should upgrade to version 1.6.1.

Linux

In the Linux kernel, the following vulnerability has been resolved: ocfs2: validate inline xattrs during inode block validation Patch series "ocfs2: validate xattr entry bounds", v7. This series validates OCFS2 xattr entry name/value bounds when xattr metadata is read and validated, before getxattr() or listxattr() can walk out-of-range entry arrays or offsets from corrupted metadata. This patch (of 2): ocfs2_validate_inode_block() verifies a dinode before OCFS2 users walk metadata from it, but inline xattr metadata is still checked only in operation-specific consumers. The existing ibody lookup helper validates inline header placement and entry count, but inode block validation does not reject entry name/value bounds. Add a flat xattr entry validator and call it from inode block validation for inline xattrs. Keep the operation paths on their existing header/count lookup checks; the full entry bounds check now runs when the inode block is validated at read time. Reject corrupted inline xattr metadata before ocfs2_xattr_ibody_get() or listxattr() can walk past the inline storage. Validation reproduced this kernel report: BUG: KASAN: use-after-free in ocfs2_xattr_find_entry+0x5a/0x170 Read of size 2 at addr ffff8881242a2000 by task python3/529 Call Trace: dump_stack_lvl+0x66/0xa0 print_report+0xce/0x630 kasan_report+0xe0/0x110 ocfs2_xattr_find_entry+0x5a/0x170 ocfs2_xattr_get_nolock+0x20a/0x820 ocfs2_xattr_get+0x10c/0x1e0 __vfs_getxattr+0xe2/0x130 vfs_getxattr+0x185/0x1b0

sanic-org sanic

Sanic is an opensource python web server/framework. Prior to version 24.12.1, and in version 25.12.0, the HTTP/1.1 response pipeline in sanic/response/types.py serializes response header names and values without rejecting carriage-return or line-feed characters. Applications that place attacker-controlled data in response.headers, file(..., filename=...), or cookie path and domain attributes can therefore emit injected headers and may split responses. Depending on application and proxy behavior, this can enable session fixation through injected cookies, cache poisoning, or security-header corruption. This issue is fixed in versions 24.12.1 and 25.12.1.

djust-org djust

djust provides Phoenix LiveView-style reactive server-side rendering for Django with Rust-powered performance. Prior to version 1.0.7, the djust live transport resolves the LiveView to mount from a client-supplied dotted path by calling `__import__(module_path, ...)`. The module is imported — running its top-level code (import side effects) — before the framework checks that the resolved object is a `LiveView` subclass and before any per-view authentication. The `LIVEVIEW_ALLOWED_MODULES` allowlist that should contain this is fail-open (`if allowed_modules:` — skipped when the setting is unset, the framework default) and uses loose `startswith` matching. An unauthenticated WebSocket client (the WS handshake does not require auth; per-view auth runs only after import + instantiate) can therefore send a `mount` / `live_redirect_mount` / `url_change` frame (or an SSE mount) with `view = "<any.importable.module>.AnyName"` and cause the server to import — and execute the top-level code of — any importable Python module by name. Version 1.0.7 fixes the issue with a fail-closed resolution gate (`djust._view_resolution.is_view_import_allowed`): a client view path resolves only if (a) its module is already loaded (`sys.modules` — so resolving runs no new code; URL-routed views loaded by URLconf at startup keep working with zero config) or (b) it matches `LIVEVIEW_ALLOWED_MODULES` on a module-segment boundary (explicit opt-in for lazily-imported views). The gate runs before `__import__` at all three sinks (+ defense-in-depth inside `_instantiate_view`). As a workaround, set `LIVEVIEW_ALLOWED_MODULES` to the narrow list of modules that contain your mountable LiveView classes. (Note: pre-patch the allowlist is `startswith`-matched and the import still precedes the subclass check, so this is mitigation, not a complete fix.)

stanfordnlp stanza

Stanza is a Stanford NLP Python library for tokenization, sentence segmentation, NER, and parsing of many human languages. Prior to 1.14.0, stanza.resources.common.unzip in stanza/resources/common.py passes downloaded model and resource archives to zipfile.ZipFile.extractall without validating member paths, and the vulnerable extraction path is reachable through stanza.download and stanza.install_corenlp. A malicious archive containing parent-directory traversal entries can write outside the intended model directory, allowing files writable by the Stanza process to be overwritten and potentially enabling code execution through modified shell configuration, SSH authorization data, Python packages, or executable scripts. This issue is fixed in version 1.14.0.

backstage · Linuxfoundation backstage plugin-techdocs-node · Linuxfoundation backstage

Backstage is an open framework for building developer portals. Prior to 1.14.6 and from 1.15.0 until 1.15.4, the @backstage/plugin-techdocs-node package insufficiently validates mkdocs.yml supplied by an authenticated user who can register or modify a TechDocs source. Unsafe Python YAML tags, markdown_extensions names and configuration, theme options, and extra_templates values can reach the documentation generator and cause unintended code execution. The resulting impact is limited to the files, credentials, network access, and other resources available to the TechDocs backend or build container. This issue is fixed in versions 1.14.6 and 1.15.4.

zopefoundation RestrictedPython

RestrictedPython is a tool that helps define a subset of the Python language for accepting program input in a trusted environment. Prior to 8.4, RestrictedPython could allow a sandbox escape when a custom import policy or globals exposed the standard library string module, the string.Formatter class, a Formatter instance, or a Formatter subclass to restricted code. The string.Formatter methods format, get_field, get_value, and vformat performed attribute and item traversal internally without passing through RestrictedPython's safer_getattr protections. Restricted code could use those live object references to reach function globals, builtins, file access, or code execution primitives, affecting confidentiality, integrity, and availability in the host environment. This issue is fixed in version 8.4.

Apache Software Foundation Apache Airflow Apache Kafka provider · Apache apache-airflow-providers-apache-kafka

Apache Airflow Apache Kafka provider versions 1.15.0 before 2.0.0 resolve dotted-path strings found in a Kafka connection's `extra` field into Python callables via `import_string`, with no allowlist, and hand them to the confluent-kafka client which invokes them. Deployments that have enabled the Kafka event producer — `dag_run_events_enabled` or `task_instance_events_enabled`, both disabled by default — build that client inside the scheduler process, so a user whose only privilege is editing Airflow connections gains arbitrary code execution in the control plane; the Airflow security model limits connection-configuration users to code execution on workers, not the scheduler. Deployments using Google Managed Kafka are not affected, because that code path overwrites any user-supplied `oauth_cb`; plain brokers and Amazon MSK are exposed. Users are recommended to upgrade to apache-airflow-providers-apache-kafka 2.0.0 or later, which adds an allowlist configuration option for connection-string callbacks.

a2ui-project a2ui

A vulnerability has been found in a2ui-project a2ui up to 0.10.7. Affected by this vulnerability is the function httpx.get of the file agent_sdks/python/a2ui_agent/src/a2ui/extensions/file_resolve/file_resolver.py of the component FileResolver. The manipulation leads to server-side request forgery. It is possible to initiate the attack remotely. The identifier of the patch is 2bb8423060308bbdea8ba468dabed4fc256d18ea. To fix this issue, it is recommended to deploy a patch.

unclecode crawl4ai

Crawl4AI before 0.9.3 contains a server-side request forgery vulnerability in PDFContentScrapingStrategy where _get_pdf_path() re-downloads targets with Python requests without egress validation. Authenticated attackers can supply URLs that redirect to internal addresses or use DNS rebinding to access internal services, exfiltrating responses through PDF text extraction in crawl results.

Google Cloud Gemini Enterprise Agent Platform SDK for Python

Bucket Squatting in Google Cloud Gemini Enterprise Agent Platform SDK for Python versions prior to 1.166.1 allows an attacker to achieve Remote Code Execution (RCE) and tenant-project token theft.

wandb

Weights & Biases wandb before 0.29.0 fails to validate the file name from server responses in the File.download function, allowing path traversal attacks. Attackers controlling the backend can supply file names with directory traversal sequences to write files outside the intended download directory, potentially enabling code execution through modification of shell startup files or Python import paths.

cyberjunky python-garminconnect

garminconnect is a Python 3 API wrapper for Garmin Connect that retrieves statistics and manages activities. Prior to 0.3.5, garminconnect/client.py Client.dump creates the OAuth token directory and garmin_tokens.json without explicit owner-only modes, so a permissive umask such as 022 can leave the directory mode at 0755 and the token file mode at 0644. garmin_tokens.json contains di_refresh_token, and another unprivileged user on a shared Linux or macOS host can read the token and obtain persistent access to the victim's Garmin Connect account, including health, fitness, activity, and device data. The Garmin.login tokenstore path is affected, and a pre-existing loosely permissioned token file remains exposed until rewritten or manually restricted. This issue is fixed in version 0.3.5.

Python Software Foundation CPython

In CPython 3.13 and earlier, the tarfile module's data and tar extraction filters are vulnerable to crafted archives containing a hard link to a symbolic link. Such archives may cause extraction to modify the permissions or modification time of a file outside the destination directory, or expose the contents of that file within the extracted tree.

AsyncFuncAI deepwiki-open

DeepWiki-Open through commit d92819a contains an arbitrary file read vulnerability in the unauthenticated /ws/chat WebSocket endpoint that accepts repo_url as a filesystem path with no containment. Attackers can supply arbitrary directory paths to read all files with supported extensions including Python, JavaScript, YAML, and JSON files containing hardcoded secrets and credentials.

CrewAI

CrewAI before fb2323b offers a Python blocklist approach that operates at the wrong level of abstraction, a different vulnerability than CVE-2026-2275. Import-time blocking of module names does not address the availability of Python's complete object graph. For example, calling ctypes.CDLL(None) loads the C library without relying in any import statements. In other words, a within-process sandbox cannot merely account for the import system and instead must account for the complete runtime of the Python interpreter.

CyberPanel

CyberPanel before 2.4.4 attempts to detect an "alais" domain (i.e., a second domain that serves the same content as a primary domain; normally spelled "alias") via an ORM query filter rather than a Python "if" statement.

IBM Langflow OSS · Langflow langflow

IBM Langflow OSS 1.0.0 through 1.11.5 could allow a remote authenticated attacker to execute arbitrary Python code due to improper authorization of custom components in stored flows.

MongoDB Python Driver

Improper neutralization of special elements in data query logic in the GridFS component of the MongoDB Python Driver can cause a caller-supplied structured file identifier to be interpreted as a query condition rather than as a literal identifier. An authenticated user who can influence the identifier passed by an affected application may obtain stored file content beyond the intended target or cause all GridFS file chunks in the affected bucket to be removed, rendering stored file content unreadable. The affected rename operation may also rename a stored file other than the intended target.

open-webui · Openwebui open webui

Open WebUI is an extensible, feature-rich, and user-friendly self-hosted AI platform. Prior to 0.11.1, POST /api/v1/retrieval/process/web and POST /api/v1/retrieval/process/web/search in backend/open_webui/retrieval/web/utils.py treated Python's globally routable address classification as proof that a destination was external. An authenticated user could make an Azure-hosted instance fetch and return content from 168.63.129.16, the Azure platform channel, as well as other reserved ranges that the standard classification did not reject. This issue is fixed in version 0.11.1.

open-webui · Openwebui open webui

Open WebUI is an extensible, feature-rich, and user-friendly self-hosted AI platform. From 0.9.6 until 0.11.1, SafePlaywrightURLLoader in backend/open_webui/retrieval/web/utils.py validated a user-controlled hostname in Python and then let the Playwright browser resolve it again in the sync and async request interceptors. An authenticated user controlling authoritative DNS could return a public address to validation and an internal address to the browser, exposing responses from internal services or cloud metadata through web search or URL ingestion. This issue is fixed in version 0.11.1.

Red Hat Ceph Storage 5 · Red Hat Ceph Storage 9 · Red Hat OpenStack Platform 17.1 · Red Hat OpenStack Platform 18.0

A flaw was found in the memcached cache plugin of the community.general Ansible collection. Although its documentation states that records are stored in JSON format, the plugin performs no explicit serialization and relies on python-memcached, which pickles values on write and unpickles them on read. Because memcached is unauthenticated and cache keys are predictable, an attacker able to reach a network-exposed or shared memcached instance can write a crafted pickle payload that is deserialized and executed on the Ansible controller when the poisoned fact cache is next read, leading to remote code execution.

gitpython-developers GitPython · Gitpython project gitpython

GitPython before 3.1.60 contains a regular expression denial of service vulnerability in Actor.name_email_regex that processes commit author and committer fields. Attackers can craft a commit object with a malformed author field containing an unterminated angle bracket to cause quadratic backtracking, exhausting CPU resources for over two minutes per commit access.

gitpython-developers GitPython · Gitpython project gitpython

GitPython before 3.1.60 fails to properly validate the git directory location, allowing attackers to impersonate the git directory using tracked files like gitdir, commondir, and HEAD. Attackers can execute arbitrary code by placing a malicious pre-commit hook in the tracked hooks directory that executes when a victim calls index.commit() on a cloned or opened repository.

Snowflake Connector for Python · Snowflake Go Driver · Snowflake JDBC Driver · Snowflake JDBC Driver (FIPS)

In affected Snowflake drivers, WORKLOAD_IDENTITY authentication requests a cloud workload-identity token and attaches it to the login request without verifying that the configured host is a Snowflake endpoint. An attacker who can modify the connection configuration can cause the driver to mint a fresh attestation and send it to a host they control. The captured token can be replayed to Snowflake for its remaining lifetime in accounts where that workload identity is already registered. On Azure, the token audience is also taken from connection configuration. Combined with an attacker-controlled host, the driver can request a Managed Identity access token scoped to a non-Snowflake Azure resource and deliver it to the attacker. That path is the only case in which impact extends beyond Snowflake; it is bounded by the token lifetime and the managed identity’s permissions. Successful exploitation requires WORKLOAD_IDENTITY authentication on a workload that already has an ambient cloud identity. Patched driver versions restrict this authenticator to recognized Snowflake hosts. Users must manually upgrade.

Red Hat Enterprise Linux 10 · Red Hat Enterprise Linux 8 · Red Hat Enterprise Linux 9 · Red Hat Cert Manager support for Red Hat OpenShift release 1.20

A flaw was found in libxml2 with Python bindings enabled. A remote attacker could exploit this vulnerability by providing a specially crafted XML document containing a Document Type Definition (DTD) with enumerated attribute values. This triggers a double-free error in the SAX attributeDecl callback handler, where a string is freed twice. This flaw can lead to a denial of service (DoS) due to a reproducible crash in Python applications using the libxml2 SAX bindings.

axolotl-ai-cloud axolotl

Axolotl before 0.19.0 contains a remote code execution vulnerability in the multipack patch path where trust_remote_code defaults to None instead of False, causing the security guard to be bypassed. Attackers can execute arbitrary Python code by crafting a malicious Hugging Face model repository selected as base_model, which is loaded with hardcoded trust_remote_code=True during AutoModelForCausalLM.from_pretrained.

lavague-ai LaVague

LaVague 0.2.35 contains a remote code execution vulnerability in PythonFromMarkdownExtractor.extract_as_object that evaluates untrusted language model output derived from web page content. Attackers can inject malicious Python code through web pages using indirect prompt injection to execute arbitrary code on the operator's host without review.