PraisonAI: SSRF via redirect-following in praisonaiagents web_crawl
PraisonAI is a multi-agent teams system. Prior to praisonaiagents 1.6.58, the web_crawl function validates only the initial URL before _crawl_with_httpx uses httpx.Client(follow_redirects=True). Redirect targets are not revalidated, so an attacker who influences a crawl target can redirect a public URL to loopback, private network, or cloud metadata services while ALLOW_LOCAL_CRAWL remains disabled. The fetched internal response is returned to the agent context. This issue is fixed in version 1.6.58.
NLTK before 3.10.3 Remote Code Execution via Unsafe Pickle Deserialization
NLTK versions before 3.10.3 contain a remote code execution vulnerability in allowlisted pickle loaders that trust entire module namespaces instead of specific safe callables. Attackers can craft malicious pickle payloads invoking dangerous in-namespace functions like ReppTokenizer._execute and numpy.f2py.crackfortran.myeval through pickle REDUCE to execute arbitrary commands during model or tokenizer artifact loading.
vLLM before 0.27.0 Denial of Service via DeepStream Backend
vLLM before 0.27.0 fails to properly classify DeepStream as a GPU backend and omits pixel-limit enforcement in its decode path. Unauthenticated attackers can activate DeepStream at request time to initialize the process-wide GPU decode pool and submit video that bypasses resource controls, causing partial denial of service for concurrent requests.
NLTK before 3.10.0 Remote Code Execution via Unsafe Pickle Deserialization
NLTK before 3.10.0 (affected versions <=3.9.4) contains an unsafe pickle deserialization vulnerability in the TransitionParser.parse() method (nltk/parse/transitionparser.py). The method calls pickle_load() with the default restricted=False, routing deserialization through WarningUnpickler, which does not override find_class() and therefore permits arbitrary class resolution. When an application loads an attacker-crafted model file, embedded pickle gadget chains execute arbitrary Python code with the privileges of the user running the application. NLTK provides a RestrictedUnpickler for safe deserialization, but it is not used by production c
NLTK before 3.10.3 SSRF Protection Bypass via Proxy
NLTK before 3.10.3 contains a server-side request forgery vulnerability in nltk.pathsec.urlopen (and callers nltk.data.load, nltk.downloader.Downloader.index/download) when an HTTP proxy is configured. pathsec.urlopen validates the requested hostname locally, but proxy-handler inheritance disables the safe HTTP/HTTPS handlers so the actual fetch is performed by the proxy against a destination that is never re-validated. An attacker can supply a validated public URL that the proxy forwards to an internal loopback-only service, allowing disclosure of internal HTTP resources, loading of forged downloader indexes, and installation of attacker-cho
NLTK before 3.10.3 Entity Expansion DoS via ElementTree
NLTK versions before 3.10.3 use xml.etree.ElementTree to parse XML in multiple modules, which honors entity declarations in document DTDs. Attackers can craft XML payloads with nested entity declarations that expand from hundreds of bytes to megabytes in memory, causing denial of service.
NLTK before 3.10.3 Arbitrary Code Execution via Graphviz dot Binary
NLTK versions before 3.10.3 fail to use validated absolute paths when invoking the Graphviz dot binary in dependencygraph.dot2img and AlignedSent._repr_svg_, allowing attackers to execute arbitrary code by placing a malicious dot binary in the search path or current working directory. Attackers can exploit bare-name binary resolution on Windows via the current working directory or on Unix-like systems via relative PATH entries to execute their binary instead of the legitimate Graphviz tool.
GitPython before 3.1.59 Arbitrary File Read via TagReference.create
GitPython before 3.1.59 contains an arbitrary file read vulnerability in TagReference.create() where a positional reference parameter bypasses the unsafe option guard. Attackers can supply a reference value like --file=<path> to read arbitrary files, with contents returned in the annotated tag message.
GitPython before 3.1.59 Arbitrary File Read via Repo.blame()
GitPython versions before 3.1.59 contain an incomplete denylist in the unsafe_git_revision_options guard that omits --contents and -S options, allowing attackers to read arbitrary files by passing these options to Repo.blame(). Attackers can supply revision values like --contents=/etc/passwd to leak file contents through the blame result returned to the caller.
GitPython before 3.1.59 Path Traversal via separate-git-dir
GitPython before 3.1.59 omits --separate-git-dir from unsafe_git_clone_options, allowing attackers to create arbitrary git directories outside the intended clone destination. Attackers can pass a separate_git_dir parameter to Repo.clone_from() or Repo.clone() to redirect repository metadata to an attacker-controlled filesystem path, enabling arbitrary directory creation and potential hook execution.
GitPython before 3.1.59 Remote Code Execution via Config Injection
GitPython before 3.1.59 fails to safely re-serialize multi-line git-config values during write operations, corrupting dormant quoted values into injected directives like core.hooksPath. Attackers can craft config files with embedded newlines that become live git directives after any unrelated GitPython config write, enabling arbitrary code execution via hook invocation.
GitPython before 3.1.59 Local File Content Disclosure via .gitmodules
GitPython before 3.1.59 fails to disable merge_includes when parsing .gitmodules, allowing attackers to disclose local file content by including arbitrary file paths via [include] directives. Attackers can craft a malicious .gitmodules file with include directives pointing to sensitive files; when repo.submodules is accessed, GitConfigParser raises MissingSectionHeaderError embedding the target file's first line verbatim in the exception message.
Wagtail: Improper restriction handling on Pages admin API
Wagtail is an open source content management system built on Django. Prior to versions 7.0.9, 7.3.4, 7.4.3, and 8.0rc2 on their respective release lines, the internal Pages admin API returns page fields declared in api_fields without sufficient access control, allowing a user with Wagtail admin access to retrieve restricted draft and live page content. This issue is fixed in versions 7.0.9, 7.3.4, 7.4.3, and 8.0rc2.
Mistune has Denial of Service — RecursionError via Excessive Emphasis Markers in Markdown
Mistune is a Python Markdown parser with renderers and plugins. Versions 3.3.0 through 3.3.2 are vulnerable to DoS through deeply nested tokens. HTML rendering creates deeply nested emphasis tokens from consecutive asterisk characters, and recursive rendering in HTMLRenderer.render_token() can exceed Python's recursion limit and raise RecursionError, allowing crafted Markdown to crash a parsing process. This issue is fixed in version 3.3.3
justhtml before 1.10.0 Denial of Service via deeply nested HTML
justhtml through 1.9.1 (fixed in 1.10.0) is vulnerable to uncontrolled recursion leading to denial of service. During JustHTML() construction, TreeBuilder.finish() unconditionally calls _populate_selectedcontent(), which recursively traverses the DOM tree via _find_elements()/_find_element() without a depth bound. An attacker who can supply HTML for parsing can provide deeply nested elements (e.g., ~1000 nested <div> tags, roughly 11 KB) to exceed CPython's default recursion limit and trigger an unhandled RecursionError, which may abort parsing, fail requests, or terminate a worker/process depending on the host application's exception handlin
justhtml before 1.12.0 Mutation XSS via Raw Text Elements
justhtml before 1.12.0 (versions <= 1.11.0) contains a mutation cross-site scripting (mXSS) vulnerability in the serialization of raw-text elements such as <style> and <script>. When a DOM tree is processed by sanitize_dom() using a custom policy that keeps these elements, text nodes inside them are serialized literally without escaping, allowing attacker-controlled text containing the matching closing tag sequence to break out of the raw-text context and inject arbitrary HTML into the serialized output. The default sanitization policy is not affected because it drops the contents of style and script.
justhtml before 1.12.0 Sanitizer Bypass via Markdown
justhtml versions <= 1.11.0 (fixed in 1.12.0) do not sufficiently escape HTML-significant characters (angle brackets) in text nodes when converting a parsed document to Markdown via to_markdown(). While a small set of Markdown metacharacters are escaped, characters such as < and > are preserved, so untrusted input that is safe in to_html() — including entity-decoded text (e.g. <script>) or text from RCDATA/RAWTEXT-parsed elements like <title>, <textarea>, <noscript>, and <plaintext> — can be emitted as raw HTML in the Markdown output, enabling a sanitizer bypass and potential cross-site scripting when that output is rendered.
justhtml before 1.16.0 Multiple Security Issues via Sanitization
justhtml before 1.16.0 contains multiple HTML sanitization bypass issues that can allow active/dangerous content (e.g., script or style) to survive sanitization, potentially leading to cross-site scripting. The issues primarily affect advanced usage rather than the default JustHTML(..., sanitize=True) path for ordinary parsed HTML: mutating or reusing sanitization policy objects (including exported defaults) could weaken later sanitization; programmatic DOM input to sanitize()/sanitize_dom() could miss mixed-case tag names (e.g., ScRiPt, StYlE); crafted programmatic doctype names could serialize into active markup; and custom policies preserv
justhtml 0.9.0 through 1.21.0 Cross-Site Scripting via code-span
justhtml versions 0.9.0 through 1.21.0 contain a cross-site scripting vulnerability in to_markdown() where inline code spans fail to account for blank lines as block boundaries. Attackers can inject blank lines into code or pre element text to break the inline span, causing sanitized HTML to be emitted unescaped and re-parsed as live Markdown by compliant renderers.
justhtml before 1.17.0 Multiple Cross-Site Scripting Vulnerabilities
justhtml before 1.17.0 contains multiple security issues in sanitization, serialization, and programmatic DOM handling. When custom policies preserve foreign namespaces (SVG/MathML), dangerous content such as HTML integration points (SVG <foreignObject>, MathML <annotation-xml encoding="text/html">) and mutation-XSS parser-differential payloads could survive sanitization and become active HTML after reparse; SVG filter="url(...)" and preserved <style> could leave resource-loading CSS (@import, background-image:url()). Programmatic script/style/Comment nodes could serialize into active markup. Additional hardening fixes address sanitize-pipeli