Open WebUI: DNS Rebinding SSRF Bypass

SummaryOpen WebUI vetted user-supplied URLs by resolving the hostname once and rejecting private, loopback and link-local addresses, then let the HTTP client resolve that hostname again at connect time. An attacker who controls the authoritative DNS for a hostname they submit can answer with a...

Open WebUI: Deletion of directories and file embeddings in other knowledge bases via sync cleanup

SummaryA user with write access to one knowledge base could delete directories, and drop file embeddings, belonging to knowledge bases they do not control. The sync cleanup endpoint verified write access on the knowledge base named in the URL and then acted on the directory and file ids supplied ...

Open WebUI: Account takeover via OAuth token exchange accepting tokens issued to any client

SummaryThe OAuth token exchange endpoint accepts a raw provider access token and validates it by calling the provider's userinfo endpoint. A userinfo endpoint reports only that a token is valid, never which OAuth client it was issued to, and the endpoint performed no audience or client check of i...

Natural Language Toolkit (NLTK): ReDoS in NLTK ReviewsCorpusReader FEATURES regex

SummaryReviewsCorpusReader extracts feature annotations of the form label followed by a bracketed signed digit e.g. a label then +2 from each review line, using the module-level FEATURES regex. The feature-label sub-pattern is unbounded — an optional greedy run of word-plus-whitespace groups...

AIOHTTP: WebSocket client accepts compressed frames without negotiated permessage-deflate

SummaryThe client accepts and decompresses frames with the RSV1 bit set even when the permessage-deflate extension was not negotiated. ImpactA client may unexpectedly decompress WebSocket frames when explicitly opted out. This could lead to additional CPU/memory consumption, but is unlikely to be...

`datamodel-code-generator` vulnerable to code injection via unescaped carriage return in `--extra-template-data` `comment` field

Summarydatamodel-code-generator is vulnerable to code injection when a developer passes an --extra-template-data file whose comment value contains a literal \r carriage return. The comment variable is rendered into a Python comment in six built-in templates with no line-terminator escaping...

datamodel-code-generator vulnerable to SSRF protection bypass via DNS rebinding

Summarydatamodel-code-generator's anti-SSRF guard validates the resolved IP of a fetch target once and then lets httpx perform its own independent DNS resolution to connect, so the validated address is never pinned. A hostname that resolves to a public IP at validation time and a private IP at...

Open WebUI: Arena task endpoints can bypass underlying model access controls

SummaryAn authenticated non-admin user with read access to an arena wrapper model can reach a restricted underlying model through task endpoints such as /api/v1/tasks/moa/completions.The normal chat route resolves arena models before the final chat dispatch and therefore re-checks the selected...

Open WebUI: Upload `metadata.knowledge_id` bypasses the knowledge-base write-access check (read-only users can add files to KB)

Open WebUI upload metadata can add files to knowledge bases without write permission SummaryOpen WebUI's file upload background processing trusts the client-supplied metadata.knowledgeid value and inserts a knowledgefile association before validating that the uploading user has write access to th...

Open WebUI: Scheduled automations continue after pending-user deactivation and stored model ACL revocation

Title: Scheduled automations continue after pending-user deactivation and stored model ACL revocation SummaryOpen WebUI documents pending as a zero-access role used for new sign-ups and deactivated users, and normal HTTP routes enforce that with getverifieduser which rejects pending, while...

PYSEC-2026-3624

PyTorch Lightning through 2.6.5, fixed in commit d710d68, contains a remote code execution vulnerability in the loadstate function that imports and executes attacker-controlled module names from checkpoint instantiator hyperparameters. Attackers can craft malicious checkpoint files that bypass...

PYSEC-2026-3450

Ollama downloadBlob Improper Validation of Array Index Denial-of-Service Vulnerability. This vulnerability allows remote attackers to create a denial-of-service condition on affected installations of Ollama. Authentication is not required to exploit this vulnerability.The specific flaw exists...

BabelDOC: Arbitrary Code Execution via CMap Pickle Deserialization in babeldoc/pdfminer/cmapdb.py

Arbitrary Code Execution via CMap Pickle Deserialization in babeldoc/pdfminer/cmapdb.py SummaryBabelDOC's vendored PDF parser babeldoc/pdfminer/cmapdb.py deserializes untrusted pickle data when loading CMap files. The loaddata method strips only NUL bytes from a PDF-controlled CMap name, then...

psd-tools vulnerable to arbitrary file write via smart-object filename

psd-tools: arbitrary file write/read via smart-object path traversal SummaryIn psd-tools all releases exposing the SmartObject API through v1.17.0, SmartObject.save writes an embedded smart object to a path taken verbatim from the PSD file. Because that name is attacker-controlled and unsanitised...

pymonocypher: Potential heap buffer overflow on nb_blocks in argon2i_32 when provided buffer is too small

ImpactThe argon2i32 implementation does not check the nbblocks size. If the caller does not provide a sufficiently large buffer based on the API contract, then argon2i32 will write past the end of the buffer and possibly corrupt the heap. PatchesFixed in 4.0.2.8, which now verifies that nbblocks ...

pyLoad: Unbounded Memory Growth Leading to DoS and Potential DDoS in EventManager

Description:The EventManager module in pyload manages a list of Client instances for subscribing to events. The addition of each unique uuid from the getevents API causes the creation of a Client instance that gets appended to the clients list. Although there is a clean method available in the...

Langroid: Sandbox Escape to Remote Code Execution via Incomplete `eval()` Mitigation in TableChatAgent

Advisory DetailsTitle: Sandbox Escape to Remote Code Execution via Incomplete eval Mitigation in TableChatAgentDescription: SummaryLangroid is vulnerable to a critical Sandbox Escape leading to Remote Code Execution RCE in its TableChatAgent and VectorStore capabilities. When these agents evaluat...

Langroid: Neo4jChatAgent executes LLM-generated Cypher without validation (prompt-to-Cypher injection; config-conditional RCE), mirroring the SQLChatAgent bug fixed in CVE-2026-25879

Neo4jChatAgent passes LLM-generated Cypher queries straight to the Neo4j driver with no validation, no statement-type allowlist, and no opt-out gate. The query text is influenceable by prompt injection direct user input or indirect content the agent reads back via RAG, so an attacker who can...

Langroid: SQLChatAgent _validate_query blocklist misses pg_read_file family enabling arbitrary file read

SummarySQLChatAgent in langroid ships a validatequery defense-in-depth layerwhose DANGEROUSSQLPATTERNS regex blocklist enumerates dangerous SQLprimitives by specific function name. The list misses the canonicalPostgreSQL filesystem-disclosure family pgreadfile, pgstatfile,pglslogdir, pglswaldir,...

Improper neutralization of argument delimiters in AWS Bedrock AgentCore Python SDK install_packages()

SummaryThe AWS Bedrock AgentCore Python SDK bedrock-agentcore is an open-source SDK that enables developers to build, deploy, and manage agents on AWS Bedrock AgentCore. An issue exists in the installpackages method of the Code Interpreter client where crafted package name arguments can bypass...

Open WebUI: Stored XSS to Account Takeover via Model Profile Images

Stored XSS to Account Takeover via Model Profile Images in Open WebUIAffected: Open WebUI = 0.9.5Bypass of: GHSA-3wgj-c2hg-vm6q, GHSA-3856-3vxq-m6fc--- TL;DROpen WebUI patched SVG XSS in user profile images and webhook profile images but forgot to apply the same fix to model profile images. The...

praisonai-platform: Any workspace member can delete the entire workspace via DELETE /workspaces/{id}

SummaryType: Authorization bypass enabling destructive action. The DELETE /workspaces/workspaceid endpoint is gated only by requireworkspacememberworkspaceid default minrole="member". Any member of the workspace can issue a single DELETE to wipe the entire workspace, including every project, issu...

praisonai-platform: Issue endpoints accept any issue_id without workspace ownership check, cross-workspace read/update/delete IDOR

SummaryType: Insecure Direct Object Reference. The issue CRUD endpoints GET / PATCH / DELETE /workspaces/workspaceid/issues/issueid gate access on requireworkspacememberworkspaceid only, then resolve issueid through IssueService.getissueid which is a primary-key lookup with no workspace constrain...

praisonai-platform: Label endpoints' unchecked label_id/issue_id enable cross-workspace label IDOR (edit, delete, link)

SummaryType: Insecure Direct Object Reference. Five label endpoints — PATCH /workspaces/workspaceid/labels/labelid, DELETE .../labels/labelid, POST .../issues/issueid/labels/labelid, DELETE .../issues/issueid/labels/labelid, GET .../issues/issueid/labels — gate access on...

praisonai-platform: list_issue_activity returns activity log for any issue regardless of workspace ownership

SummaryType: Insecure Direct Object Reference. The GET /workspaces/workspaceid/issues/issueid/activity endpoint is gated by requireworkspacememberworkspaceid and dispatches to ActivityService.listforissueissueid, which executes SELECT FROM activity WHERE issueid = :issueid with no workspace...

PraisonAI CLI automatically resolves @url mentions in prompt text and can read loopback URLs into model context

SummaryPraisonAI's direct-prompt CLI automatically expands @url: mentions in raw prompt text before agent execution begins.If a prompt contains @url:, the CLI calls MentionsParser.process.... The @url: handler then performs a direct urllib.request.urlopen request to the attacker-controlled URL an...

OpenStack Neutron has an Incorrect Authorization issue

In OpenStack Neutron before 28.0.1, the tagging controller enforces plural policy action names on single-tag write operations while the defined policy rules use singular names. The mismatched names evaluate as allowed under the default policy, permitting a project reader to create and update tags...

hermes-agent has an Injection issue

A vulnerability was found in NousResearch hermes-agent 2026.4.23. The impacted element is the function scancontextcontent of the file agent/promptbuilder.py. The manipulation results in injection. The attack may be performed from remote. The exploit has been made public and could be used. The...

OpenStack Ironic has an Incorrect Resource Transfer Between Spheres

An issue was discovered in idrac in OpenStack Ironic before 35.0.1. During import, a user invoking molds can request authorization to be sent to a remote endpoint. The credential forwarded is a time-limited Keystone token which provides access to all OpenStack services Ironic is authorized for; o...

django-mdeditor is Missing Authentication for Critical Function

All versions of the package django-mdeditor are vulnerable to Missing Authentication for Critical Function in the image upload endpoint. An attacker can upload malicious files and achieve arbitrary code execution since this endpoint lacks authentication protection and proper sanitisation of file...

Graphiti vulnerable to Cypher Injection via unsanitized node_labels in search filters

SummaryGraphiti versions before 0.28.2 contained a Cypher injection vulnerability in shared search-filter construction for non-Kuzu backends. Attacker-controlled label values supplied through SearchFilters.nodelabels were concatenated directly into Cypher label expressions without validation.In M...

multipart vulnerable to ReDoS in `parse_options_header()`

SummaryThe parseoptionsheader function in multipart.py uses a regular expression with an ambiguous alternation, which can cause exponential backtracking ReDoS when parsing maliciously crafted HTTP or multipart segment headers. This can be abused for denial of service DoS attacks against web...

NVIDIA NeMo Framework Deserializes Untrusted Data

NVIDIA NeMo Framework contains a vulnerability where an attacker could cause remote code execution by convincing a user to load a maliciously crafted file. A successful exploit of this vulnerability might lead to code execution, denial of service, information disclosure, and data tampering...

Keras has a Local File Disclosure via HDF5 External Storage During Keras Weight Loading

SummaryTensorFlow / Keras continues to honor HDF5 “external storage” and ExternalLink features when loading weights. A malicious .weights.h5 or a .keras archive embedding such weights can direct loadweights to read from an arbitrary readable filesystem path. The bytes pulled from that path popula...

TensorFlow has Heap-buffer-overflow in AvgPoolGrad

Impactpythonimport osos.environ'TFENABLEONEDNNOPTS' = '0'import tensorflow as tfprinttf.versionwith tf.device"CPU": ksize = 1, 40, 128, 1 strides = 1, 128, 128, 30 padding = "SAME" dataformat = "NHWC" originputshape = 11, 9, 78, 9 grad = tf.saturatecasttf.random.uniform16, 16, 16, 16, minval=-128...

`CHECK_EQ` fail in `tf.raw_ops.TensorListResize`

ImpactIf tf.rawops.TensorListResize is given a nonscalar value for input size, it results CHECK fail which can be used to trigger a denial of service attack.pythonimport numpy as npimport tensorflow as tfa = datastructures.tftensorlistnewelements = tf.constantvalue=3, 4, 5b = np.zeros0, 2, 3,...

TensorFlow vulnerable to segfault in `QuantizedAvgPool`

ImpactIf QuantizedAvgPool is given mininput or maxinput tensors of a nonzero rank, it results in a segfault that can be used to trigger a denial of service attack.pythonimport tensorflow as tfksize = 1, 2, 2, 1strides = 1, 2, 2, 1padding = "SAME"input = tf.constant1, shape=1,4,4,2,...

PYSEC-2026-3448

Crawl4AI before 0.8.8 contains credential exfiltration vulnerabilities in the Docker API server that allow attackers to redirect LLM API calls to attacker-controlled endpoints and read arbitrary environment variables. Attackers can exploit the unauthenticated /md, /llm, and /llm/job endpoints by...

askbot inexhaustive permissions check allows any user to modify a different user's profile picture

All versions of askbot before and including 0.12.2 allow an attacker authenticated with normal user permissions to modify the profile picture of other application users. This issue affects askbot: 0.12.2...

GuardDog Path Traversal Vulnerability Leads to Arbitrary File Overwrite and RCE

SummaryA path traversal vulnerability exists in GuardDog's safeextract function that allows malicious PyPI packages to write arbitrary files outside the intended extraction directory, leading to Arbitrary File Overwrite and Remote Code Execution on systems running GuardDog.CWE: CWE-22 Improper...

AIOHTTP has unicode match groups in regexes for ASCII protocol elements

SummaryThe parser allows non-ASCII decimals to be present in the Range header. ImpactThere is no known impact, but there is the possibility that there's a method to exploit a request smuggling vulnerability.----Patch:...

Picklescan does not block ctypes

SummaryPicklescan doesnt flag ctypes module as a dangerous module, which is a huge issue. ctypes is basically a foreign function interface library and can be used to Load DLLs Call C functions directly Manipulate memory raw pointers.This can allow attackers to achieve RCE by invoking direct...

Weblate has an arbitrary file read via symbolic links

ImpactIt was possible to read arbitrary files from the server file system using crafted symbolic links in the repository. ResourcesThanks to Jason Marcello for responsible disclosure...

FastAPI SSP is vulnerable to Cross-site Request Forgery (CSRF) through improper OAuth parameter validation

Versions of the package fastapi-sso before 0.19.0 are vulnerable to Cross-site Request Forgery CSRF due to the improper validation of the OAuth state parameter during the authentication callback. While the getloginurl method allows for state generation, it does not persist the state or bind it to...

motionEye vulnerable to RCE via unsanitized motion config parameter

SummaryA command injection vulnerability in MotionEye allows attackers to achieve Remote Code Execution RCE by supplying malicious values in configuration fields exposed via the Web UI. Because MotionEye writes user-supplied values directly into Motion configuration files without sanitization,...

pypdf can exhaust RAM via manipulated LZWDecode streams

ImpactAn attacker who uses this vulnerability can craft a PDF which leads to large memory usage. This requires parsing the content stream of a page using the LZWDecode filter. PatchesThis has been fixed in pypdf==6.1.3. WorkaroundsIf you cannot upgrade yet, consider applying the changes from PR...

PyLoad vulnerable to SQL Injection via API /json/add_package in add_links parameter

SummaryThe parameter addlinks in the API /json/addpackage is vulnerable to SQL Injection. SQL injection vulnerabilities can lead to sensitive data leakage. Details- Affected file：https://github.com/pyload/pyload/blob/develop/src/pyload/core/database/filedatabase.pyL271- Affected...

Quivr unauthenticated Denial of Service (DoS) via Multipart Boundary

A Denial of Service DoS vulnerability in the file upload feature of stangirard/quivr v0.0.298 allows unauthenticated attackers to cause excessive resource consumption by appending characters to the end of a multipart boundary in an HTTP request. This leads to the server continuously processing ea...

Open WebUI has SSRF in /openai/models

The /openai/models endpoint in open-webui/open-webui version 0.3.8 is vulnerable to Server-Side Request Forgery SSRF. An attacker can change the OpenAI URL to any URL without checks, causing the endpoint to send a request to the specified URL and return the output. This vulnerability allows the...

FastChat Server-Side Request Forgery vulnerability

A Server-Side Request Forgery SSRF vulnerability was identified in the lm-sys/fastchat web server, specifically in the affected version git 2c68a13. This vulnerability allows an attacker to access internal server resources and data that are otherwise inaccessible, such as AWS metadata credentials...