Open WebUI versions prior to 0.8.11 allow authenticated users to access notes belonging to other users via Indirect Object Reference (IDOR) in the /api/v1/notes/{note_id} endpoint, enabling unauthorized disclosure of potentially sensitive private data. The vulnerability is exploitable both when the notes feature is enabled in the UI and when disabled but re-enabled via /api/config endpoint manipulation, requiring only valid user authentication and UUID enumeration or guessing.
Authenticated admin users in pyLoad-ng can bypass the CVE-2026-33509 fix by setting the storage_folder to the Flask session directory (/tmp/pyLoad/flask), then downloading and reusing session files of other users via the /files/get/ endpoint to achieve account takeover. The original patch failed to block access to the session cache directory, leaving it accessible through the directory traversal protection bypass. Publicly available proof-of-concept code confirms the bypass is functional.
Type Confusion in V8 in Google Chrome prior to 148.0.7778.168 allowed a remote attacker to obtain potentially sensitive information from process memory via a crafted HTML page. (Chromium security severity: Medium)
Use after free in Google Lens in Google Chrome prior to 148.0.7778.168 allowed a remote attacker who had compromised the renderer process to obtain potentially sensitive information from process memory via a crafted HTML page. (Chromium security severity: High)
Command injection in @apostrophecms/cli apos create command allows arbitrary command execution when a user supplies specially crafted input during the interactive password prompt. The vulnerability exists in lib/commands/create.js line 186, where user-supplied password input is passed directly into a shell exec() call without sanitization or escaping, enabling attackers to inject shell metacharacters (;, &&, $()) to execute arbitrary commands with the privileges of the user running the CLI. Exploitation requires user interaction (UI:R) and high privilege context (PR:H), but publicly available proof-of-concept demonstrates successful arbitrary code execution on Ubuntu systems with Node.js.
Open WebUI fails to authorize model update requests, allowing authenticated users to modify private models belonging to other users and alter their access controls. Tested on version 0.5.4, this broken access control vulnerability affects all versions up to 0.5.6 and is exploitable via a direct POST request to the /api/v1/models/model/update endpoint without requiring special privileges beyond basic authentication.
Open WebUI before version 0.9.0 allows authenticated users to bypass API key endpoint restrictions by submitting requests via the `x-api-key` header instead of the `Authorization` header, enabling full access to protected endpoints including message processing and potentially admin functionality. The vulnerability affects deployments where admins have configured API key restrictions to limit which endpoints specific keys can access. A proof-of-concept demonstrates that the same API key correctly rejected via `Authorization` header (403 Forbidden) is fully processed via `x-api-key` header (200 OK with LLM response), completely undermining the intended access control model. No active public exploitation is reported, but the vulnerability is straightforward to exploit and has been verified against Open WebUI v0.8.11.
Time-based blind SQL injection in the Taskbuilder - Project Management & Task Management Tool With Kanban Board WordPress plugin (all versions through 5.0.6) enables authenticated attackers with Subscriber-level access or above to exfiltrate arbitrary data from the underlying database via the 'project_search' parameter. The vulnerability stems from unsanitized user input being passed into an unprepared SQL query, exposing confidential database contents including user credentials, configuration data, and application records. No public exploit code has been identified at time of analysis, EPSS probability is very low at 0.03%, and CISA has not added this to the Known Exploited Vulnerabilities catalog.
Privilege escalation in Essential Addons for Elementor (all versions ≤ 6.5.13) allows authenticated WordPress users with Author-level access or above to create new accounts with elevated roles such as Editor by exploiting the plugin's `register_user` function, which applies an incomplete role denylist that blocks only 'administrator' while leaving other privileged roles unguarded. The network-accessible, low-complexity attack vector (AV:N/AC:L/PR:L) makes this realistic for any site with the plugin's registration widget exposed and populated with low-trust authors. No public exploit has been identified at time of analysis and CISA KEV status is absent, but the plugin's broad WordPress deployment increases aggregate exposure.
Denial-of-service in GitLab Enterprise Edition allows a crafted file upload to exhaust service availability through improper deserialization validation. The vulnerability spans an exceptionally wide range, affecting all GitLab EE instances from version 11.9 through the 18.11 line until patched releases. There is no public exploit identified at time of analysis and EPSS sits at 0.02% (4th percentile), indicating low observed exploitation pressure, though the breadth of the affected version range means unpatched installations represent a meaningful attack surface for availability disruption.
Infinite loop denial-of-service in OpenStack Ironic's image handling allows low-privileged authenticated attackers to exhaust conductor resources by supplying file:///dev/zero as an image URL, triggering unbounded checksum calculations that never terminate. All Ironic versions from 0 through 35.x prior to commit a3f6d73 are affected. No public exploit independently confirmed but SSVC data indicates proof-of-concept exists; EPSS sits at 0.01% (2nd percentile), consistent with low widespread exploitation likelihood despite the poc signal.
Cross-Site Request Forgery in GitLab CE/EE allows an unauthenticated attacker to create unauthorized Jira subscriptions within a targeted authenticated user's namespace by tricking the victim into clicking a specially crafted link. All GitLab installations from version 11.10 through the pre-patch 18.x releases are affected across both Community and Enterprise editions. No public exploit exists and this is not listed in CISA KEV; however, the broad version range spanning over seven years of releases and the prevalence of Jira integrations in enterprise GitLab deployments make patching a meaningful priority.
Improper authorization in GitLab CE/EE exposes confidential issue content to any authenticated user on public projects across three active release branches. The flaw (CWE-288) allows a low-privileged, authenticated user - including users with no project membership - to read issue content that project owners explicitly marked confidential and restricted. With patch versions released on May 13, 2026 and no public exploit or KEV listing identified at time of analysis, the immediate threat is bounded, though the confidentiality impact is rated High by CVSS given the potential for sensitive business or security-relevant issue data to be disclosed.
Remote authenticated actors with S3 write access can achieve code execution in Amazon SageMaker Triton inference containers by replacing model artifacts with malicious pickle payloads that are deserialized without integrity verification. Affected versions are SDK v2 before v2.257.2 and v3 before v3.8.0. The vulnerability requires high-privilege S3 access to the model artifact path but carries severe impact including arbitrary code execution within inference containers. No public exploit code or active exploitation has been identified at time of analysis.
Stored Cross-Site Scripting in Royal Elementor Addons and Templates plugin for WordPress up to version 1.7.1058 allows authenticated attackers with Contributor-level privileges to inject arbitrary JavaScript via the 'title_tag' parameter in multiple widgets, with execution occurring when any user accesses the affected page. The vulnerability stems from insufficient input sanitization and output escaping on user-controlled parameters.
Stored cross-site scripting in CC Child Pages WordPress plugin up to version 2.1.1 allows authenticated attackers with Contributor-level access and above to inject arbitrary JavaScript via the 'more' parameter, which executes in the browsers of users viewing the compromised pages. CVSS 6.4 reflects the authentication gate but cross-site scope (C:L/I:L/S:C). No public exploit code or active exploitation confirmed at time of analysis.
Stored Cross-Site Scripting in the Meta Field Block WordPress plugin (versions ≤1.5.2) allows authenticated contributors to inject persistent malicious scripts via the unsanitized 'tagName' block attribute. Any user who subsequently visits a page containing the injected block will execute the attacker's arbitrary JavaScript in their browser context. A vendor-released patch (1.5.3) is available; no public exploit has been identified and EPSS and SSVC signals both indicate low exploitation probability at this time.
Stored Cross-Site Scripting in the Bold Page Builder WordPress plugin (versions up to and including 5.6.8) allows authenticated attackers holding contributor-level access or higher to inject persistent malicious JavaScript via the 'text' attribute of the bt_bb_button shortcode. The script executes in the browser of any user who subsequently loads the compromised page, including administrators, enabling session hijacking or unauthorized action execution under victim identity. No active exploitation is confirmed - the vulnerability is absent from CISA KEV, carries an EPSS score of 0.03% (8th percentile), and SSVC designates exploitation status as none.
Stored cross-site scripting in The Plus Addons for Elementor WordPress plugin (all versions through 6.4.11) allows authenticated contributors to persistently inject arbitrary JavaScript via the `menu_hover_click` parameter of the Navigation Menu Lite widget. The CVSS scope change (S:C) confirms the payload executes in visiting users' browser contexts rather than the attacker's, enabling session hijacking, credential harvesting, or admin account takeover against any user who loads an affected page. EPSS is very low at 0.03% (8th percentile) and SSVC confirms no known exploitation, but the low privilege bar - contributor access, which is routinely granted on multi-author WordPress sites - elevates practical risk above what the medium CVSS score alone suggests.
Stored Cross-Site Scripting in Envira Gallery Lite (WordPress plugin, versions ≤ 1.12.4) enables authenticated attackers with Author-level access to inject persistent arbitrary JavaScript into gallery pages via the REST API. The attack exploits a context mismatch: the `arrows` gallery parameter is omitted from `sanitize_config_values()` sanitization, and `gallery_init()` escapes it with `esc_attr()` — a function designed for HTML attribute contexts, not inline JavaScript — allowing JavaScript expression injection that executes in any visitor's browser. No public exploit or active exploitation has been identified; EPSS is 0.01% and CISA KEV status is absent, placing this firmly in the low-immediate-risk tier despite its persistent, cross-user impact.
Remote code execution in Archon 0.1.0 enables attackers to execute arbitrary commands and steal API keys when victims access a malicious HTML page. The attack exploits insufficient input validation (CWE-94) to control the Archon UI, run prompts on behalf of authenticated users, and exfiltrate all information displayed in the interface. With EPSS score of 0.04% (13th percentile) and no confirmed active exploitation, this represents a web-to-client attack requiring social engineering but offering significant post-exploitation capabilities once triggered.
Argument injection in dbt-mcp v1.15.1 through v1.17.0 allows MCP clients to inject arbitrary dbt command-line flags such as --profiles-dir, --project-dir, and --target via unsanitized node_selection and resource_type parameters, enabling attackers to redirect dbt's configuration and database operations to attacker-controlled locations. The vulnerability is exploitable via two independent vectors in the _run_dbt_command() function and has been verified by proof-of-concept code demonstrating arbitrary dbt profile injection. Vendor-released patch available in v1.17.1.
Cleartext transmission of sensitive information in KDDI's Android parental-control app 'あんしんフィルター for au' (Anshin Filter for au) exposes communications to interception and modification by man-in-the-middle attackers. Versions prior to 4.9_b0003 transmit data over unencrypted channels, enabling an attacker with a privileged network position to read sensitive user data or tamper with app communications. Disclosed by JPCERT/CC (JVN#24167657), no public exploit code exists and CISA has not listed this in KEV; EPSS of 0.01% (3rd percentile) and SSVC exploitation status of 'none' confirm negligible current threat activity.
Unauthenticated attackers with knowledge of the KVM key download endpoint URL can retrieve sensitive cryptographic keys without authentication, compromising confidentiality of encrypted communications and system secrets. This affects AMD KVM implementations and requires no special user interaction, though attackers must possess prior knowledge of the specific endpoint URL. No public exploit code or active exploitation has been identified at time of analysis.
pyzipper before version 0.4.0 fails to use AE-2 encryption format due to an operator precedence bug, causing CRC32 checksums to be stored unencrypted in ZIP headers. Attackers with access to encrypted archives can extract plaintext CRC32 values and conduct brute-force attacks on small or low-entropy files to recover their content without decrypting the AES encryption itself. Large or high-entropy files remain practically safe under current computational constraints, but the vulnerability represents a cryptographic bypass for files under approximately 20 bytes.
Reflected Cross-Site Scripting in the MapGeo - Interactive Geo Maps WordPress plugin (all versions through 1.6.27) allows unauthenticated remote attackers to inject arbitrary JavaScript via the unsanitized 'map' parameter of the display-map shortcode. Exploitation requires tricking an authenticated or anonymous site visitor into clicking a crafted URL, after which the injected script executes in the victim's browser within the plugin's page scope. No public exploit code or CISA KEV listing exists at time of analysis; EPSS probability is 0.06%, placing this in the 19th percentile of exploitation likelihood.
Reflected Cross-Site Scripting in the GLS Shipping for WooCommerce WordPress plugin (all versions through 1.4.0) allows unauthenticated remote attackers to inject arbitrary JavaScript via the unsanitized 'failed_orders' parameter, executing in victims' browsers when they interact with a crafted link. The CVSS vector (PR:N, S:C) confirms no authentication is needed by the attacker, though user interaction is required, and the Changed scope rating means successful exploitation can affect resources beyond the plugin's own context - such as session tokens or other page content. No active exploitation is identified at time of analysis; EPSS stands at 0.06% (17th percentile) and CISA SSVC assigns exploitation status of 'none', making this a theoretical rather than emergent threat.
Cross-site scripting (XSS) vulnerability in mistune's render_toc_ul() function allows attackers to inject arbitrary HTML and JavaScript into table-of-contents output by crafting malicious heading IDs. When heading identifiers are derived from user-supplied text (standard practice for readable slug anchors), an attacker can break out of the href attribute context with a payload like `x"><script>alert(document.cookie)</script><a href="`, causing the script block to execute in the rendered TOC. The vulnerability requires user interaction (UI:R) to view the poisoned TOC but affects all users of the generated page. Vendor-released patch available in mistune 3.2.1.
Cross-site scripting (XSS) vulnerability in Northern.tech CFEngine Enterprise before versions 3.21.8, 3.24.3, and 3.27.0 allows remote attackers to inject malicious scripts into web application responses. User interaction is required to trigger the vulnerability, limiting real-time automation of attacks but enabling account compromise, session hijacking, or credential theft against CFEngine administrators and users. No public exploit code or active exploitation has been identified at the time of analysis.
OS command injection in Fleet's software installer pipeline allows arbitrary code execution as root (macOS/Linux) or SYSTEM (Windows) on managed endpoints when a specially crafted software package is uninstalled. The vulnerability exists because package metadata fields are not sanitized before being incorporated into auto-generated uninstall scripts. An attacker with the ability to upload packages to Fleet can exploit this by embedding malicious commands in package metadata fields, resulting in code execution with elevated privileges when endpoints execute the uninstall operation. Patch version 4.81.1 available.
Stack exhaustion in MongoDB PHP driver allows remote denial of service when processing deeply nested BSON documents from untrusted sources. Unauthenticated attackers can crash applications by sending maliciously crafted BSON payloads with excessive nesting levels, affecting all versions of the PHP driver that parse BSON without depth limits. The vulnerability requires high attack complexity but results in complete availability loss.
Portainer's custom template file endpoint (GET /api/custom_templates/{id}/file) exposes template file contents to any authenticated user due to a completely absent authorization check in the customTemplateFile handler - a check that every other custom template endpoint correctly implements. Authenticated users at any privilege level can enumerate sequential integer template IDs to read Docker Compose files belonging to templates they have no explicit access to, potentially harvesting embedded secrets such as database connection strings, API tokens, and registry credentials. No public exploit or CISA KEV listing has been identified at time of analysis; however, exploitation requires only a valid session and sequential ID guessing, making it trivially scriptable against any unpatched multi-tenant Portainer instance.
electerm's sync encryption uses deterministic AES-192-CBC with a fixed zero IV, constant KDF salt, and no message authentication code, allowing attackers to crack common passwords across multiple installations and perform undetected bit-flip attacks on synced bookmark and profile data. Affects electerm versions prior to 3.9.5. No public exploit code identified at time of analysis, but the cryptographic weaknesses are fundamental and exploitable without specialized tooling.
Improper privilege management in AMD's KVM key download component allows authenticated local attackers to swap tokens and exfiltrate sensitive cryptographic keys due to insufficient access controls, potentially enabling unauthorized access to privileged resources and compromising system confidentiality. The vulnerability requires authenticated access (PR:L) but carries high confidentiality impact (VC:H), making it a significant risk in multi-tenant or shared-access environments.
Regular expression denial of service (ReDoS) in Svelte 5.51.5 through 5.55.6 allows attackers to cause application hang or crash by passing unconstrained-length tag names to the `<svelte:element>` component, triggering exponential regex evaluation time in the runtime tag validation logic. The vulnerability requires applications to accept user-controlled tag input without length or content restrictions.
GitLab's Jira integration exposes out-of-scope Jira issues to authenticated GitLab users across all editions (CE and EE) from version 13.7 through the patched releases, due to the integration's project-scope filter operating only as a UI display control rather than an enforced access boundary. The Changed scope (S:C) in the CVSS vector reflects that impact crosses into Atlassian Jira - a component outside GitLab's own trust domain - allowing confidential Jira issue data to leak beyond intended project boundaries. No public exploit exists and no active exploitation has been confirmed; EPSS is negligible at 0.01% (3rd percentile), placing this firmly in low-exploitation-probability territory despite a broad version range spanning over five years of releases.
Path traversal in Portainer's backup restore feature allows arbitrary file write to the server filesystem via crafted tar.gz archives. Versions prior to 2.39.0 (and 2.33.8 in the LTS branch) use unsafe path joining in the ExtractTarGz function, permitting entries like `../../etc/cron.d/evil` to escape the extraction root. An authenticated administrator who restores a malicious archive or whose credentials are compromised can write files to any path accessible to the Portainer process, potentially establishing persistence on the host. Exploitation requires administrative privileges within Portainer, reducing practical impact.
Integer overflow in OpenImageIO TGA image decoder allows local attackers to trigger out-of-bounds buffer read and denial of service. When processing TGA image files, the bounds check in TGAInput::decode_pixel computes k + palbytespp using unsigned 32-bit arithmetic; specifically, when k equals 0xFFFFFFFC and palbytespp equals 4, the sum wraps to zero, bypassing the palette size validation. The subsequent palette access then uses the unwrapped value as an array index, reading approximately 4 GB past the palette buffer start, causing a segmentation fault. Affects OpenImageIO versions prior to 3.0.18.0 and 3.1.13.0; requires local file access and user interaction to open a malicious TGA file. CVSS score of 5.5 reflects local-only attack vector with high availability impact but no confidentiality or integrity impact.
Discretionary access control bypass in Chrome Remote Desktop (Chromoting) allows adjacent network attackers to achieve limited confidentiality, integrity, and availability impact through a malicious file requiring user interaction. Google released Chrome 148.0.7778.168 to address this medium-severity flaw. EPSS score of 0.01% (1st percentile) and CISA SSVC assessment indicate low real-world exploitation probability with no observed exploitation activity. The adjacent network attack vector (AV:A) significantly constrains attacker positioning compared to typical remote vulnerabilities.
Incorrect security UI in Fullscreen in Google Chrome prior to 148.0.7778.168 allowed a remote attacker to perform UI spoofing via a crafted HTML page. (Chromium security severity: Medium)
Mass assignment vulnerability in Open WebUI v0.9.2 allows authenticated attackers to spoof user identities and manipulate model evaluation data by injecting a `user_id` field into feedback requests. The `POST /api/v1/evaluations/feedback` endpoint fails to properly validate and segregate server-set values from user-supplied input, enabling attackers to create feedback records attributed to arbitrary users and corrupt Elo-based model leaderboard rankings. Patch available in v0.9.5.
Stored cross-site scripting (XSS) in Open WebUI versions up to 0.9.2 allows authenticated users to upload malicious Office files (Excel, DOCX, PPT) that execute arbitrary JavaScript in the browsers of any user previewing the file. The vulnerability stems from rendering user-supplied HTML from file-preview components using Svelte's `{@html}` directive without DOMPurify sanitization, despite DOMPurify being available and correctly applied in 39% of the codebase's other rendering locations. This is a regression of a previously patched vulnerability (GHSA-jwf8-pv5p-vhmc) that was fixed in v0.8.0 but reintroduced after that release.
Stored cross-site scripting in Open WebUI versions before 0.8.0 allows authenticated users to inject malicious data URIs into the profile_image_url field, enabling JavaScript execution in the application origin. The most critical attack path involves uploading a base64-encoded SVG data URI that, when loaded via the /api/v1/users/{user_id}/profile/image endpoint, executes scripts with access to localStorage and enables full account takeover of any user viewing the malicious profile image, including administrators. Two independent reporters demonstrated distinct vectors: one via HTML data URIs in new tabs (limited scope), and one via SVG data URIs served by the application origin (account takeover). No public exploit code or active exploitation has been confirmed, but the vulnerability requires only authenticated access and user interaction (viewing a profile image).
Script injection in SanitizerAPI in Google Chrome on Android prior to 148.0.7778.168 allowed a remote attacker to inject arbitrary scripts or HTML (UXSS) via a crafted HTML page. (Chromium security severity: High)
Cisco Catalyst SD-WAN Manager web UI fails to properly redact sensitive information in device configurations and templates, allowing authenticated users with read-only permissions to extract and leverage privileged credentials to escalate their access and modify system configurations. The vulnerability affects all versions of the product and requires only network access and valid (albeit minimal) read-only credentials; successful exploitation grants attackers high-privileged administrative capability over the SD-WAN fabric.
Privilege escalation in Cisco Catalyst SD-WAN Manager allows authenticated users with read-only permissions to elevate privileges to high-privileged user level through exposure of sensitive session information in audit logs. An attacker with initial read-only access can extract high-privilege session credentials from audit logs and impersonate an administrator, bypassing intended access controls. CVSS score 5.4 (medium) reflects the requirement for initial authentication, though the ease of escalation (AC:L) and direct path to administrative capability represent significant risk in multi-tenant or shared SD-WAN deployments.
Open WebUI versions 0.8.10 and earlier allow authenticated users to bypass model access control by appending ?bypass_filter=true to POST requests to /openai/chat/completions or /ollama/api/chat endpoints. The vulnerability exposes an internal-only FastAPI function parameter to external HTTP clients via query string binding, permitting any authenticated user to invoke admin-restricted models regardless of their assigned access grants. Vendor-released patch: v0.8.11 (March 2026). No public exploit code identified beyond the PoC in the advisory, but exploitation is trivial for any authenticated user.
Missing authorization in PostgreSQL CREATE TYPE allows authenticated users to hijack search_path resolution and force other database users to execute arbitrary SQL functions chosen by the attacker. An authenticated attacker can create a malicious user-defined type in a schema that appears earlier in a victim's search_path than legitimate extension or system types, causing the victim's queries to execute attacker-controlled functions instead of intended ones. This affects PostgreSQL versions 14.x before 14.23, 15.x before 15.18, 16.x before 16.14, 17.x before 17.10, and 18.x before 18.4. While CVSS 5.4 is moderate, the attack requires authenticated database access and carries real risk in multi-tenant or shared PostgreSQL environments where privilege escalation or lateral movement is the goal.
Missing authorization controls in the WP Encryption WordPress plugin (all versions through 7.8.5.10) allow any authenticated subscriber-level user to invoke the `wple_basic_get_requests` function and tamper with SSL lifecycle configuration - resetting SSL setup state, falsely marking SSL as complete, and altering plan selection options. The affected plugin automates Let's Encrypt certificate provisioning for WordPress sites, meaning successful exploitation can silently break or misrepresent a site's HTTPS posture. No public exploit code exists and no active exploitation is confirmed (not in CISA KEV); EPSS at 0.02% and SSVC exploitation rating of 'none' place this firmly in the low-urgency tier despite the low privilege bar.
Cross-site scripting in GitLab CE/EE 18.11.x (before 18.11.3) enables an authenticated low-privilege attacker to inject unsanitized content that executes arbitrary JavaScript within another user's browser session. The CVSS Scope Changed (S:C) flag reflects the cross-user impact boundary: the vulnerability allows one GitLab principal to affect a distinct security context belonging to a different user, including potentially higher-privileged accounts such as project owners or instance administrators. No public exploit code exists and no active exploitation has been identified - EPSS at 0.02% (5th percentile) and SSVC exploitation status of 'none' both confirm low immediate urgency - though GitLab's wide enterprise deployment makes patching to 18.11.3 advisable.