Schwachstellen im Stack Python · Seite 32

MLFLOW_ALLOW_PICKLE_DESERIALIZATION=False safety control bypassed by mlflow.statsmodels flavor — RCE via crafted model artifact

## Summary MLflow introduced `MLFLOW_ALLOW_PICKLE_DESERIALIZATION` as a security control to prevent unsafe `pickle.load` execution during model loading, in response to CVE-2024-37052 through CVE-2024-37060. When set to `False`, operators expect all pickle deserialization to be blocked. The most recent related fix (#21188) patched a bypass in the pyfunc flavor. However, the `mlflow.statsmodels` flavor completely omits this guard. An attacker who places a crafted MLmodel artifact into any accessible artifact store can trigger arbitrary code execution on any process that calls `mlflow.pyfunc.load_model()` against the malicious model — **even w

• Auf 3.15.0 aktualisieren

Hugging Face Transformers library writes remote code to disk prior to consent check

A vulnerability in Hugging Face Transformers (versions 4.57.0 to 5.16.1) allows remote Python files to be written to local disk without user consent when using GenerativePreTrainedModel.load_custom_generate(). The function fetches and caches a remote module file before performing the required trust_remote_code consent check, inverting the security model enforced by other code-loading paths (such as AutoConfig, AutoModel, and AutoTokenizer). As a result, attacker‑controlled Python code from custom_generate/generate.py is copied into the user’s ~/.cache/huggingface/modules directory even if the user declines the trust prompt. Although execution

pypdf: Inefficient handling of non-whitespace inputs in read_until_whitespace

pypdf is a free and open-source pure-python PDF library. Prior to 6.15.0, an attacker can craft a PDF that causes long runtimes when the pypdf/_utils.py function read_until_whitespace reads a stream containing a long run of bytes without whitespace. The function repeatedly performs immutable bytes concatenation in a one-byte loop, causing quadratic processing cost for the long non-whitespace input. This issue is fixed in version 6.15.0.

• Auf 6.15.0 aktualisieren
• Auf 6.19.0-1 aktualisieren

Tornado: Urlencoded body parsing omits max_num_fields, so one request can stall the event loop

Tornado is a Python web framework and asynchronous networking library. Prior to 6.5.8, Tornado parses application/x-www-form-urlencoded request bodies with urllib.parse.parse_qs in tornado/escape.py without passing max_num_fields. RequestHandler._execute in tornado/web.py parses the body before handler dispatch through HTTPServerRequest._parse_body and parse_body_arguments in tornado/httputil.py, so an unauthenticated request body containing millions of separator-delimited fields can synchronously stall the single-threaded event loop and delay every connection. The body is bounded only by max_buffer_size, which defaults to 104857600 bytes. Th

• Auf 6.5.8 aktualisieren

AIIR verification and policy gates could report success without enforcing the control (fail-open)

### Summary Several of AIIR's verification and policy paths could return a success/"verified" result without actually enforcing the control they represent — they could **fail open** rather than fail closed. For a tool whose purpose is trustworthy verification, a consumer relying on these gates may have treated unverified or non-conforming input as verified. Found during an internal adversarial hardening review of AIIR (not a third-party audit). All paths are fixed in **1.7.0**. ### Affected paths - A `require_signing` policy gate could be satisfied by a forgeable/empty field, so an unsigned or forged-bundle receipt could pass a "signing req

• Auf 1.7.0 aktualisieren

plone.app.event: Denial of service via iCalendar import

plone.app.event provides the event content type for Plone. Prior to versions 5.2.4 and 6.0.1, the iCalendar import in src/plone/app/event/ical/importer.py accepts insufficiently restricted calendar and event URLs, does not adequately bound downloaded bytes or imported events, and commits work per event. A logged-in editor can make the server request internal network resources or local calendar files, exhaust resources and take the site offline, and store a malicious event URL that executes script in another user's browser. The fix restricts accepted URLs, applies MAXIMUM_ICAL_IMPORT_SIZE_BYTES and MAXIMUM_ICAL_IMPORT_EVENTS limits, uses trans

• Auf 5.2.4 aktualisieren
• Auf 6.0.1 aktualisieren

Protego: Exponential backtracking ReDoS in robots.txt URL wildcard matching

Protego is a pure-Python robots.txt parser with support for modern conventions. Prior to 0.6.2, protego._urlpattern._URLPattern._prepare_pattern_for_regex translates every asterisk in an Allow or Disallow directive into a lazy regular-expression wildcard, so a directive containing many asterisks creates exponential backtracking. After protego.Protego.parse processes a crafted robots.txt file, protego.Protego.can_fetch can spend an attacker-controlled period matching a near-miss URL and deny service to the crawler. The vulnerable path is src/protego/_urlpattern.py in the _URLPattern match logic. This issue is fixed in version 0.6.2.

• Auf 0.6.2 aktualisieren
• Auf 0.6.2+dfsg-1 aktualisieren

plone.app.portlets: Denial of service via RSS feed portlet

plone.app.portlets provides portlets and a Plone-specific user interface for plone.portlets. Prior to 5.0.8, 6.0.4, and 7.0.2, a member who can add an RSS portlet can set its feed URL to a very large response, causing src/plone/app/portlets/portlets/rss.py to download and retain excessive data in memory and deny service. The same RSS URL handling accepts internal hosts, IP addresses, single-word domains, and explicit ports, allowing server-side requests that can probe internal network services and open ports. A malicious feed item can also supply a JavaScript URL that is retained as the item link and can execute script when used by a victim.

• Auf 5.0.8 aktualisieren
• Auf 6.0.4 aktualisieren
• Auf 7.0.2 aktualisieren

Piccolo Admin: Privilege escalation - admin to superuser via session-token disclosure in GET /api/tables/sessions/.

Piccolo Admin is an admin interface and content management system for Python, built on top of Piccolo. Prior to 1.14.0, piccolo_admin/endpoints.py uses superuser_validators to block PUT, PATCH, DELETE, and POST requests by non-superusers but permits GET requests to configured user and session tables, while piccolo_api/session_auth/tables.py exposes SessionsBase.token because the token column is not secret. In deployments that add the Sessions and User tables to create_admin, a non-superuser administrator can call GET /api/tables/sessions/, obtain another user's live session token, replay it as the Cookie id value to impersonate a superuser, a

• Auf 1.14.0 aktualisieren

WsgiDAV: Blind SQL injection in the MySQL provider

WsgiDAV is a generic and extendable WebDAV server based on WSGI. Prior to 4.3.5, the sample MySQLBrowserProvider in wsgidav/samples/mysql_dav_provider.py concatenates the record key parsed from a request URL directly into SQL WHERE clauses. The affected _exists_record_by_primary_key, _get_field_by_primary_key, and _get_record_by_primary_key methods are part of a shipped example provider that is not enabled by default. An attacker who can access a share explicitly configured with this non-default provider can inject SQL through a normal GET request; anonymously exposed read shares permit a status-code oracle and extraction of arbitrary data re

• Auf 4.3.5 aktualisieren

CVE-2026-79720

Reflected XSS in Netron versions <=9.1.2 on desktop application through unsanitized node names allows an attacker to hide certain nodes, perform port scanning or abuse a Chrome n-day to achieve Remote Code Execution.

• Auf 9.1.3 aktualisieren

CVE-2026-79719

Reflected XSS in Netron versions <=9.1.2 on desktop application through unsanitized node names allows an attacker to hide certain nodes, perform port scanning or abuse a Chrome n-day to achieve Remote Code Execution.

• Auf 9.1.3 aktualisieren

CVE-2026-79718

Reflected XSS in Netron versions <=9.1.2 on desktop application through unsanitized node names allows an attacker to hide certain nodes, perform port scanning or abuse a Chrome n-day to achieve Remote Code Execution.

• Auf 9.1.3 aktualisieren

NLTK before 3.10.3 Hardlink File Overwrite via downloader

NLTK versions before 3.10.3 contain a filesystem containment bypass vulnerability in the Downloader.download and Downloader.incr_download methods that allows attackers to overwrite files outside the install root through pre-existing hardlinks. Attackers with write access to a shared downloader directory can create hardlinks pointing to outside-root files that are then overwritten during normal package extraction, mutating files outside the intended install tree.

• Auf 3.10.3 aktualisieren
• Auf 3.10.3-1 aktualisieren

NLTK through 3.10.3 Path Traversal via Model-Artifact APIs

NLTK through 3.10.3 contains a path traversal vulnerability in model-artifact APIs that bypass pathsec enforcement by using raw file operations on caller-controlled paths. Attackers can read or write files outside allowed sandbox roots through TransitionParser, AveragedPerceptron, PerceptronTagger, and maxent parameter APIs when pathsec is enabled.

• Auf 3.10.3 aktualisieren

NLTK before 3.10.3 Regular Expression Denial of Service via Pl196xCorpusReader

NLTK before 3.10.3 contains a regular expression denial of service vulnerability in Pl196xCorpusReader that allows attackers to cause quadratic CPU consumption by supplying malformed TEI blocks with many unmatched opening tags. Attackers can exploit lazy regex patterns in the read_block method through public APIs like words() and tagged_words() to force repeated rescans and achieve near-quadratic runtime growth.

• Auf 3.10.3 aktualisieren
• Auf 3.10.3-1 aktualisieren

NLTK before 3.10.3 Denial of Service via Uncontrolled Recursion

NLTK before 3.10.3 contains an uncontrolled recursion vulnerability in nltk.featstruct.FeatStructReader that allows unauthenticated attackers to cause a denial of service by supplying deeply nested feature-structure input. Attackers can craft trivial payloads with nested brackets that exceed Python's recursion limit and raise an unhandled RecursionError, crashing applications that parse user-supplied feature structures or feature grammars.

• Auf 3.10.3 aktualisieren
• Auf 3.10.3-1 aktualisieren

NLTK before 3.10.3 Quadratic CPU Exhaustion via XMLCorpusView

NLTK versions before 3.10.3 contain a quadratic CPU exhaustion vulnerability in XMLCorpusView._read_xml_fragment() that rescans accumulated XML fragments on every 1 KiB block read. Attackers can provide malformed XML corpus files to cause severe CPU consumption and denial of service through affected readers like BNCCorpusReader.

• Auf 3.10.3 aktualisieren
• Auf 3.10.3-1 aktualisieren

nltk PorterStemmer before 3.10.3 Quadratic-time DoS

nltk PorterStemmer in versions <= 3.10.2 (fixed in 3.10.3) contains an inefficient-algorithmic-complexity denial of service in PorterStemmer.stem(). The _is_consonant() helper walks backward over the entire run of trailing 'y' characters on every call, and _measure() invokes it for each stem position, causing O(n^2) behavior. A single ~20-50 KB untrusted token consisting of a long run of the letter 'y' followed by a matching suffix (e.g., 'ness') can pin a CPU core for seconds to minutes, causing availability impact.

• Auf 3.10.3 aktualisieren
• Auf 3.10.3-1 aktualisieren

openssl_encrypt before 1.4.9 Denial of Service via KDF

openssl_encrypt before 1.4.9 fails to validate KDF cost parameters in encrypted file metadata and keystore headers, allowing attackers to trigger unbounded memory allocation. Attackers can craft malicious encrypted files declaring arbitrarily large Argon2, scrypt, or balloon KDF parameters to exhaust system memory and crash the process without authentication.

• Auf 1.4.9 aktualisieren