Veille CVE

Semaine du 5 octobre au 11 octobre 2026

34 CVE

Critique 6 Élevée 10 Moyenne 16 Faible 2 Non évaluée 0 Exploitées (KEV) Tri : plus récentes « Python python » ✕
  1. CVE-2026-108597 4.8 Moyenne

    cohere-ai cohere-python

    Cohere Python SDK 5.11.0 through 7.2.0 contains a path traversal (tar slip) vulnerability in _s3_models_dir_to_tarfile that allows arbitrary file write via unvalidated tarfile.extractall calls. Attackers who can write model archives to the victim's S3 prefix can include absolute paths or ../ members to overwrite files on the SDK host.

  2. CVE-2026-108264 9.1 Critique

    wizarrrr wizarr

    Wizarr is an advanced user invitation and management system for Jellyfin, Plex, Emby, and other media servers. Prior to 2026.9.1, wizard step Markdown supplied through the editor or imported bundles was evaluated by app/blueprints/wizard/routes.py in the application's non-sandboxed Jinja2 environment with application globals exposed. An authenticated user able to create steps, or an administrator importing an untrusted bundle through POST /settings/wizard/import, could execute arbitrary Python when GET /wizard/{server}/{idx} rendered the stored step; app/jinja_filters.py and app/services/wizard_widgets.py contained additional evaluation sinks. This could execute operating-system commands as the application user, disclose the Flask SECRET_KEY, access connected service credentials and the database, and produce stored cross-site scripting. This issue is fixed in 2026.9.1.

  3. CVE-2026-108263 9.9 Critique

    iflytek astron-agent

    Astron Agent is an agentic workflow platform for building and running AI agents. Prior to 1.1.2, the default workflow code-node path through /console-api/workflow/code/run and /workflow/v1/run selects LocalExecutor in core/workflow/engine/nodes/code/code_node.py when CODE_EXEC_TYPE is not explicitly changed. LocalExecutor supplies complete Python builtins to dynamic code execution without the documented sandbox restrictions. An authenticated low-privilege tenant can execute code as root in the core-workflow container and use shared service and database credentials to bypass application-level tenant checks, read or modify other tenants' data, and disrupt shared services. This issue is fixed in version 1.1.2.

  4. CVE-2026-108258 6.9 Moyenne

    posit-dev py-shiny

    Shiny for Python is a framework for building interactive web applications in Python. From 1.4.0 until 1.6.4, bookmark restore accepts a client-supplied state_id and joins it into the server-side shiny_bookmarks directory without validating that it is a single safe path segment. An unauthenticated request can use parent-directory segments or an absolute path to make the server open input.json and values.json outside the bookmark store, even when bookmark_store is set to disable. In applications configured with bookmark_store set to server and using ui.input_file(), the restore handler can additionally copy and expose an attacker-selected file from an attacker-selected directory. This issue is fixed in version 1.6.4.

  5. CVE-2026-75597 5.3 Moyenne

    pyload

    pyLoad is a free and open-source download manager written in Python. Prior to 0.5.0b3.dev101, the `/web/<path:filename>` route in `src/pyload/webui/app/blueprints/app_blueprint.py` renders Jinja2 templates without any authentication requirement. Every equivalent direct route (`/logs`, `/settings`, `/queue`, `/dashboard`, etc.) is protected by `@login_required`, but the underlying templates for all of these pages are accessible unauthenticated via this endpoint. Combined with an exception attribute typo in `src/pyload/webui/app/handlers.py` (`exc.desc` instead of `exc.description`), internal Jinja2 variable names are leaked in HTTP 500 response bodies to unauthenticated callers. An attacker can also enumerate all valid template names by observing 200 vs 500 response differentiation. Version 0.5.0b3.dev101 contains a patch.

  6. CVE-2026-48484 6.5 Moyenne

    pyload

    pyLoad is a free and open-source download manager written in Python. Prior to 0.5.0b3.dev101, the API `rpc` function in `api_blueprint.py` handles `multipart/form-data` uploads by reading the whole content of the uploaded file into memory with `file.read()`. This occurs before the data is sent to the underlying function. Since there is no size limit set at this point, a large file upload can exhaust the server's available memory which led to process termination. Version 0.5.0b3.dev101 contains a patch.

  7. CVE-2026-107779 9.8 Critique

    dromara skyeye

    Dromara Skyeye through commit 003549ae5615bd114ba5bb8ddf6a8e8ead97c321 contains a missing authentication vulnerability in bundled xxl-job-admin JobInfoController endpoints annotated with @PermissionLimit(limit = false). Unauthenticated attackers can POST GLUE_SHELL, GLUE_PYTHON, or GLUE_POWERSHELL jobs with attacker-supplied glueSource to /jobinfo/addAndStart, executing commands on the executor host or stopping and deleting jobs.

  8. CVE-2026-106435 4.7 Moyenne

    MongoDB Python Driver

    The MongoDB Python Driver's binary accelerator can read outside a buffer when an application decodes malformed BSON containing a truncated regular-expression element without a trailing NUL byte. An actor who can supply BSON to the documented decode or decode_all API can cause the application process to terminate when the C extension is loaded. The driver's normal database wire-protocol path does not reach this code.

  9. CVE-2026-107377 7.5 Élevée

    datamodel-code-generator

    datamodel-code-generator generates Python data models from schema definitions. From 0.59.0 until 0.81.0, an attacker-controlled Protobuf schema can supply absolute or parent-directory paths captured by WEAK_IMPORT_PATTERN and consumed by _write_missing_weak_imports in src/datamodel_code_generator/parser/protobuf.py. Exploitation requires a victim or automated job to process the attacker-controlled schema with Protobuf input support, which requires the grpcio-tools package. The paths escape the weak_imports temporary directory before protoc runs, allowing creation of directory trees and new files or overwrite of existing writable files with a generated Protobuf syntax declaration. The effect persists when later Protobuf compilation fails. The written content is limited to a proto2 or proto3 syntax declaration, and direct arbitrary code execution has not been demonstrated. This issue is fixed in version 0.81.0.

  10. CVE-2026-107295 7.6 Élevée

    pydantic-ai · pydantic-ai-slim · Pydantic pydantic ai

    Pydantic AI is a Python agent framework for building applications and workflows with Generative AI. From 1.34.0 until 1.107.4 and 2.28.0, the Agent.to_web() and clai web development chat endpoint has missing request content-type validation. A website visited by a developer can submit a browser-compatible request to a loopback-hosted chat server, causing the served agent to run and execute tools with the privileges and credentials of the local process; client-relayed approval decisions also leave requires_approval=True tools exposed. Binding to localhost does not prevent a browser page from reaching the loopback address. This issue is fixed in versions 1.107.4 and 2.28.0.

  11. CVE-2026-107294 6.5 Moyenne

    pydantic-ai · pydantic-ai-slim · Pydantic pydantic ai

    Pydantic AI is a Python agent framework for building applications and workflows with Generative AI. From 1.77.0 until 1.107.2 and 2.24.0, web_fetch_tool, the WebFetch local fallback, and remote FileUrl media downloads buffer the complete HTTP response body before enforcing content-size controls. An attacker-influenced URL can stream an arbitrarily large response that exhausts process memory and crashes the worker; affected media types include ImageUrl, DocumentUrl, VideoUrl, and AudioUrl. SSRF protections remain effective, and the impact is limited to availability. This issue is fixed in versions 1.107.2 and 2.24.0.

  12. CVE-2026-107293 5.3 Moyenne

    pydantic-ai · pydantic-ai-slim · Pydantic pydantic ai

    Pydantic AI is a Python agent framework for building applications and workflows with Generative AI. From 0.3.4 until 1.107.4 and 2.27.1, OpenTelemetry instrumentation configured with InstrumentationSettings(include_content=False) can export retry prompts outside tool calls in gen_ai.input.messages and pydantic_ai.all_messages. Agents using NativeOutput, PromptedOutput, or output validators on text output can therefore disclose validation feedback, including invalid model values quoted by that feedback, to readers of the telemetry backend. Tool-call retries and deployments that do not use include_content=False are not affected by this specific path. This issue is fixed in versions 1.107.4 and 2.27.1.

  13. CVE-2026-107292 6.4 Moyenne

    pydantic-ai · pydantic-ai-slim · Pydantic pydantic ai

    Pydantic AI is a Python agent framework for building applications and workflows with Generative AI. From 1.34.0 until 2.30.0, the Agent.to_web() and clai web development chat server does not validate the Host header, allowing a website visited by a developer to use DNS rebinding to reach a loopback-hosted agent as a same-origin service. The hostile page can read the served UI and submit chat requests that execute agent tools with the local process's privileges and credentials, causing data disclosure or unwanted side effects. Binding to localhost, Origin checks, and CSRF tokens do not prevent the same-origin DNS rebinding path. This issue is fixed in versions 1.107.5 and 2.30.0.

  14. CVE-2026-107291 4.3 Moyenne

    pydantic-ai · pydantic-ai-slim · Pydantic pydantic ai

    Pydantic AI is a Python agent framework for building applications and workflows with Generative AI. From 0.3.4 until 1.107.6 and 2.44.0, OpenTelemetry instrumentation configured with InstrumentationSettings(include_content=False) can still export sensitive agent content through exception.message and exception.stacktrace events, error status descriptions, and model_request_parameters containing instructions or the prompted_output_template. The exposed data is available to readers of the configured telemetry backend and can include tool feedback, provider error bodies, runtime instructions, and structured-output templates even though message attributes are redacted. This issue does not grant new access to agent data, and deployments that do not use include_content=False are not affected by the setting bypass. This issue is fixed in versions 1.107.6 and 2.44.0.

  15. CVE-2026-107290 6.5 Moyenne

    pydantic-ai · pydantic-ai-slim · Pydantic pydantic ai

    Pydantic AI is a Python agent framework for building applications and workflows with Generative AI. From 1.77.0 until 1.107.6 and 2.44.0, the local web_fetch_tool and the WebFetch local fallback process server-controlled responses with quadratic title extraction, whitespace normalization, and ordered-list numbering. An attacker-controlled page of modest size can therefore block the event loop for an extended period, stalling other agent runs and requests, while unsupported codecs or excessive HTML or JSON nesting can abort an individual run. This issue is fixed in versions 1.107.6 and 2.44.0.

  16. CVE-2026-107289 6.8 Moyenne

    pydantic-ai · pydantic-ai-slim · Pydantic pydantic ai

    Pydantic AI is a Python agent framework for building applications and workflows with Generative AI. From 1.56.0 until 1.107.6 and 2.44.0, applications that opt attacker-influenced URLs into local network access through FileUrl with force_download='allow-local' or web_fetch_tool with allow_local_urls=True can bypass the cloud-metadata blocklist by appending an IPv6 zone identifier to an IPv6 metadata address. IPv6Address equality and hashing include the zone identifier, so the blocklist comparison fails even though the network stack ignores the zone on a non-link-local destination and reaches the metadata service, potentially exposing cloud IAM credentials. The opt-in settings are disabled by default, and the issue requires an IPv6-enabled environment. This issue is fixed in versions 1.107.6 and 2.44.0.

  17. CVE-2026-107288 3.7 Faible

    pydantic-ai · pydantic-ai-slim · Pydantic pydantic ai

    Pydantic AI is a Python agent framework for building applications and workflows with Generative AI. From 1.77.0 until 1.107.6 and 2.44.0, the local web_fetch_tool and the WebFetch local fallback compare blocked_domains entries with a URL hostname before both values are normalized to the form used by getaddrinfo. An attacker-influenced model can use an equivalent IDNA spelling, non-ASCII label separator, case variation, or trailing root label that resolves to a blocked host but does not match the configured string, causing the application to fetch that host with its own privileges. allowed_domains fails closed for unmatched spellings, and private-IP and cloud-metadata protections remain effective. This issue is fixed in versions 1.107.6 and 2.44.0.

  18. CVE-2026-107287 6.5 Moyenne

    pydantic-ai · pydantic-ai-slim · Pydantic pydantic ai

    Pydantic AI is a Python agent framework for building applications and workflows with Generative AI. From 1.77.0 until 1.107.7 and 2.52.0, the local web_fetch_tool and the WebFetch local fallback can consume excessive CPU and memory during HTML-to-Markdown conversion of attacker-controlled HTML containing deeply nested block elements. Conversion repeatedly reprocesses accumulated text and can greatly expand intermediate output before the returned-content limit is applied, allowing a model-directed fetch to delay other work in the process. Provider-native web fetching is not affected. This issue is fixed in versions 1.107.7 and 2.52.0.

  19. CVE-2026-107286 7.5 Élevée

    pydantic-ai · pydantic-ai-slim · Pydantic pydantic ai

    Pydantic AI is a Python agent framework for building applications and workflows with Generative AI. From 2.10.0 until 2.53.0, streamed requests made through ConcurrencyLimitedModel or limit_model_concurrency can retain shared concurrency slots because anyio.CapacityLimiter associates an acquired slot with the borrowing task while streaming cleanup can run in a different task. Early stream termination, cancellation, consumer exceptions, or complete stream_text() consumption with debounce_by=0.1 can therefore leave capacity occupied, eventually preventing later requests that share the long-lived limiter from proceeding and causing a denial of service. Agent-level max_concurrency and non-streaming model requests are not affected. This issue is fixed in version 2.53.0.

  20. CVE-2026-62176 9.1 Critique

    MervinPraison PraisonAI

    PraisonAI is a multi-agent teams system. Prior to version 4.6.78, the `deploy/api.py` module generates Python server code by directly interpolating the `agents_file` parameter into an f-string that is then written to a file and executed via `subprocess.Popen()`. An attacker who controls the `agents_file` value (via CLI argument, configuration, or upstream API) can inject arbitrary Python code. Version 4.6.78 patches the issue.

  21. CVE-2026-20032 4.4 Moyenne

    Cisco NX-OS Software

    A vulnerability in the Python interpreter of Cisco NX-OS Software could allow an authenticated, local attacker with low privileges to escape the Python sandbox and gain unauthorized access to the underlying operating system of an affected&nbsp;device. This vulnerability is due to insufficient validation of user-supplied input. An attacker could exploit this vulnerability by manipulating specific functions within the Python interpreter. A successful exploit could allow an attacker to escape the Python sandbox and execute arbitrary commands on the underlying operating system with the privileges of the authenticated user.

  22. CVE-2026-107204 9.8 Critique

    LMCache

    LMCache through 0.5.5 contains an unauthenticated remote code execution vulnerability that allows remote attackers to execute Python code by posting scripts to the /run_script endpoint. Attackers can recover real builtins through the injected FastAPI app object, bypassing the guarded __import__, to import os and run operating system commands as the LMCache process.

  23. CVE-2026-106558 8.8 Élevée

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

    Backstage is an open framework for building developer portals. Prior to 1.14.8, 1.15.6, and 2.0.1, the @backstage/plugin-techdocs-node package improperly validated mapping-style markdown_extensions configuration. An authenticated attacker who can register or influence an SCM-backed documentation source may bypass TechDocs sanitization and cause Python objects to be imported and instantiated in the generator runtime, leading to arbitrary code execution. Earlier fixes in versions 1.14.6 and 1.15.4 did not fully address the supported mapping representation of markdown_extensions. Impact is greatest when documentation generation runs with backend credentials, filesystem access, or internal network access. This issue is fixed in versions 1.14.8, 1.15.6, and 2.0.1.

  24. CVE-2026-105812 9.0 Critique

    aws bedrock-agentcore-starter-toolkit

    Improper control of code generation in the agent import functionality of Amazon Bedrock AgentCore Starter Toolkit before 0.3.14 might allow an authenticated same-account actor to execute arbitrary code when a user imports and runs or deploys a Bedrock Agent, via crafted configuration values incorporated into generated Python source without safe literal encoding. To remediate this issue, users should upgrade to version 0.3.14. Because this issue persists into generated source, upgrading alone is not sufficient: agents imported with an affected version must be re-imported with version 0.3.14 or later and their local and deployed output artifacts replaced.

  25. CVE-2026-106454 4.3 Moyenne

    twisted

    Twisted is an event-based framework for internet applications, supporting Python 3.6+. In 25.5.0 and earlier, wildcardToRegexp() in twisted/mail/imap4.py translates the IMAP asterisk and percent wildcards but passes all other characters from an authenticated client's LIST or LSUB pattern directly to re.compile(), allowing nested or otherwise expensive regular expression constructs to cause catastrophic backtracking when matched against mailbox names. Because Twisted uses a cooperative single-threaded reactor, the blocking match suspends all server input and output for the duration of the match. No fixed release is available as of this review.

  26. CVE-2026-106441 7.8 Élevée

    hydra-ecosystem hydra

    Hydra is a framework for elegantly configuring complex applications. Prior to 1.3.6 and 1.4.0.dev9, Hydra passes Python logging configuration to logging.config.dictConfig() without applying Hydra's target policy to handler class values or formatter, filter, handler, queue, and listener factories. An attacker who controls Hydra logging configuration can therefore select an importable class or factory and cause it to be invoked with the application's privileges, even in versions where instantiate() is protected because the logging path does not use instantiate(). This issue is fixed in versions 1.3.6 and 1.4.0.dev9.

  27. CVE-2026-106440 7.8 Élevée

    hydra-ecosystem hydra

    Hydra is a framework for elegantly configuring complex applications. From 1.2.0 until 1.3.0 and 1.4.0.dev10, the hydra-optuna-sweeper package accepts a configuration-controlled dotted path in hydra.sweeper.custom_search_space, resolves it with hydra.utils.get_method(), and later invokes the returned callable in the Hydra controller process. Because get_method() is a trusted-input lookup helper and does not apply the execution policy used by instantiate(), an attacker who controls Optuna sweep configuration or command-line overrides can select importable Python code for execution with the application's privileges, including bypassing a trusted execution whitelist on affected Hydra 1.4 development releases. This issue is fixed in versions 1.3.0 and 1.4.0.dev10.

  28. CVE-2026-106439 8.5 Élevée

    hydra-ecosystem hydra

    Hydra is a framework for elegantly configuring complex applications. From 1.3.4 until 1.3.7 and 1.4.0.dev10, Hydra stores legacy instantiate target blocklists and related execution-policy collections in mutable module-level state. An attacker who controls multiple sibling target entries can resolve hydra._internal.target_policy.UNCONTROLLED_EXECUTION_TARGETS.discard through instantiate(), remove a denied target, and then invoke that target because sibling nodes are processed in insertion order against the same modified policy. The mutation persists in process-global state and can enable code execution with the application's privileges, while a narrow execution whitelist supplied by trusted Python code is not bypassed by the reported direct mutation path. This issue is fixed in versions 1.3.7 and 1.4.0.dev10.

  29. CVE-2026-105801 8.4 Élevée

    openapi-generators openapi-python-client

    openapi-python-client generates Python clients from OpenAPI documents. Prior to 0.29.1, the generator does not safely neutralize malicious OpenAPI document content before rendering string, docstring, and f-string contexts in generated Python. The generated Python client can contain attacker-controlled Python that executes when a user imports the client, affecting the importing environment's integrity and potentially its confidentiality and availability. This issue is fixed in version 0.29.1.

  30. CVE-2026-105782 7.5 Élevée

    scrapy

    Scrapy is a high-level web crawling and scraping framework for Python. From 1.4.0 until 2.14.2, RefererMiddleware in scrapy/spidermiddlewares/referer.py treated a Referrer-Policy response-header value that resembled a Python import path as a referrer policy class, imported the referenced object, and called it. A malicious website could supply a callable such as sys.exit and terminate a crawler processing the response. This issue is fixed in version 2.14.2.

  31. CVE-2026-105757 6.5 Moyenne

    vllm-project vllm · Vllm vllm

    vLLM is an inference and serving engine for large language models. Prior to 0.30.0, structured-output request failures can escape request-scoped validation and reach the EngineCore fatal-error path. A per-request backend mismatch can re-raise a grammar compilation exception, padding produced by the ngram_gpu speculative-decoding mode can pass a negative token to guidance validation, and the Rust frontend can admit empty structured-output values that the Python frontend rejects, allowing ordinary constrained-generation requests to terminate the shared engine. This issue is fixed in version 0.30.0.

  32. CVE-2026-105315 4.7 Moyenne

    django-haystack

    A vulnerability has been found in django-haystack up to 3.3.0. Affected is the function _to_python of the file haystack/backends/elasticsearch_backend.py of the component more_like_this Template Tag Handler. Such manipulation of the argument result_class leads to improper neutralization of directives in dynamically evaluated code. The attack can be launched remotely. The exploit has been disclosed to the public and may be used. Upgrading to version 3.4.0 is able to address this issue. The name of the patch is eb05f193c9771a68dcc8cfac6674a0d48a52ee9d. It is suggested to upgrade the affected component.

  33. CVE-2026-105314 7.5 Élevée

    Papermerge

    Papermerge 3.5.3 allows remote code execution by a standard user via directory traversal in a /api/documents/upload call. A Python .pth file can be written to site-packages, and its code is executed upon the next start of the Python interpreter.

  34. CVE-2026-105245 3.7 Faible

    sgl-project sglang

    A vulnerability has been found in sgl-project sglang up to 0.5.21. This issue affects the function server_info of the file python/sglang/srt/entrypoints/http_server.py of the component HTTP Endpoint. Such manipulation of the argument api_key leads to cleartext transmission of sensitive information. It is possible to launch the attack remotely. Attacks of this nature are highly complex. The exploitability is assessed as difficult. The exploit has been disclosed to the public and may be used. The pull request to fix this issue awaits acceptance.