Advisories

A security flaw has been discovered in funadmin up to 7.1.0-rc4. Affected by this issue is the function repass of the file app/frontend/controller/Member.php. Performing a manipulation of the argument forget_code/vercode results in weak password recovery. Remote exploitation of the attack is possible. The attack's complexity is rated as high. The exploitation is known to be difficult. The exploit has been released to the public and may be used for attacks. …

A weakness has been identified in funadmin up to 7.1.0-rc4. This affects the function setConfig of the file app/backend/controller/Ajax.php of the component Configuration Handler. Executing a manipulation can lead to improper authorization. The attack can be executed remotely. The exploit has been made available to the public and could be used for attacks. The vendor was contacted early about this disclosure but did not respond in any way.

A vulnerability was identified in funadmin up to 7.1.0-rc4. Affected by this vulnerability is the function getMember of the file app/frontend/view/login/forget.html. Such manipulation leads to information disclosure. The attack may be launched remotely. The exploit is publicly available and might be used. The vendor was contacted early about this disclosure but did not respond in any way.

A Denial of Service vulnerability was identified in Moodle’s TeX formula editor. When rendering TeX content using mimetex, insufficient execution time limits could allow specially crafted formulas to consume excessive server resources. An authenticated user could abuse this behavior to degrade performance or cause service interruption.

A flaw was identified in Moodle’s backup restore functionality where specially crafted backup files were not properly validated during processing. If a malicious backup file is restored, it could lead to unintended execution of server-side code. Since restore capabilities are typically available to privileged users, exploitation requires authenticated access. Successful exploitation could result in full compromise of the Moodle server.

This vulnerability allows remote attackers to bypass authentication on affected installations of MLflow. Authentication is not required to exploit this vulnerability. The specific flaw exists within the basic_auth.ini file. The file contains hard-coded default credentials. An attacker can leverage this vulnerability to bypass authentication and execute arbitrary code in the context of the administrator.

MLflow Tracking Server Artifact Handler Directory Traversal Remote Code Execution Vulnerability. This vulnerability allows remote attackers to execute arbitrary code on affected installations of MLflow Tracking Server. Authentication is not required to exploit this vulnerability. The specific flaw exists within the handling of artifact file paths. The issue results from the lack of proper validation of a user-supplied path prior to using it in file operations. An attacker can leverage …

When a DAG failed during parsing, Airflow’s error-reporting in the UI could include the full kwargs passed to the operators. If those kwargs contained sensitive values (such as secrets), they might be exposed in the UI tracebacks to authenticated users who had permission to view that DAG. The issue has been fixed in Airflow 3.1.5rc1 and 2.11.1, and users are strongly advised to upgrade to prevent potential disclosure of sensitive …

There is a potential vulnerability in Traefik managing HTTP/3 connections. More details in the CVE-2025-68121.

There is a potential vulnerability in Traefik managing HTTP/3 connections. More details in the CVE-2025-68121.

There is a potential vulnerability in Traefik managing HTTP/3 connections. More details in the CVE-2025-68121.

A Stored Cross-Site Scripting (XSS) vulnerability in Sync-in Server before 1.9.3 allows an authenticated attacker to execute arbitrary JavaScript in a victim's browser. By uploading a crafted SVG file containing a malicious payload, an attacker can access and exfiltrate sensitive information, including the user's session cookies.

A Timing-based username enumeration in Basic Authentication vulnerability due to early response on invalid usernames could allow attackers to identify valid users and focus their efforts on targeted brute-force or credential-stuffing attacks.

Ray’s dashboard HTTP server blocks browser-origin POST/PUT but does not cover DELETE, and key DELETE endpoints are unauthenticated by default. If the dashboard/agent is reachable (e.g., –dashboard-host=0.0.0.0), a web page via DNS rebinding or same-network access can issue DELETE requests that shut down Serve or delete jobs without user interaction. This is a drive-by availability impact.

Potential unintentional disclosure of local files from the packaging machine into a generated .skill artifact. Requires local execution of the packaging script on attacker-controlled skill contents.

Local ACP sessions may become less responsive when very large prompts are submitted Larger-than-expected model usage/cost when oversized text is forwarded No privilege escalation and no direct remote attack path in the default ACP model

Affected Packages / Versions openclaw npm package versions <= 2026.2.17. Vulnerability Cron webhook delivery in src/gateway/server-cron.ts used fetch() directly, so webhook targets could reach private/metadata/internal endpoints without SSRF policy checks. Fix Commit(s) 99db4d13e 35851cdaf Thanks @Adam55A-code for reporting.

In setups where Discord moderation actions are enabled and the bot has the necessary guild permissions, a non-admin user could request moderation actions by spoofing sender identity fields.

Email properties (such as to, subject, html, text, and attachments) are not reset between sends when a single client instance is reused across multiple .send() calls. This can cause properties from a previous send to leak into a subsequent one, potentially delivering content or recipient addresses to unintended parties. Applications sending emails to different recipients in sequence — such as transactional flows like password resets or notifications — are affected.

Stored Cross-Site Scripting (XSS) in the _genai/_evals_visualization component of Google Cloud Vertex AI SDK (google-cloud-aiplatform) versions from 1.98.0 up to (but not including) 1.131.0 allows an unauthenticated remote attacker to execute arbitrary JavaScript in a victim's Jupyter or Colab environment via injecting script escape sequences into model evaluation results or dataset JSON data.

Predictable bucket naming in Vertex AI Experiments in Google Cloud Vertex AI from version 1.21.0 up to (but not including) 1.133.0 on Google Cloud Platform allows an unauthenticated remote attacker to achieve cross-tenant remote code execution, model theft, and poisoning via pre-creating predictably named Cloud Storage buckets (Bucket Squatting). This vulnerability was patched and no customer action is needed.

Fickling's check_safety() API and –check-safety CLI flag incorrectly rate as LIKELY_SAFE pickle files that open outbound TCP connections at deserialization time using stdlib network-protocol constructors: smtplib.SMTP, imaplib.IMAP4, ftplib.FTP, poplib.POP3, telnetlib.Telnet, and nntplib.NNTP. The bypass exploits two independent root causes described below.

A dot (.) in a DOCTYPE entity name is treated as a regex wildcard during entity replacement, allowing an attacker to shadow built-in XML entities (&lt;, &gt;, &amp;, &quot;, &apos;) with arbitrary values. This bypasses entity encoding and leads to XSS when parsed output is rendered.

This affects versions of the package bn.js before 4.12.3 and 5.2.3. Calling maskn(0) on any BN instance corrupts the internal state, causing toString(), divmod(), and other methods to enter an infinite loop, hanging the process indefinitely.

AVideo allows Markdown in video comments and uses Parsedown (v1.7.4) without Safe Mode enabled. Markdown links are not sufficiently sanitized, allowing javascript: URIs to be rendered as clickable links. An authenticated low-privilege attacker can post a malicious comment that injects persistent JavaScript. When another user clicks the link, the attacker can perform actions such as session hijacking, privilege escalation (including admin takeover), and data exfiltration.

The zumba/json-serializer library allows deserialization of PHP objects from JSON using a special @type field. Prior to version 3.2.3, the deserializer would instantiate any class specified in the @type field without restriction. When processing untrusted JSON input, this behavior may allow an attacker to instantiate arbitrary classes available in the application. If a vulnerable application passes attacker-controlled JSON into JsonSerializer::unserialize() and contains classes with dangerous magic methods (such as __wakeup() …

The asm! block enabled by the off-by-default asm feature, when enabled on ARMv8 targets, misspecified the operand type for all of its operands, using in for pointers and values which were subsequently mutated by operations performed within the assembly block.

On February 17, 2026 at 3:26 AM PT, an unauthorized party used a compromised npm publish token to publish an update to Cline CLI on the NPM registry: cline@2.3.0. The published package contains a modified package.json with an added postinstall script: "postinstall": "npm install -g openclaw@latest" This causes openclaw (an unrelated, non-malicious open source package) to be globally installed when cline@2.3.0 is installed. No other files were modified – the …

When using <svelte:element this={tag}> in server-side rendering, the provided tag name is not validated or sanitized before being emitted into the HTML output. If the tag string contains unexpected characters, it can result in HTML injection in the SSR output. Client-side rendering is not affected.

In server-side rendering, attribute spreading on elements (e.g. <div {…attrs}>) enumerates inherited properties from the object's prototype chain rather than only own properties. In environments where Object.prototype has already been polluted — a precondition outside of Svelte's control — this can cause unexpected attributes to appear in SSR output or cause SSR to throw errors. Client-side rendering is not affected.

In certain circumstances, the server-side rendering output of an <option> element does not properly escape its content, potentially allowing HTML injection in the SSR output. Client-side rendering is not affected.

Versions of svelte prior to 5.51.5 are vulnerable to cross-site scripting (XSS) during server-side rendering. When using spread syntax to render attributes from untrusted data, event handler properties are included in the rendered HTML output. If an application spreads user-controlled or external data as element attributes, an attacker can inject malicious event handlers that execute in victims' browsers.

Stored XSS vulnerability in html fieldtypes allow authenticated users with field management permissions to inject malicious JavaScript that executes when viewed by higher-privileged users.

PyO3 0.28.1 added support for #[pyclass(extends=PyList)] struct NativeSub (and other native types) when targeting Python 3.12 and up with the abi3 feature. It was discovered that subclasses of such classes would use the type of the subclass when attempting to access to data of NativeSub contained within Python objects, amounting to memory corruption. PyO3 0.28.2 fixed the issue by using the type of (e.g.) NativeSub correctly.

A prototype pollution vulnerability exists in the the npm package swiper (>=6.5.1, < 12.1.2). Despite a previous fix that attempted to mitigate prototype pollution by checking whether user input contained a forbidden key, it is still possible to pollute Object.prototype via a crafted input using Array.prototype. The exploit works across Windows and Linux and on Node and Bun runtimes. This issue is fixed in version 12.1.2

The hot spot attributes configuration property allowed any attribute to be set, including HTML event handler attributes, allowing for potential XSS attacks. This affects websites hosting the standalone viewer HTML file and any other use of untrusted JSON config files (bypassing the protections of the escapeHTML parameter). As certain events fire without any additional user interaction, visiting a standalone viewer URL that points to a malicious config file—without additional user …

tools.exec.safeBins could be bypassed for filesystem access when sort output flags (-o / –output) or recursive grep flags were allowed through safe-bin execution paths.

tools.exec.safeBins could be bypassed for filesystem access when sort output flags (-o / –output) or recursive grep flags were allowed through safe-bin execution paths.

An information disclosure vulnerability in OpenClaw's tools.exec.safeBins approval flow allowed a file-existence oracle. When safe-bin validation examined candidate file paths, command allow/deny behavior could differ based on whether a path already existed on the host filesystem. An attacker could probe for file presence by comparing outcomes for existing vs non-existing filenames.

An information disclosure vulnerability in OpenClaw's tools.exec.safeBins approval flow allowed a file-existence oracle. When safe-bin validation examined candidate file paths, command allow/deny behavior could differ based on whether a path already existed on the host filesystem. An attacker could probe for file presence by comparing outcomes for existing vs non-existing filenames.

The sandbox identifier cache key for Docker/browser sandbox configuration used SHA-1 to hash normalized configuration payloads. SHA-1 is deprecated for cryptographic use and has known collision weaknesses. In this code path, deterministic IDs are used to decide whether an existing sandbox container can be reused safely. A collision in this hash could let one configuration be interpreted as another under the same sandbox cache identity, increasing the risk of cache …

The sandbox identifier cache key for Docker/browser sandbox configuration used SHA-1 to hash normalized configuration payloads. SHA-1 is deprecated for cryptographic use and has known collision weaknesses. In this code path, deterministic IDs are used to decide whether an existing sandbox container can be reused safely. A collision in this hash could let one configuration be interpreted as another under the same sandbox cache identity, increasing the risk of cache …

The web_fetch tool could be used to crash the OpenClaw Gateway process (OOM / resource exhaustion) by fetching and attempting to parse attacker-controlled web pages with oversized response bodies or pathological HTML nesting.

The web_fetch tool could be used to crash the OpenClaw Gateway process (OOM / resource exhaustion) by fetching and attempting to parse attacker-controlled web pages with oversized response bodies or pathological HTML nesting.

In affected versions, when apply_patch was enabled and the agent ran without filesystem sandbox containment, crafted paths could cause file writes/deletes outside the configured workspace directory.

In affected versions, when apply_patch was enabled and the agent ran without filesystem sandbox containment, crafted paths could cause file writes/deletes outside the configured workspace directory.

Impact: An RCE vulnerability has been identified in Microsoft Semantic Kernel Python SDK, specifically within the InMemoryVectorStore filter functionality. Patches: The problem has been fixed in python-1.39.4. Users should upgrade this version or higher. Workarounds: Avoid using InMemoryVectorStore for production scenarios. References: Release python-1.39.4 · microsoft/semantic-kernel · GitHub PR to block use of dangerous attribute names that must not be accessed in filter expressions

htmlEscaped in leaf-kit will only escape html special characters if the extended grapheme clusters match, which allows bypassing escaping by using an extended grapheme cluster containing both the special html character and some additional characters. In the case of html attributes, this can lead to XSS if there is a leaf variable in the attribute that is user controlled.

A flaw was identified in the Docker v2 authentication endpoint of Keycloak, where tokens continue to be issued even after a Docker registry client has been administratively disabled. This means that turning the client “Enabled” setting to OFF does not fully prevent access. As a result, previously valid credentials can still be used to obtain authentication tokens. This weakens administrative controls and could allow unintended access to container registry resources.

An issue in Kata with Cloud Hypervisor allows a user of the container to modify the file system used by the Guest micro VM ultimately achieving arbitrary code execution as root in said VM. The current understinding is this doesn’t impact the security of the Host or of other containers / VMs running on that Host (note that arm64 QEMU lacks NVDIMM read-only support: It is believed that until the …

Kargo's authorization model includes a promote verb – a non-standard Kubernetes "dolphin verb" – that gates the ability to advance Freight through a promotion pipeline. This verb exists to separate the ability to manage promotion-related resources from the ability to trigger promotions, enabling fine-grained access control over what is often a sensitive operation. The promote verb is correctly enforced in Kargo's legacy gRPC API. However, three endpoints in the newer …

The batch resource creation endpoints of both Kargo's legacy gRPC API and newer REST API accept multi-document YAML payloads. When either endpoint creates a Project resource, creation of subsequent resources from that same payload belonging in that Project's underlying Kubernetes namespace, by design, proceeds using the API server's own permissions. The creator of a new Project automatically becomes its administrator, but those permissions are granted asynchronously by the management controller. …

User control of the argument of the addJS method allows an attacker to inject arbitrary PDF objects into the generated document. By crafting a payload that escapes the JavaScript string delimiter, an attacker can execute malicious actions or alter the document structure, impacting any user who opens the generated PDF. import { jsPDF } from "jspdf"; const doc = new jsPDF(); // Payload: // 1. ) closes the JS string. …

User control of properties and methods of the Acroform module allows users to inject arbitrary PDF objects, such as JavaScript actions. If given the possibility to pass unsanitized input to the following property, a user can inject arbitrary PDF objects, such as JavaScript actions, which are executed when the victim hovers over the radio option. AcroformChildClass.appearanceState Example attack vector: import { jsPDF } from "jspdf" const doc = new jsPDF(); …

User control of the first argument of the addImage method results in denial of service. If given the possibility to pass unsanitized image data or URLs to the addImage method, a user can provide a harmful GIF file that results in out of memory errors and denial of service. Harmful GIF files have large width and/or height entries in their headers, wich lead to excessive memory allocation. Other affected methods …

The basicAuth and bearerAuth middlewares previously used a comparison that was not fully timing-safe. The timingSafeEqual function used normal string equality (===) when comparing hash values. This comparison may stop early if values differ, which can theoretically cause small timing differences. The implementation has been updated to use a safer comparison method.

The application fails to properly enforce role-based authorization during account creation. Although the system validates that the specified role exists, it does not verify whether the current user has sufficient privileges to assign highly privileged roles such as admin. As a result, an authenticated user with the editor role can create a new account with administrative privileges, leading to full administrative access and complete compromise of the CMS.

When the session object is accessed, Flask should set the Vary: Cookie header. This instructs caches not to cache the response, as it may contain information specific to a logged in user. This is handled in most cases, but some forms of access such as the Python in operator were overlooked. The severity depends on the application's use of the session, and the cache's behavior regarding cookies. The risk depends …

The origin validation uses startsWith() for comparison, allowing attackers to bypass the check by registering a domain that shares a common prefix with an allowed origin. The getAllowedOrigin() function checks if the Referer header starts with any allowed origin: // https://github.com/feathersjs/feathers/blob/dove/packages/authentication-oauth/src/strategy.ts#L75 const allowedOrigin = origins.find((current) => referer.toLowerCase().startsWith(current.toLowerCase())); This comparison is insufficient as it only validates the prefix. This is exploitable when the origins array is configured and an attacker registers …

The redirect query parameter is appended to the base origin without validation, allowing attackers to steal access tokens via URL authority injection. This leads to full account takeover, as the attacker obtains the victim's access token and can impersonate them. The application constructs the final redirect URL by concatenating the base origin with the user-supplied redirect parameter: // https://github.com/feathersjs/feathers/blob/dove/packages/authentication-oauth/src/service.ts#L158C3-L176C4 const { redirect } = query; … session.redirect = redirect; // …

All HTTP request headers are stored in the session cookie, which is signed but not encrypted, exposing internal proxy/gateway headers to clients. The OAuth service stores the complete headers object in the session: // https://github.com/feathersjs/feathers/blob/dove/packages/authentication-oauth/src/service.ts#L173 session.headers = headers; The session is persisted using cookie-session, which base64-encodes the data. While the cookie is signed to prevent tampering, the contents are readable by anyone by simply decoding the base64 value. Under specific …

The ebay_set_user_tokens tool allows updating the .env file with new tokens. The updateEnvFile function in src/auth/oauth.ts blindly appends or replaces values without validating them for newlines or quotes. This allows an attacker to inject arbitrary environment variables into the configuration file.

Under certain circumstances, serializing sparse arrays using uneval or stringify could cause CPU and/or memory exhaustion. When this occurs on the server, it results in a DoS. This is extremely difficult to take advantage of in practice, as an attacker would have to manage to create a sparse array on the server — which is impossible in every mainstream wire format — and then that sparse array would have to …

Under certain circumstances, unevaling untrusted data can produce output code that will create objects with polluted prototypes when later evaled, meaning the output data can be a different shape from the input data.

A command injection vulnerability exists in Deno's node:child_process implementation.

Dagu's default configuration ships with authentication disabled. The POST /api/v2/dag-runs endpoint accepts an inline YAML spec and executes its shell commands immediately with no credentials required — any dagu instance reachable over the network is fully compromised by default.

Users hosting D-Tale publicly can be vulnerable to remote code execution allowing attackers to run malicious code on the server.

Versions of @sveltejs/kit prior to 2.52.2 with remote functions enabled are vulnerable to CPU exhaustion. Malformed form data can cause the server to become unresponsive while processing a request, resulting in denial of service. Only applications using both experimental.remoteFunctions and form are vulnerable.

When verifying artifact signatures using a certificate, Cosign first verifies the certificate chain using the leaf certificate's "not before" timestamp and later checks expiry of the leaf certificate using either a signed timestamp provided by the Rekor transparency log or from a timestamp authority, or using the current time. The root and all issuing certificates are assumed to be valid during the leaf certificate's validity. An issuing certificate with a …

Host Policies will incorrectly permit traffic from Pods on other nodes when all of the following configurations are enabled: Native Routing WireGuard Node Encryption (beta) These options are disabled by default in Cilium.

Centrifugo v6.6.0 binary is compiled with Go 1.25.5 and statically links github.com/quic-go/webtransport-go v0.9.0, having 7 known CVEs Go standard library — compiled with Go 1.25.5: | CVE | Severity | CVSS | Fixed In | |—–|———-|——|———-| | CVE-2025-68121 | CRITICAL | 10.0 | Go 1.25.7, 1.24.13 | | CVE-2025-61726 | HIGH | 7.5 | Go 1.25.6, 1.24.12 | | CVE-2025-61728 | MEDIUM | 6.5 | Go 1.25.6, 1.24.12 | | …

A malicious actor with administrative privileges can upload an arbitrary file to a user-controlled location within the deployment via a system REST API. Successful uploads may lead to remote code execution. By leveraging the vulnerability, a malicious actor may perform Remote Code Execution by uploading a specially crafted payload.

Versions of @sveltejs/adapter-vercel prior to 6.3.2 are vulnerable to cache poisoning. An internal query parameter intended for Incremental Static Regeneration (ISR) is accessible on all routes, allowing an attacker to cause sensitive user-specific responses to be cached and served to other users. Successful exploitation requires a victim to visit an attacker-controlled link while authenticated. Existing deployments are protected by Vercel's WAF, but users should upgrade as soon as possible.

Werkzeug's safe_join function allows Windows device names as filenames if when preceded by other path segments. This was previously reported as https://github.com/pallets/werkzeug/security/advisories/GHSA-hgf8-39gv-g3f2, but the added filtering failed to account for the fact that safe_join accepts paths with multiple segments, such as example/NUL. send_from_directory uses safe_join to safely serve files at user-specified paths under a directory. If the application is running on Windows, and the requested path ends with a special …

Versions of @sveltejs/kit prior to 2.52.2 with remote functions enabled can be vulnerable to memory exhaustion. Malformed form data can cause the server process to crash due to excessive memory allocation, resulting in denial of service. Only applications using both experimental.remoteFunctions and form are vulnerable.

The padding extension was incorrectly removed in utls for the non-pq variant of Chrome 120 fingerprint. Chrome removed this extension only when sending pq keyshares. Only this fingerprint is affected since newer fingerprints have pq keyshares by default and older fingerprints have this extension. Affected symbols: HelloChrome_120 Fix commit: 8fe0b08e9a0e7e2d08b268f451f2c79962e6acd0 Thanks to telegram @acgdaily for reporting this issue.

There is a fingerprint mismatch with Chrome when using GREASE ECH, having to do with ciphersuite selection. When Chrome selects the preferred ciphersuite in the outer ClientHello and the ciphersuite for ECH, it does so consistently based on hardware support. That means, for example, if it prefers AES for the outer ciphersuite, it would also use AES for ECH. The Chrome parrot in utls hardcodes AES preference for outer ciphersuites …

A command injection vulnerability in the wifiNetworks() function allows an attacker to execute arbitrary OS commands via an unsanitized network interface parameter in the retry code path.

A query injection vulnerability exists in the @langchain/langgraph-checkpoint-redis package's filter handling. The RedisSaver and ShallowRedisSaver classes construct RediSearch queries by directly interpolating user-provided filter keys and values without proper escaping. RediSearch has special syntax characters that can modify query behavior, and when user-controlled data contains these characters, the query logic can be manipulated to bypass intended access controls.

An attacker who uses this vulnerability can craft a PDF which leads to long runtimes. This requires a malformed /FlateDecode stream, where the byte-by-byte decompression is used.

An attacker who uses this vulnerability can craft a PDF which leads to long runtimes and large memory consumption. This requires parsing the /ToUnicode entry of a font with unusually large values, for example during text extraction.

An attacker who uses this vulnerability can craft a PDF which leads to an infinite loop. This requires accessing the children of a TreeObject, for example as part of outlines.

This is a scanning bypass to scan_pytorch function in picklescan. As we can see in the implementation of get_magic_number() that uses pickletools.genops(data) to get the magic_number with the condition opcode.name includes INT or LONG, but the PyTorch's implemtation simply uses pickle_module.load() to get this magic_number. For this implementation difference, we then can embed the magic_code into the PyTorch file via dynamic eval on the __reduce__ trick, which can make the …

An issue was discovered in OpenStack Nova before 30.2.2, 31 before 31.2.1, and 32 before 32.1.1. By writing a malicious QCOW header to a root or ephemeral disk and then triggering a resize, a user may convince Nova's Flat image backend to call qemu-img without a format restriction, resulting in an unsafe image resize operation that could destroy data on the host system. Only compute nodes using the Flat image …

OpenClaw versions <= 2026.2.9 construct transcript file paths using an unsanitized sessionId and also accept sessionFile paths without enforcing that they stay within the agent sessions directory. A crafted sessionId and/or sessionFile (example: ../../etc/passwd) can cause path traversal when the gateway performs transcript file read/write operations. Preconditions: an attacker must be able to authenticate to the gateway (gateway token/password). By default the gateway binds to loopback (local-only); configurations that expose …

OpenClaw versions <= 2026.2.9 construct transcript file paths using an unsanitized sessionId and also accept sessionFile paths without enforcing that they stay within the agent sessions directory. A crafted sessionId and/or sessionFile (example: ../../etc/passwd) can cause path traversal when the gateway performs transcript file read/write operations. Preconditions: an attacker must be able to authenticate to the gateway (gateway token/password). By default the gateway binds to loopback (local-only); configurations that expose …

normalizeForHash in src/agents/sandbox/config-hash.ts recursively sorted arrays that contained only primitive values. This made order-sensitive sandbox configuration arrays hash to the same value even when order changed. In OpenClaw sandbox flows, this hash is used to decide whether existing sandbox containers should be recreated. As a result, order-only config changes (for example Docker dns and binds array order) could be treated as unchanged and stale containers could be reused. This is …

Prompt injection may alter agent behavior and could lead to unintended tool use or disclosure of sensitive information.

Disclosure of a Telegram bot token allows an attacker to impersonate the bot and take over Bot API access.

OpenClaw CLI process cleanup used system-wide process enumeration and pattern matching to terminate processes without verifying they were owned by the current OpenClaw process. On shared hosts, unrelated processes could be terminated if they matched the pattern.

OpenClaw CLI process cleanup used system-wide process enumeration and pattern matching to terminate processes without verifying they were owned by the current OpenClaw process. On shared hosts, unrelated processes could be terminated if they matched the pattern.

On macOS, the Claude CLI keychain credential refresh path constructed a shell command to write the updated JSON blob into Keychain via security add-generic-password -w …. Because OAuth tokens are user-controlled data, this created an OS command injection risk. The fix avoids invoking a shell by using execFileSync("security", argv) and passing the updated keychain payload as a literal argument.

On macOS, the Claude CLI keychain credential refresh path constructed a shell command to write the updated JSON blob into Keychain via security add-generic-password -w …. Because OAuth tokens are user-controlled data, this created an OS command injection risk. The fix avoids invoking a shell by using execFileSync("security", argv) and passing the updated keychain payload as a literal argument.

A configuration injection issue in the Docker tool sandbox could allow dangerous Docker options (bind mounts, host networking, unconfined profiles) to be applied, enabling container escape or host data access.

Base64-backed media inputs could be decoded into Buffers before enforcing decoded-size budgets. An attacker supplying oversized base64 payloads can force large allocations, causing memory pressure and denial of service.

Base64-backed media inputs could be decoded into Buffers before enforcing decoded-size budgets. An attacker supplying oversized base64 payloads can force large allocations, causing memory pressure and denial of service.

Base64-backed media inputs could be decoded into Buffers before enforcing decoded-size budgets. An attacker supplying oversized base64 payloads can force large allocations, causing memory pressure and denial of service.

Base64-backed media inputs could be decoded into Buffers before enforcing decoded-size budgets. An attacker supplying oversized base64 payloads can force large allocations, causing memory pressure and denial of service.

OpenClaw previously accepted untrusted PATH sources in limited situations. In affected versions, this could cause OpenClaw to resolve and execute an unintended binary ("command hijacking") when running host commands. This issue primarily matters when OpenClaw is relying on allowlist/safe-bin protections and expects PATH to be trustworthy.

OpenClaw previously accepted untrusted PATH sources in limited situations. In affected versions, this could cause OpenClaw to resolve and execute an unintended binary ("command hijacking") when running host commands. This issue primarily matters when OpenClaw is relying on allowlist/safe-bin protections and expects PATH to be trustworthy.

A Twilio webhook signature-verification bypass in the voice-call extension could allow unauthenticated webhook requests when a specific ngrok free-tier compatibility option is enabled.

A Twilio webhook signature-verification bypass in the voice-call extension could allow unauthenticated webhook requests when a specific ngrok free-tier compatibility option is enabled.

Telegram allowlist authorization could match on @username (mutable/recyclable) instead of immutable numeric sender IDs.

Telegram allowlist authorization could match on @username (mutable/recyclable) instead of immutable numeric sender IDs.

Telegram allowlist authorization could match on @username (mutable/recyclable) instead of immutable numeric sender IDs.

Telegram allowlist authorization could match on @username (mutable/recyclable) instead of immutable numeric sender IDs.

When Slack DMs are configured with dmPolicy=open, the Slack slash-command handler incorrectly treated any DM sender as command-authorized. This allowed any Slack user who could DM the bot to execute privileged slash commands via DM, bypassing intended allowlist/access-group restrictions.

When Slack DMs are configured with dmPolicy=open, the Slack slash-command handler incorrectly treated any DM sender as command-authorized. This allowed any Slack user who could DM the bot to execute privileged slash commands via DM, bypassing intended allowlist/access-group restrictions.

Shared-agent, multi-user, less-trusted environments: session-tool access could expose transcript content across peer sessions. Single-agent or trusted environments: practical impact is limited. Telegram webhook mode: account-level secret wiring could be missed unless an explicit monitor webhook secret override was provided.

Inter-session messages sent via sessions_send could be interpreted as direct end-user instructions because they were persisted as role: "user" without provenance metadata.

Under iMessage groupPolicy=allowlist, group authorization could be satisfied by sender identities coming from the DM pairing store, broadening DM trust into group contexts.

Under iMessage groupPolicy=allowlist, group authorization could be satisfied by sender identities coming from the DM pairing store, broadening DM trust into group contexts.

The Feishu extension could fetch attacker-controlled remote URLs in two paths without SSRF protections: sendMediaFeishu(mediaUrl) Feishu DocX markdown image URLs (write/append -> image processing)

The Feishu extension could fetch attacker-controlled remote URLs in two paths without SSRF protections: sendMediaFeishu(mediaUrl) Feishu DocX markdown image URLs (write/append -> image processing)

openclaw could start the sandbox browser bridge server without authentication. When the sandboxed browser is enabled, openclaw runs a local (loopback) HTTP bridge that exposes browser control endpoints (for example /profiles, /tabs, /tabs/open, /agent/*). Due to missing auth wiring in the sandbox initialization path, that bridge server accepted requests without requiring gateway auth.

openclaw could start the sandbox browser bridge server without authentication. When the sandboxed browser is enabled, openclaw runs a local (loopback) HTTP bridge that exposes browser control endpoints (for example /profiles, /tabs, /tabs/open, /agent/*). Due to missing auth wiring in the sandbox initialization path, that bridge server accepted requests without requiring gateway auth.

Authenticated attackers can read arbitrary files from the Gateway host by supplying absolute paths or path traversal sequences to the browser tool's upload action. The server passed these paths to Playwright's setInputFiles() APIs without restricting them to a safe root. Severity remains High due to the impact (arbitrary local file read on the Gateway host), even though exploitation requires authenticated access.

OpenClaw’s browser control API accepted user-supplied output paths for trace/download files without consistently constraining writes to OpenClaw-managed temporary directories.

OpenClaw’s browser control API accepted user-supplied output paths for trace/download files without consistently constraining writes to OpenClaw-managed temporary directories.

OpenClaw browser download helpers accepted an unsanitized output path. When invoked via the browser control gateway routes, this allowed path traversal to write downloads outside the intended OpenClaw temp downloads directory. This issue is not exposed via the AI agent tool schema (no download action). Exploitation requires authenticated CLI access or an authenticated gateway RPC token.

The BlueBubbles extension accepted attacker-controlled local filesystem paths via mediaPath and could read arbitrary local files from disk before sending them as media attachments.

The BlueBubbles extension accepted attacker-controlled local filesystem paths via mediaPath and could read arbitrary local files from disk before sending them as media attachments.

Command injection in the maintainer/dev script scripts/update-clawtributors.ts.

A bug in download skill installation allowed targetDir values from skill frontmatter to resolve outside the per-skill tools directory if not strictly validated. In the admin-only skills.install flow, this could write files outside the intended install sandbox.

When multiple Google Chat webhook targets are registered on the same HTTP path, and request verification succeeds for more than one target, inbound webhook events could be routed by first-match semantics. This can cause cross-account policy/context misrouting.

When multiple Google Chat webhook targets are registered on the same HTTP path, and request verification succeeds for more than one target, inbound webhook events could be routed by first-match semantics. This can cause cross-account policy/context misrouting.

When multiple Google Chat webhook targets are registered on the same HTTP path, and request verification succeeds for more than one target, inbound webhook events could be routed by first-match semantics. This can cause cross-account policy/context misrouting.

When multiple Google Chat webhook targets are registered on the same HTTP path, and request verification succeeds for more than one target, inbound webhook events could be routed by first-match semantics. This can cause cross-account policy/context misrouting.

OpenClaw's exec-approvals allowlist supports a small set of "safe bins" intended to be stdin-only (no positional file arguments) when running tools.exec.host=gateway|node with security=allowlist. In affected configurations, the allowlist validation checked pre-expansion argv tokens, but execution used a real shell (sh -c) which expands globs and environment variables. This allowed safe bins like head, tail, or grep to read arbitrary local files via tokens such as * or $HOME/… without triggering …

OpenClaw's exec-approvals allowlist supports a small set of "safe bins" intended to be stdin-only (no positional file arguments) when running tools.exec.host=gateway|node with security=allowlist. In affected configurations, the allowlist validation checked pre-expansion argv tokens, but execution used a real shell (sh -c) which expands globs and environment variables. This allowed safe bins like head, tail, or grep to read arbitrary local files via tokens such as * or $HOME/… without triggering …

The manual Chutes OAuth login flow could accept attacker-controlled callback input in a way that bypassed OAuth CSRF state validation, potentially resulting in credential substitution.

The manual Chutes OAuth login flow could accept attacker-controlled callback input in a way that bypassed OAuth CSRF state validation, potentially resulting in credential substitution.

Discovery beacons (Bonjour/mDNS and DNS-SD) include TXT records such as lanHost, tailnetDns, gatewayPort, and gatewayTlsSha256. TXT records are unauthenticated. Prior to the fix, some clients treated TXT values as authoritative routing/pinning inputs: iOS and macOS: used TXT-provided host hints (lanHost/tailnetDns) and ports (gatewayPort) to build the connection URL. iOS and Android: allowed the discovery-provided TLS fingerprint (gatewayTlsSha256) to override a previously stored TLS pin. On a shared/untrusted LAN, an attacker …

Stored XSS in the OpenClaw Control UI when rendering assistant identity (name/avatar) into an inline <script> tag without script-context-safe escaping. A crafted value containing </script> could break out of the script tag and execute attacker-controlled JavaScript in the Control UI origin.

The optional Tlon (Urbit) extension previously accepted a user-provided base URL for authentication and used it to construct an outbound HTTP request, enabling server-side request forgery (SSRF) in affected deployments.

The optional Tlon (Urbit) extension previously accepted a user-provided base URL for authentication and used it to construct an outbound HTTP request, enabling server-side request forgery (SSRF) in affected deployments.

OpenClaw Gateway supports hook mappings with optional JavaScript/TypeScript transform modules. In affected versions, the gateway did not sufficiently constrain configured module paths before passing them to dynamic import(). Under some configurations, a user who can modify gateway configuration could cause the gateway process to load and execute an unintended local module.

OpenClaw Gateway supports hook mappings with optional JavaScript/TypeScript transform modules. In affected versions, the gateway did not sufficiently constrain configured module paths before passing them to dynamic import(). Under some configurations, a user who can modify gateway configuration could cause the gateway process to load and execute an unintended local module.

Multiple webhook handlers accepted and buffered request bodies without a strict unified byte/time limit. A remote unauthenticated attacker could send oversized payloads and cause memory pressure, degrading availability.

Multiple webhook handlers accepted and buffered request bodies without a strict unified byte/time limit. A remote unauthenticated attacker could send oversized payloads and cause memory pressure, degrading availability.

Multiple webhook handlers accepted and buffered request bodies without a strict unified byte/time limit. A remote unauthenticated attacker could send oversized payloads and cause memory pressure, degrading availability.

Multiple webhook handlers accepted and buffered request bodies without a strict unified byte/time limit. A remote unauthenticated attacker could send oversized payloads and cause memory pressure, degrading availability.

URL-backed media fetch handling allocated the entire response payload in memory (arrayBuffer) before enforcing maxBytes, allowing oversized responses to cause memory exhaustion.

URL-backed media fetch handling allocated the entire response payload in memory (arrayBuffer) before enforcing maxBytes, allowing oversized responses to cause memory exhaustion.

Archive extraction lacked strict resource budgets, allowing high-expansion ZIP/TAR archives to consume excessive CPU/memory/disk during install/update flows.

Archive extraction lacked strict resource budgets, allowing high-expansion ZIP/TAR archives to consume excessive CPU/memory/disk during install/update flows.

Archive extraction lacked strict resource budgets, allowing high-expansion ZIP/TAR archives to consume excessive CPU/memory/disk during install/update flows.

Archive extraction lacked strict resource budgets, allowing high-expansion ZIP/TAR archives to consume excessive CPU/memory/disk during install/update flows.

Browser-facing localhost mutation routes accepted cross-origin browser requests without explicit Origin/Referer validation. Loopback binding reduces remote exposure but does not prevent browser-initiated requests from malicious origins.

Browser-facing localhost mutation routes accepted cross-origin browser requests without explicit Origin/Referer validation. Loopback binding reduces remote exposure but does not prevent browser-initiated requests from malicious origins.

A security vulnerability has been discovered in how the input.parsed_path field is constructed. HTTP request paths are treated as full URIs when parsed; interpreting leading path segments prefixed with double slashes (//) as authority components, and therefore dropping them from the parsed path. This creates a path interpretation mismatch between authorization policies and backend servers, enabling attackers to bypass access controls by crafting requests where the authorization filter evaluates a …

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.

NVIDIA NeMo Framework contains a vulnerability where malicious data could cause remote code execution. A successful exploit of this vulnerability might lead to code execution, escalation of privileges, information disclosure, and data tampering.

Nokogiri's CRuby extension fails to check the return value from xmlC14NExecute in the method Nokogiri::XML::Document#canonicalize and Nokogiri::XML::Node#canonicalize. When canonicalization fails, an empty string is returned instead of raising an exception. This incorrect return value may allow downstream libraries to accept invalid or incomplete canonicalized XML, which has been demonstrated to enable signature validation bypass in SAML libraries. JRuby is not affected, as the Java implementation correctly raises RuntimeError on canonicalization …

A critical vulnerability exists in the NLTK downloader component of nltk/nltk, affecting all versions. The _unzip_iter function in nltk/downloader.py uses zipfile.extractall() without performing path validation or security checks. This allows attackers to craft malicious zip packages that, when downloaded and extracted by NLTK, can execute arbitrary code. The vulnerability arises because NLTK assumes all downloaded packages are trusted and extracts them without validation. If a malicious package contains Python files, …

minimatch is vulnerable to Regular Expression Denial of Service (ReDoS) when a glob pattern contains many consecutive * wildcards followed by a literal character that doesn't appear in the test string. Each * compiles to a separate [^/]*? regex group, and when the match fails, V8's regex engine backtracks exponentially across all possible splits. The time complexity is O(4^N) where N is the number of * characters. With N=15, a …

A flaw has been found in mingSoft MCMS 6.1.1. The affected element is an unknown function of the file /ms/file/uploadTemplate.do of the component Template Archive Handler. Executing a manipulation of the argument File can lead to unrestricted upload. The attack can be launched remotely. The exploit has been published and may be used.

SQL Injection in IPv6 Address Search functionality via address parameter* A SQL injection vulnerability exists in the ajax_table.php endpoint. The application fails to properly sanitize or parameterize user input when processing IPv6 address searches. Specifically, the address parameter is split into an address and a prefix, and the prefix portion is directly concatenated into the SQL query string without validation. This allows an attacker to inject arbitrary SQL commands, potentially …

A time-based blind SQL injection vulnerability exists in address-search.inc.php via the address parameter. When a crafted subnet prefix is supplied, the prefix value is concatenated directly into an SQL query without proper parameter binding, allowing an attacker to manipulate query logic and infer database information through time-based conditional responses.

The unit parameter in Custom OID functionality lacks strip_tags() sanitization while other fields (name, oid, datatype) are sanitized. The unsanitized value is stored in the database and rendered without HTML escaping, allowing Stored XSS.

A stored Cross-Site Scripting (XSS) vulnerability exists in LibreNMS (<= 25.12.0) in the creation of Alert Rules. This allows a user with the admin role to inject malicious JavaScript, which will be executed when the alert rules page is viewed.

reflected xss via email field

/port-groups name Stored Cross-Site Scripting HTTP POST Request-URI(s): "/port-groups" Vulnerable parameter(s): "name" Attacker must be authenticated with "admin" privileges. When a user adds a port group, an HTTP POST request is sent to the Request-URI "/port-groups". The name of the newly created port group is stored in the value of the name parameter. After the port group is created, the entry is displayed along with some relevant buttons like Edit …

/device-groups name Stored Cross-Site Scripting HTTP POST Request-URI(s): "/device-groups" Vulnerable parameter(s): "name" Attacker must be authenticated with "admin" privileges. When a user adds a device group, an HTTP POST request is sent to the Request-URI "/device-groups". The name of the newly created device group is stored in the value of the name parameter. After the device group is created, the entry is displayed along with some relevant buttons like Rediscover …

A critical security vulnerability exists in the LibreDesk Webhooks module that allows an authenticated "Application Admin" to compromise the underlying cloud infrastructure or internal corporate network where this service is being hosted. The application fails to validate destination URLs for webhooks. This allows an attacker to force the server to make HTTP requests to arbitrary internal destinations.

TensorFlow / 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 load_weights() to read from an arbitrary readable filesystem path. The bytes pulled from that path populate model tensors and become observable through inference or subsequent re-save operations. Keras “safe mode” only guards object deserialization and does not cover weight I/O, so this behaviour …

Jenkins 2.483 through 2.550 (both inclusive), LTS 2.492.1 through 2.541.1 (both inclusive) does not escape the user-provided description of the "Mark temporarily offline" offline cause, resulting in a stored cross-site scripting (XSS) vulnerability exploitable by attackers with Agent/Configure or Agent/Disconnect permission.

Jenkins 2.550 and earlier, LTS 2.541.1 and earlier accepts Run Parameter values that refer to builds the user submitting the build does not have access to, allowing attackers with Item/Build and Item/Configure permission to obtain information about the existence of jobs, the existence of builds, and if a specified build exists, its display name.

This is a remote code execution (RCE) vulnerability. Node.js automatically imports **/*.plugin.{js,mjs} files including those from node_modules, so any malicious package with a .plugin.js file could execute arbitrary code when installed or required. All projects using this loading behavior are affected, especially those installing untrusted packages.

A weakness has been identified in huggingface smolagents 1.24.0. Impacted is the function requests.get/requests.post of the component LocalPythonExecutor. Executing a manipulation can lead to server-side request forgery. It is possible to launch the attack remotely. The exploit has been made available to the public and could be used for attacks. The vendor was contacted early about this disclosure but did not respond in any way.

Through a flaw in the ECIES cryptography implementation, an attacker may be able to extract bits of the p2p node key.

An attacker can cause high memory usage by sending a specially-crafted p2p message. More details to be released later.

A vulnerable node can be forced to shutdown/crash using a specially crafted message. More details to be released later.

A SQL injection vulnerability existed in Ghost's Content API that allowed unauthenticated attackers to read arbitrary data from the database.

(*Point).MultiScalarMult failed to initialize its receiver. If the method was called on an initialized point that is not the identity point, MultiScalarMult produced an incorrect result. If the method was called on an uninitialized point, the behavior was undefined. In particular, if the receiver was the zero value, MultiScalarMult returned an invalid point that compared Equal to every point. Note that MultiScalarMult is a rarely used advanced API. For example, …

fabric.js applies escapeXml() to text content during SVG export (src/shapes/Text/TextSVGExportMixin.ts:186) but fails to apply it to other user-controlled string values that are interpolated into SVG attribute markup. When attacker-controlled JSON is loaded via loadFromJSON() and later exported via toSVG(), the unescaped values break out of XML attributes and inject arbitrary SVG elements including event handlers.

Command Injection via Unsanitized locate Output in versions() — systeminformation Package: systeminformation (npm) Tested Version: 5.30.7 Affected Platform: Linux Author: Sebastian Hildebrandt Weekly Downloads: ~5,000,000+ Repository: https://github.com/sebhildebrandt/systeminformation Severity: Medium CWE: CWE-78 (OS Command Injection) The Vulnerable Code Path Inside the versions() function, when detecting the PostgreSQL version on Linux, the code does this: // lib/osinfo.js — lines 770-776 exec('locate bin/postgres', (error, stdout) => { if (!error) { const postgresqlBin = …

tar.extract() in Node tar allows an attacker-controlled archive to create a hardlink inside the extraction directory that points to a file outside the extraction root, using default options. This enables arbitrary file read and write as the extracting user (no root, no chmod, no preservePaths). Severity is high because the primitive bypasses path protections and turns archive extraction into a direct filesystem access primitive.

The SSH management console did not validate the passed input while adding the SSH host key, which could lead to an argument injection to ssh-add.

Gogs exposes unauthenticated file upload endpoints by default. When the global RequireSigninView setting is disabled (default), any remote user can upload arbitrary files to the server via /releases/attachments and /issues/attachments. This enables the instance to be abused as a public file host, potentially leading to disk exhaustion, content hosting, or delivery of malware. CSRF tokens do not mitigate this attack due to same-origin cookie issuance.

The #[contractimpl] macro contains a bug in how it wires up function calls. In Rust, you can define functions on a type in two ways: Directly on the type as an inherent function: impl MyContract { fn value() { … } } Through a trait impl Trait for MyContract { fn value() { … } } These are two separate functions that happen to share the same name. Rust has …

Rack::Directory generates an HTML directory index where each file entry is rendered as a clickable link. If a file exists on disk whose basename begins with the javascript: scheme (e.g. javascript:alert(1)), the generated index includes an anchor whose href attribute is exactly javascript:alert(1). Clicking this entry executes arbitrary JavaScript in the context of the hosting application. This results in a client-side XSS condition in directory listings generated by Rack::Directory.

Description: A vulnerability in the API Server of Skill Scanner could allow a unauthenticated, remote attacker to interact with the server API and either trigger a denial of service (DoS) condition or upload arbitrary files. This vulnerability is due to an erroneous binding to multiple interfaces. An attacker could exploit this vulnerability by sending API requests to a device exposing the affected API Server. A successful exploit could allow the …

Rack::Directory’s path check used a string prefix match on the expanded path. A request like /../root_example/ can escape the configured root if the target path starts with the root string, allowing directory listing outside the intended root.

Deleting a user account with SFTP access or changing the user's password does not immediately terminate existing SFTP sessions, allowing continued filesystem access after credentials are revoked. This can result in unintended and unauthorized access to server files even after administrators believe access has been fully invalidated.

Deleting a user account with SFTP access or changing the user's password does not immediately terminate existing SFTP sessions, allowing continued filesystem access after credentials are revoked. This can result in unintended and unauthorized access to server files even after administrators believe access has been fully invalidated.

A missing authorization check in multiple controllers allows any user with access to a node secret token to fetch information about any server on a Pterodactyl instance, even if that server is associated with a different node. This issue stems from missing logic to verify that the node requesting server data is the same node that the server is associated with. Any authenticated Wings node can retrieve server installation scripts …

On Windows nodes, exec requests were executed via cmd.exe /d /s /c <rawCommand>. In allowlist/approval-gated mode, the allowlist analysis did not model Windows cmd.exe parsing and metacharacter behavior. A crafted command string could cause cmd.exe to interpret additional operations (for example command chaining via &, or expansion via %…% / !…!) beyond what was allowlisted/approved.

On Windows nodes, exec requests were executed via cmd.exe /d /s /c <rawCommand>. In allowlist/approval-gated mode, the allowlist analysis did not model Windows cmd.exe parsing and metacharacter behavior. A crafted command string could cause cmd.exe to interpret additional operations (for example command chaining via &, or expansion via %…% / !…!) beyond what was allowlisted/approved.

The OpenClaw Nostr channel plugin (optional, disabled by default, installed separately) exposes profile management HTTP endpoints under /api/channels/nostr/:accountId/profile (GET/PUT) and /api/channels/nostr/:accountId/profile/import (POST). In affected versions, these routes were dispatched via the gateway plugin HTTP layer without requiring gateway authentication, allowing unauthenticated remote callers to read or mutate the Nostr profile and persist changes to the gateway config. Profile updates are also published as a signed Nostr kind:0 event using the …

The OpenClaw Nostr channel plugin (optional, disabled by default, installed separately) exposes profile management HTTP endpoints under /api/channels/nostr/:accountId/profile (GET/PUT) and /api/channels/nostr/:accountId/profile/import (POST). In affected versions, these routes were dispatched via the gateway plugin HTTP layer without requiring gateway authentication, allowing unauthenticated remote callers to read or mutate the Nostr profile and persist changes to the gateway config. Profile updates are also published as a signed Nostr kind:0 event using the …

The gateway WebSocket connect handshake could allow skipping device identity checks when auth.token was present but not yet validated.

The gateway WebSocket connect handshake could allow skipping device identity checks when auth.token was present but not yet validated.

The Chrome extension relay (ensureChromeExtensionRelayServer) previously treated wildcard hosts (0.0.0.0 / ::) as loopback, which could make it bind the relay HTTP/WS server to all interfaces when a wildcard cdpUrl was passed.

The Chrome extension relay (ensureChromeExtensionRelayServer) previously treated wildcard hosts (0.0.0.0 / ::) as loopback, which could make it bind the relay HTTP/WS server to all interfaces when a wildcard cdpUrl was passed.

In affected versions, the Browser Relay /cdp WebSocket endpoint did not require an authentication token. As a result, a website running in the browser could potentially connect to the local relay (via loopback WebSocket) and use CDP to access cookies from other open tabs and run JavaScript in the context of other tabs.

In affected versions, the Browser Relay /cdp WebSocket endpoint did not require an authentication token. As a result, a website running in the browser could potentially connect to the local relay (via loopback WebSocket) and use CDP to access cookies from other open tabs and run JavaScript in the context of other tabs.

In affected versions, the Browser Relay /cdp WebSocket endpoint did not require an authentication token. As a result, a website running in the browser could potentially connect to the local relay (via loopback WebSocket) and use CDP to access cookies from other open tabs and run JavaScript in the context of other tabs.

In affected versions, the Browser Relay /cdp WebSocket endpoint did not require an authentication token. As a result, a website running in the browser could potentially connect to the local relay (via loopback WebSocket) and use CDP to access cookies from other open tabs and run JavaScript in the context of other tabs.

In the optional Twitch channel plugin (extensions/twitch), allowFrom is documented as a hard allowlist of Twitch user IDs, but it was not enforced as a hard gate. If allowedRoles is unset or empty, the access control path defaulted to allow, so any Twitch user who could mention the bot could reach the agent dispatch pipeline. Scope note: This only affects deployments that installed and enabled the Twitch plugin. Core OpenClaw …

In the optional Twitch channel plugin (extensions/twitch), allowFrom is documented as a hard allowlist of Twitch user IDs, but it was not enforced as a hard gate. If allowedRoles is unset or empty, the access control path defaulted to allow, so any Twitch user who could mention the bot could reach the agent dispatch pipeline. Scope note: This only affects deployments that installed and enabled the Twitch plugin. Core OpenClaw …

skills.status could disclose secrets to operator.read clients by returning raw resolved config values in configChecks for skill requires.config paths.

In certain reverse-proxy / forwarding setups, webhook verification can be bypassed if untrusted forwarded headers are accepted.

In certain reverse-proxy / forwarding setups, webhook verification can be bypassed if untrusted forwarded headers are accepted.

In certain reverse-proxy / forwarding setups, webhook verification can be bypassed if untrusted forwarded headers are accepted.

In certain reverse-proxy / forwarding setups, webhook verification can be bypassed if untrusted forwarded headers are accepted.

A mismatch between rawCommand and command[] in the node host system.run handler could cause allowlist/approval evaluation to be performed on one command while executing a different argv.

NOTE: This only affects deployments that enable the optional MS Teams extension (Teams channel). If you do not use MS Teams, you are not impacted. When OpenClaw downloads inbound MS Teams attachments / inline images, it may retry a URL with an Authorization: Bearer <token> header after receiving 401 or 403. Because the default download allowlist uses suffix matching (and includes some multi-tenant suffix domains), a message that references an …

NOTE: This only affects deployments that enable the optional MS Teams extension (Teams channel). If you do not use MS Teams, you are not impacted. When OpenClaw downloads inbound MS Teams attachments / inline images, it may retry a URL with an Authorization: Bearer <token> header after receiving 401 or 403. Because the default download allowlist uses suffix matching (and includes some multi-tenant suffix domains), a message that references an …

OpenClaw macOS desktop client registers the openclaw:// URL scheme. For openclaw://agent deep links without an unattended key, the app shows a confirmation dialog that previously displayed only the first 240 characters of the message, but executed the full message after the user clicked "Run". At the time of writing, the OpenClaw macOS desktop client is still in beta. An attacker could pad the message with whitespace to push a malicious …

In openclaw versions prior to 2026.2.13, OpenClaw logged certain WebSocket request headers (including Origin and User-Agent) without neutralization or length limits on the "closed before connect" path. If an unauthenticated client can reach the gateway and send crafted header values, those values may be written into core logs. Under workflows where logs are later read or interpreted by an LLM (for example via AI-assisted debugging), this can increase the risk …

In affected versions, OpenClaw's optional @openclaw/voice-call plugin Telnyx webhook handler could accept unsigned inbound webhook requests when telnyx.publicKey was not configured, allowing unauthenticated callers to forge Telnyx events. This only impacts deployments where the Voice Call plugin is installed, enabled, and the webhook endpoint is reachable from the attacker (for example, publicly exposed via a tunnel/proxy).

The issue is not deterministic session keys by itself. The exploitable path was accepting externally supplied sessionKey values on authenticated hook ingress, allowing a hook token holder to route messages into chosen sessions.

An authentication bypass in the optional voice-call extension/plugin allowed unapproved or anonymous callers to reach the voice-call agent when inbound policy was set to allowlist or pairing. Deployments that do not install/enable the voice-call extension are not affected.

An authentication bypass in the optional voice-call extension/plugin allowed unapproved or anonymous callers to reach the voice-call agent when inbound policy was set to allowlist or pairing. Deployments that do not install/enable the voice-call extension are not affected.

Exec approvals allowlist bypass via command substitution/backticks inside double quotes.

Exec approvals allowlist bypass via command substitution/backticks inside double quotes.

In OpenClaw versions prior to 2026.2.12, the gateway accepted an untrusted sessionFile path when resolving the session transcript file. This could allow an authenticated gateway client to create and append OpenClaw session transcript records at an arbitrary path on the gateway host.

In OpenClaw versions prior to 2026.2.12, the gateway accepted an untrusted sessionFile path when resolving the session transcript file. This could allow an authenticated gateway client to create and append OpenClaw session transcript records at an arbitrary path on the gateway host.

The BlueBubbles webhook handler previously treated any request whose socket remoteAddress was loopback (127.0.0.1, ::1, ::ffff:127.0.0.1) as authenticated. When OpenClaw Gateway is behind a reverse proxy (Tailscale Serve/Funnel, nginx, Cloudflare Tunnel, ngrok), the proxy typically connects to the gateway over loopback, allowing unauthenticated remote requests to bypass the configured webhook password. This could allow an attacker who can reach the proxy endpoint to inject arbitrary inbound BlueBubbles message/reaction events.

The BlueBubbles webhook handler previously treated any request whose socket remoteAddress was loopback (127.0.0.1, ::1, ::ffff:127.0.0.1) as authenticated. When OpenClaw Gateway is behind a reverse proxy (Tailscale Serve/Funnel, nginx, Cloudflare Tunnel, ngrok), the proxy typically connects to the gateway over loopback, allowing unauthenticated remote requests to bypass the configured webhook password. This could allow an attacker who can reach the proxy endpoint to inject arbitrary inbound BlueBubbles message/reaction events.

In Telegram webhook mode, if channels.telegram.webhookSecret is not set, OpenClaw may accept webhook HTTP requests without verifying Telegram’s secret token header. In deployments where the webhook endpoint is reachable by an attacker, this can allow forged Telegram updates (for example spoofing message.from.id). Note: Telegram webhook mode is not enabled by default. It is enabled only when channels.telegram.webhookUrl is configured.

OpenClaw's SSRF protection could be bypassed using full-form IPv4-mapped IPv6 literals such as 0:0:0:0:0:ffff:7f00:1 (which is 127.0.0.1). This could allow requests that should be blocked (loopback / private network / link-local metadata) to pass the SSRF guard. Vulnerable component: SSRF guard (src/infra/net/ssrf.ts) Issue type: SSRF protection bypass

When Telegram webhook mode is enabled without a configured webhook secret, OpenClaw may accept unauthenticated HTTP POST requests at the Telegram webhook endpoint and trust attacker-controlled update JSON. This can allow forged Telegram updates that spoof message.from.id / chat.id, potentially bypassing sender allowlists and executing privileged bot commands.

When Telegram webhook mode is enabled without a configured webhook secret, OpenClaw may accept unauthenticated HTTP POST requests at the Telegram webhook endpoint and trust attacker-controlled update JSON. This can allow forged Telegram updates that spoof message.from.id / chat.id, potentially bypassing sender allowlists and executing privileged bot commands.

OpenClaw's plugin installation path derivation could be abused by a malicious plugin package.json name to escape the intended extensions directory and write files to a parent directory.

OpenClaw's plugin installation path derivation could be abused by a malicious plugin package.json name to escape the intended extensions directory and write files to a parent directory.

OpenClaw Matrix DM allowlist matching could be bypassed in certain configurations. Matrix support ships as an optional plugin (not bundled with the core install), so this only affects deployments that have installed and enabled the Matrix plugin.

OpenClaw Matrix DM allowlist matching could be bypassed in certain configurations. Matrix support ships as an optional plugin (not bundled with the core install), so this only affects deployments that have installed and enabled the Matrix plugin.

The Feishu extension previously allowed sendMediaFeishu to treat attacker-controlled mediaUrl values as local filesystem paths and read them directly.

Google Chat allowlisting supports matching by sender email in addition to immutable sender resource name (users/<id>). This weakens identity binding if a deployment assumes allowlists are strictly keyed by immutable principals.

Google Chat allowlisting supports matching by sender email in addition to immutable sender resource name (users/<id>). This weakens identity binding if a deployment assumes allowlists are strictly keyed by immutable principals.

The Gateway tool accepted a tool-supplied gatewayUrl without sufficient restrictions, which could cause the OpenClaw host to attempt outbound WebSocket connections to user-specified targets.

In affected versions, the optional BlueBubbles iMessage channel plugin could accept webhook requests as authenticated based only on the TCP peer address being loopback (127.0.0.1, ::1, ::ffff:127.0.0.1) even when the configured webhook secret was missing or incorrect. This does not affect the default iMessage integration unless BlueBubbles is installed and enabled.

In affected versions, the optional BlueBubbles iMessage channel plugin could accept webhook requests as authenticated based only on the TCP peer address being loopback (127.0.0.1, ::1, ::ffff:127.0.0.1) even when the configured webhook secret was missing or incorrect. This does not affect the default iMessage integration unless BlueBubbles is installed and enabled.

A gateway client authenticated with a device token scoped only to operator.write (without operator.approvals) could approve/deny pending exec approval requests by sending a chat message containing the built-in /approve command. exec.approval.resolve is correctly scoped to operator.approvals for direct RPC calls, but the /approve command path invoked it via an internal privileged gateway client.

A gateway client authenticated with a device token scoped only to operator.write (without operator.approvals) could approve/deny pending exec approval requests by sending a chat message containing the built-in /approve command. exec.approval.resolve is correctly scoped to operator.approvals for direct RPC calls, but the /approve command path invoked it via an internal privileged gateway client.

Versions of the openclaw npm package prior to 2026.2.2 could be coerced into fetching arbitrary http(s) URLs during attachment/media hydration. An attacker who can influence the media URL (for example via model-controlled sendAttachment or auto-reply media URLs) could trigger SSRF to internal resources and exfiltrate the fetched bytes as an outbound attachment.

Versions of the openclaw npm package prior to 2026.2.2 could be coerced into fetching arbitrary http(s) URLs during attachment/media hydration. An attacker who can influence the media URL (for example via model-controlled sendAttachment or auto-reply media URLs) could trigger SSRF to internal resources and exfiltrate the fetched bytes as an outbound attachment.

A server-side request forgery (SSRF) vulnerability in the Image tool allowed attackers to force OpenClaw to make HTTP requests to arbitrary internal or restricted network targets.

When the Slack integration is enabled, Slack channel metadata (topic/description) could be incorporated into the model's system prompt.

In affected versions of the optional Nextcloud Talk plugin (installed separately; not bundled with the core OpenClaw install), an untrusted webhook field (actor.name, display name) could be treated as an allowlist identifier. An attacker could change their Nextcloud display name to match an allowlisted user ID and bypass DM or room allowlists.

In affected versions of the optional Nextcloud Talk plugin (installed separately; not bundled with the core OpenClaw install), an untrusted webhook field (actor.name, display name) could be treated as an allowlist identifier. An attacker could change their Nextcloud display name to match an allowlisted user ID and bypass DM or room allowlists.

Indico makes outgoing requests to user-provides URLs in various places. This is mostly intentional and part of Indico's functionality, but of course it is never intended to let you access "special" targets such as localhost or cloud metadata endpoints.

There is a Cross-Site-Scripting vulnerability when uploading certain file types as materials.

An issue was discovered in httpsig-hyper where Digest header verification could incorrectly succeed due to misuse of Rust's matches! macro. Specifically, the comparison: if matches!(digest, _expected_digest) treated _expected_digest as a pattern binding rather than a value comparison, resulting in unconditional success of the match expression. As a consequence, digest verification could incorrectly return success even when the computed digest did not match the expected value. Applications relying on Digest verification …

A broken access control vulnerability in Gogs allows authenticated users with write access to any repository to modify labels belonging to other repositories. The UpdateLabel function in the Web UI (internal/route/repo/issue.go) fails to verify that the label being modified belongs to the repository specified in the URL path, enabling cross-repository label tampering attacks.

An access control bypass vulnerability in Gogs web interface allows any repository collaborator with Write permissions to delete protected branches (including the default branch) by sending a direct POST request, completely bypassing the branch protection mechanism. This vulnerability enables privilege escalation from Write to Admin level, allowing low-privilege users to perform dangerous operations that should be restricted to administrators only. Although Git Hook layer correctly prevents protected branch deletion via …

The POST /:owner/:repo/issues/comments/:id/delete endpoint does not verify that the comment belongs to the repository specified in the URL. This allows a repository administrator to delete comments from any other repository by supplying arbitrary comment IDs, bypassing authorization controls.

The XML parser can be forced to do an unlimited amount of entity expansion. With a very small XML input, it’s possible to make the parser spend seconds or even minutes processing a single request, effectively freezing the application.

Multiple shared maps are accessed without consistent synchronization across goroutines. Under concurrent activity, Go runtime can trigger fatal error: concurrent map read and map write, causing C2 process crash (availability loss).

On Windows, Echo’s middleware.Static using the default filesystem allows path traversal via backslashes, enabling unauthenticated remote file read outside the static root.

A critical cryptographic vulnerability in the TypeScript SDK's BRC-104 authentication implementation caused incorrect signature data preparation, resulting in signature incompatibility between SDK implementations and potential authentication bypass scenarios.

Improper Input Validation vulnerability in Apache Tomcat Native, Apache Tomcat. When using an OCSP responder, Tomcat Native (and Tomcat's FFM port of the Tomcat Native code) did not complete verification or freshness checks on the OCSP response which could allow certificate revocation to be bypassed. This issue affects Apache Tomcat: from 11.0.0-M1 through 11.0.17, from 10.1.0-M7 through 10.1.51, from 9.0.83 through 9.0.114. Apache Tomcat users are recommended to upgrade to …

Improper Input Validation vulnerability in Apache Tomcat Native, Apache Tomcat. When using an OCSP responder, Tomcat Native did not complete verification or freshness checks on the OCSP response which could allow certificate revocation to be bypassed. The vulnerable code is in the process_ocsp_response() function in sslutils.c, which was missing calls to OCSP_basic_verify(), OCSP_check_validity(), and OCSP_check_nonce(). This issue affects Apache Tomcat Native: from 1.3.0 through 1.3.4, from 2.0.0 through 2.0.11. The …

Improper Input Validation vulnerability in Apache Tomcat Native, Apache Tomcat. When using an OCSP responder, Tomcat Native (and Tomcat's FFM port of the Tomcat Native code) did not complete verification or freshness checks on the OCSP response which could allow certificate revocation to be bypassed. This issue affects Apache Tomcat Native: from 1.3.0 through 1.3.4, from 2.0.0 through 2.0.11; Apache Tomcat: from 11.0.0-M1 through 11.0.17, from 10.1.0-M7 through 10.1.51, from …

Improper Input Validation vulnerability in Apache Tomcat Native, Apache Tomcat. When using an OCSP responder, Tomcat Native (and Tomcat's FFM port of the Tomcat Native code) did not complete verification or freshness checks on the OCSP response which could allow certificate revocation to be bypassed. This issue affects Apache Tomcat Native: from 1.3.0 through 1.3.4, from 2.0.0 through 2.0.11; Apache Tomcat: from 11.0.0-M1 through 11.0.17, from 10.1.0-M7 through 10.1.51, from …

Improper Input Validation vulnerability in Apache Tomcat Native, Apache Tomcat. When using an OCSP responder, Tomcat Native (and Tomcat's FFM port of the Tomcat Native code) did not complete verification or freshness checks on the OCSP response which could allow certificate revocation to be bypassed. The vulnerable code is in the OpenSSLEngine class within the org.apache.tomcat.util.net.openssl.panama package, which is shipped in the tomcat-coyote-ffm artifact. This issue affects Apache Tomcat: from …

Improper Input Validation vulnerability in Apache Tomcat. Tomcat did not limit HTTP/0.9 requests to the GET method. If a security constraint was configured to allow HEAD requests to a URI but deny GET requests, the user could bypass that constraint on GET requests by sending a (specification invalid) HEAD request using HTTP/0.9. This issue affects Apache Tomcat: from 11.0.0-M1 through 11.0.14, from 10.1.0-M1 through 10.1.49, from 9.0.0.M1 through 9.0.112. Older, …

Improper Input Validation vulnerability in Apache Tomcat. Tomcat did not limit HTTP/0.9 requests to the GET method. If a security constraint was configured to allow HEAD requests to a URI but deny GET requests, the user could bypass that constraint on GET requests by sending a (specification invalid) HEAD request using HTTP/0.9. This issue affects Apache Tomcat: from 11.0.0-M1 through 11.0.14, from 10.1.0-M1 through 10.1.49, from 9.0.0.M1 through 9.0.112. Older, …

Improper Input Validation vulnerability in Apache Tomcat. Tomcat did not limit HTTP/0.9 requests to the GET method. If a security constraint was configured to allow HEAD requests to a URI but deny GET requests, the user could bypass that constraint on GET requests by sending a (specification invalid) HEAD request using HTTP/0.9. This issue affects Apache Tomcat: from 11.0.0-M1 through 11.0.14, from 10.1.0-M1 through 10.1.49, from 9.0.0.M1 through 9.0.112. Older, …

Improper Input Validation vulnerability. This issue affects Apache Tomcat: from 11.0.0-M1 through 11.0.14, from 10.1.0-M1 through 10.1.49, from 9.0.0-M1 through 9.0.112. The following versions were EOL at the time the CVE was created but are known to be affected: 8.5.0 through 8.5.100. Older EOL versions are not affected. Tomcat did not validate that the host name provided via the SNI extension was the same as the host name provided in …

Improper Input Validation vulnerability. This issue affects Apache Tomcat: from 11.0.0-M1 through 11.0.14, from 10.1.0-M1 through 10.1.49, from 9.0.0-M1 through 9.0.112. The following versions were EOL at the time the CVE was created but are known to be affected: 8.5.0 through 8.5.100. Older EOL versions are not affected. Tomcat did not validate that the host name provided via the SNI extension was the same as the host name provided in …

Improper Input Validation vulnerability. This issue affects Apache Tomcat: from 11.0.0-M1 through 11.0.14, from 10.1.0-M1 through 10.1.49, from 9.0.0-M1 through 9.0.112. The following versions were EOL at the time the CVE was created but are known to be affected: 8.5.0 through 8.5.100. Older EOL versions are not affected. Tomcat did not validate that the host name provided via the SNI extension was the same as the host name provided in …

Apache NiFi 1.1.0 through 2.7.2 are missing authorization when updating configuration properties on extension components that have specific Required Permissions based on the Restricted annotation. The Restricted annotation indicates additional privileges required to add the annotated component to the flow configuration, but framework authorization did not check restricted status when updating a component previously added. The missing authorization requires a more privileged user to add a restricted component to the …

Emails sent by pretix can utilize placeholders that will be filled with customer data. For example, when {name} is used in an email template, it will be replaced with the buyer's name for the final email. This mechanism contained two security-relevant bugs: It was possible to exfiltrate information about the pretix system through specially crafted placeholder names such as {event.init.code.co_filename}}. This way, an attacker with the ability to control email …

A security vulnerability has been detected in MindsDB up to 25.14.1. This vulnerability affects the function clear_filename of the file mindsdb/utilities/security.py of the component File Upload. Such manipulation leads to server-side request forgery. The attack may be performed from remote. The exploit has been disclosed publicly and may be used.

Mattermost versions 11.1.x <= 11.1.2, 10.11.x <= 10.11.9, 11.2.x <= 11.2.1 and Mattermost Plugin Zoom versions <=1.11.0 fail to validate user identity and post ownership in the {{/api/v1/askPMI}} endpoint which allows unauthorized users to start Zoom meetings as any user and overwrite arbitrary posts via direct API calls with manipulated user IDs and post data.. Mattermost Advisory ID: MMSA-2025-00534

Mattermost versions 11.1.x <= 11.1.2, 10.11.x <= 10.11.9, 11.2.x <= 11.2.1 and Mattermost Plugin Zoom versions <=1.11.0 fail to validate the authenticated user when processing {{/plugins/zoom/api/v1/channel-preference}}, which allows any logged-in user to change Zoom meeting restrictions for arbitrary channels via crafted API requests.. Mattermost Advisory ID: MMSA-2025-00558

Mattermost versions 11.1.x <= 11.1.2, 10.11.x <= 10.11.9, 11.2.x <= 11.2.1 fail to sanitize sensitive data in WebSocket messages which allows authenticated users to exfiltrate password hashes and MFA secrets via profile nickname updates or email verification events. Mattermost Advisory ID: MMSA-2025-00560

Mattermost versions 11.1.x <= 11.1.2, 10.11.x <= 10.11.9, 11.2.x <= 11.2.1 fail to sanitize sensitive data in WebSocket messages which allows authenticated users to exfiltrate password hashes and MFA secrets via profile nickname updates or email verification events. Mattermost Advisory ID: MMSA-2025-00560

Mattermost versions 11.1.x <= 11.1.2, 10.11.x <= 10.11.9, 11.2.x <= 11.2.1 fail to properly validate team membership when processing channel mentions which allows authenticated users to determine the existence of teams and their URL names via posting channel shortlinks and observing the channel_mentions property in the API response. Mattermost Advisory ID: MMSA-2025-00563

Mattermost versions 11.1.x <= 11.1.2, 10.11.x <= 10.11.9, 11.2.x <= 11.2.1 fail to properly validate team membership when processing channel mentions which allows authenticated users to determine the existence of teams and their URL names via posting channel shortlinks and observing the channel_mentions property in the API response. Mattermost Advisory ID: MMSA-2025-00563

Mattermost versions 11.1.x <= 11.1.2, 10.11.x <= 10.11.9, 11.2.x <= 11.2.1 fail to properly validate login method restrictions which allows an authenticated user to bypass SSO-only login requirements via userID-based authentication. Mattermost Advisory ID: MMSA-2025-00548

Mattermost versions 11.1.x <= 11.1.2, 10.11.x <= 10.11.9, 11.2.x <= 11.2.1 fail to properly validate login method restrictions which allows an authenticated user to bypass SSO-only login requirements via userID-based authentication. Mattermost Advisory ID: MMSA-2025-00548

Mattermost versions 10.11.x <= 10.11.9 fail to enforce invite permissions when updating team settings, which allows team administrators without proper permissions to bypass restrictions and add users to their team via API requests. Mattermost Advisory ID: MMSA-2025-00561

Mattermost versions 10.11.x <= 10.11.9 fail to enforce invite permissions when updating team settings, which allows team administrators without proper permissions to bypass restrictions and add users to their team via API requests. Mattermost Advisory ID: MMSA-2025-00561

Versions of the package directorytree/imapengine before 1.22.3 are vulnerable to Improper Neutralization of Special Elements in Output Used by a Downstream Component ('Injection') via the id() function in ImapConnection.php due to improperly escaping user input before including it in IMAP ID commands. This allows attackers to read or delete victim's emails, terminate the victim's session or execute any valid IMAP command on victim's mailbox by including quote characters " or …

A flaw was found in Wildfly Elytron integration. The component does not implement sufficient measures to prevent multiple failed authentication attempts within a short time frame, making it more susceptible to brute force attacks via CLI.

The below gist hangs while attempting to format a long list of tuples. This was found while drafting a regression test for Dja ngo 5.2's composite primary key feature, which allows querying composite fields with tuples.

For some messages, rPGP returned incorrectly decrypted data without signaling that integrity protection was invalid.

It was possible to trigger an unhandled edge case in the Rust Crypto rsa crate through rPGP packet parsing functionality, and crash the process that runs rPGP. This problem has been patched in a new rsa version. The new release of rPGP ensures a patched version of the rsa crate is in use, which prevents this issue.

Previous rPGP versions could be caused to crash with a "stack overflow" when parsing messages that contain deeply nested message layers, such as messages with many signatures. rPGP 0.19.0 resolves this issue with a more robust message handling implementation (via https://github.com/rpgp/rpgp/pull/625).

Mattermost versions 11.1.x <= 11.1.2, 10.11.x <= 10.11.9, 11.2.x <= 11.2.1 fail to validate user permissions when creating Jira issues from Mattermost posts, which allows an authenticated attacker with access to the Jira plugin to read post content and attachments from channels they do not have access to via the /create-issue API endpoint by providing the post ID of an inaccessible post.. Mattermost Advisory ID: MMSA-2025-00550

Mattermost versions 10.11.x <= 10.11.9 fail to properly validate channel membership at the time of data retrieval which allows a deactivated user to learn team names they should not have access to via a race condition in the /common_teams API endpoint.. Mattermost Advisory ID: MMSA-2025-00549

Two path traversal vulnerabilities in the local block adapter allow authenticated users to read and write files outside their designated storage boundaries.

A Critical Broken Authentication vulnerability exists in Known 1.6.2. The application leaks the password reset token within a hidden HTML input field on the password reset page. This allows any unauthenticated attacker to retrieve the reset token for any user by simply querying the user's email, leading to full Account Takeover (ATO) without requiring access to the victim's email inbox.

Summary A Reflected Cross-Site Scripting (XSS) vulnerability was discovered in the AI Playground's OAuth callback handler. The error_description query parameter was directly interpolated into an HTML script tag without proper escaping, allowing attackers to execute arbitrary JavaScript in the context of the victim's session. Root cause The OAuth callback handler in site/ai-playground/src/server.ts directly interpolated the authError value, sourced from the error_description query parameter, into an inline <script> tag. Impact An …

A Reflected Cross-Site Scripting (XSS) vulnerability was discovered in the AI Playground's OAuth callback handler. The error_description query parameter was directly interpolated into an HTML script tag without proper escaping, allowing attackers to execute arbitrary JavaScript in the context of the victim's session.

When Renovate spawns child processes, their access to environment variables is filtered to an allowlist, to prevent unauthorized access to privileged credentials that the Renovate process has access to. Since 42.68.1 (2025-12-30), this filtering had been inadvertently removed, and so any child processes spawned from these versions will have had access to any environment variables that Renovate has access to. This could lead to insider attackers and outside attackers being …

We publish a GitHub security advisory for any releases whose CHANGELOG includes bug-fixes, and encourage our users to upgrade. The latest releases of the hpke-rs and hpke-rs-rust-crypto crates contain the following bug-fixes:

We publish a GitHub security advisory for any releases whose CHANGELOG includes bug-fixes, and encourage our users to upgrade. The latest releases of the hpke-rs and hpke-rs-rust-crypto crates contain the following bug-fixes:

beautiful-mermaid versions prior to 0.1.3 contain an SVG attribute injection issue that can lead to cross-site scripting (XSS) when rendering attacker-controlled Mermaid diagrams. User-controlled values from Mermaid style and classDef directives are interpolated into SVG attribute values without proper escaping, allowing crafted input to break out of an attribute context and inject arbitrary SVG elements/attributes into the rendered output. When the generated SVG is embedded in a web page, this …

Improper Control of Generation of Code ('Code Injection') vulnerability in Apache Avro Java SDK when generating specific records from untrusted Avro schemas. This issue affects Apache Avro Java SDK: all versions through 1.11.4 and version 1.12.0. Users are recommended to upgrade to version 1.12.1 or 1.11.5, which fix the issue.

Improper Control of Generation of Code ('Code Injection') vulnerability in Apache Avro Java SDK when generating specific records from untrusted Avro schemas. This issue affects Apache Avro Java SDK: all versions through 1.11.4 and version 1.12.0. Users are recommended to upgrade to version 1.12.1 or 1.11.5, which fix the issue.

It appeared to be typosquatting existing crate polymarket-client-sdk (sdks vs sdk) and attempting to steal credentials from local files. The malicious crate had 1 version published on 2026-02-09 and had been downloaded only 33 times. There were no crates depending on this crate on crates.io. Thanks to Roland Peelen for finding and reporting this to the crates.io team!

It's possible using comments to inject CSS that would transform the full wiki in a link area leading to a malicious page. All versions of XWiki are impacted by this kind of attack.

An attacker can cause excessive memory consumption in webtransport-go's session implementation by sending a WT_CLOSE_SESSION capsule containing an excessively large Application Error Message. The implementation does not enforce the draft-mandated limit of 1024 bytes on this field, allowing a peer to send an arbitrarily large message payload that is fully read and stored in memory. This allows an attacker to consume an arbitrary amount of memory. The attacker must transmit …

An attacker can cause unbounded memory consumption repeatedly creating and closing many WebTransport streams. Closed streams were not removed from an internal session map, preventing garbage collection of their resources.

An attacker can cause a denial of service in webtransport-go by preventing or indefinitely delaying WebTransport session closure. A malicious peer can withhold QUIC flow control credit on the CONNECT stream, blocking transmission of the WT_CLOSE_SESSION capsule and causing the close operation to hang.

Yoke ATC implements multiple Admission Webhook endpoints (/validations/{airway}, /validations/resources, /validations/flights.yoke.cd, /validations/airways.yoke.cd, etc.) that process AdmissionReview requests. These endpoints do not implement TLS client certificate authentication or request source validation. Any client that can reach the ATC service within the cluster can send requests directly to these endpoints, bypassing the Kubernetes API Server's authentication and authorization mechanisms.

A remote, unauthenticated client can bypass Traefik entrypoint respondingTimeouts.readTimeout by sending the 8-byte Postgres SSLRequest (STARTTLS) prelude and then stalling, causing connections to remain open indefinitely and enabling file-descriptor and goroutine exhaustion denial of service. This triggers during protocol detection before routing, so it is reachable on an entrypoint even when no Postgres/TCP routers are configured (the PoC uses only an HTTP router).

In SurrealDB instances with the scripting capability enabled (–allow-scripting), users with the ability to run arbitrary queries can trigger a server crash due to a memory-safety bug in the underlying JS engine. The SurrealDB instance terminates instantly, requiring a manual restart. The query consists of using built-in string functions to construct a large string and passing it to the JavaScript runtime for compilation. The exact string size required to trigger …

The arrayLimit option in qs does not enforce limits for comma-separated values when comma: true is enabled, allowing attackers to cause denial-of-service via memory exhaustion. This is a bypass of the array limit enforcement, similar to the bracket notation bypass addressed in GHSA-6rw7-vpxm-498p (CVE-2025-15284).

The serialize function used to compile MDX in next-mdx-remote is vulnerable to arbitrary code execution due to insufficient sanitization of MDX content.

A vulnerability has been identified in the NeuVector scanner where the scanner process accepts registry and controller credentials as command-line arguments, potentially exposing sensitive credentials to local users. This may allow unauthorized access to registries or the NeuVector controller, potentially enabling image manipulation, information disclosure, or further lateral movement within the environment. Important: For the exposure of credentials not related to Rancher NeuVector, the final impact severity for confidentiality, integrity …

Versions of the package markdown-it from 13.0.0 and before 14.1.1 are vulnerable to Regular Expression Denial of Service (ReDoS) due to the use of the regex /*+$/ in the linkify function. An attacker can supply a long sequence of * characters followed by a non-matching character, which triggers excessive backtracking and may lead to a denial-of-service condition.

MagicLink stores serialized action objects in the magic_links.action database column and deserializes them without integrity validation or class allowlisting in src/MagicLink.php and src/Actions/ResponseAction.php. An attacker with the ability to manipulate database records (e.g., via SQL injection or compromised admin access) could inject malicious serialized objects containing arbitrary closures, leading to Remote Code Execution (RCE) when the magic link is visited.

The html.Parse function in golang.org/x/net/html has quadratic parsing complexity when processing certain inputs, which can lead to Denial of Service (DoS) if an attacker provides specially crafted HTML content.

FrankenPHP’s CGI path splitting logic improperly handles Unicode characters during case conversion. The logic computes the split index (for finding .php) on a lowercased copy of the request path but applies that byte index to the original path. Because strings.ToLower() in Go can increase the byte length of certain UTF-8 characters (e.g., Ⱥ expands when lowercased), the computed index may not align with the correct position in the original string. …

When running FrankenPHP in worker mode, the $_SESSION superglobal is not correctly reset between requests. This allows a subsequent request processed by the same worker to access the $_SESSION data of the previous request (potentially belonging to a different user) before session_start() is called.

Duplicate Advisory This advisory has been withdrawn because it is a duplicate of GHSA-3m4q-jmj6-r34q. This link is maintained to preserve external references. Original Description Arbitrary file read in the model loading mechanism (HDF5 integration) in Keras versions 3.0.0 through 3.13.1 on all supported platforms allows a remote attacker to read local files and disclose sensitive information via a crafted .keras model file utilizing HDF5 external dataset references.

A timing-based user enumeration vulnerability exists in the password reset functionality. When an invalid reset_url parameter is provided, the response time differs by approximately 500ms between existing and non-existing users, enabling reliable user enumeration.

A timing-based user enumeration vulnerability exists in the password reset functionality. When an invalid reset_url parameter is provided, the response time differs by approximately 500ms between existing and non-existing users, enabling reliable user enumeration.

A vulnerability in CediPay allows attackers to bypass input validation in the transaction API. Affected users: All deployments running versions prior to the patched release. Risk: Exploitation could result in unauthorized transactions, exposure of sensitive financial data, and compromise of payment integrity. Severity: High — potential financial loss and reputational damage. Patches The issue has been fixed in version 1.2.3. Users should upgrade to 1.2.3 or later immediately. All versions …

In accordance with our security policy for libcrux, we publish a GitHub security advisory for any releases whose CHANGELOG includes bug-fixes, and encourage our users to upgrade. The latest releases of the libcrux-ecdh, libcrux-ed25519 and libcrux-psq crates contain the following bug-fixes:

In accordance with our security policy for libcrux, we publish a GitHub security advisory for any releases whose CHANGELOG includes bug-fixes, and encourage our users to upgrade. The latest releases of the libcrux-ecdh, libcrux-ed25519 and libcrux-psq crates contain the following bug-fixes:

In accordance with our security policy for libcrux, we publish a GitHub security advisory for any releases whose CHANGELOG includes bug-fixes, and encourage our users to upgrade. The latest releases of the libcrux-ecdh, libcrux-ed25519 and libcrux-psq crates contain the following bug-fixes:

Yoke ATC allows users to override the Flight WASM module URL via the overrides.yoke.cd/flight annotation on Custom Resources. The controller only checks if the user has update permission on airways resources but does not validate the WASM URL source. An attacker with CR create/update permissions can inject a malicious WASM URL, causing the ATC controller to download and execute arbitrary code.

This crate was used as a dependency by finch_cli_rust and finch-rst and contained a malware payload to exfiltrate credentials. The malicious crate had 1 version published on 2025-12-08 and had been downloaded 22 times. Other than the other crates above that were part of the attack, no other crates depedended on this crate. Thanks to Matthias Zepper of NGI Sweden for reporting this to the crates.io team!

This attempts to typosquat the existing crate finch to steal credentials from local files. The malicious crate had 1 version published on 2025-12-08 and had been downloaded 21 times. There were no crates depending on this crate on crates.io. Thanks to Matthias Zepper of NGI Sweden for reporting this to the crates.io team!

This attempts to typosquat the existing crate finch_cli to steal credentials from local files. The malicious crate had 1 version published on 2025-12-08 and had been downloaded 18 times. There were no crates depending on this crate on crates.io. Thanks to Matthias Zepper of NGI Sweden for reporting this to the crates.io team!

npm @farmfe/core before 1.7.6 is Missing Origin Validation in WebSocket. The development (hot module reloading) server does not validate origin when connecting to a WebSocket client. This allows attackers to surveil developers running Farm who visit their webpage and steal source code that is leaked by the WebSocket server.

The task preview component creates a unparented div. The div's innerHtml is set to the unescaped description of the task

Users without permission to view assets are able are able to download them and view their metadata. Logged-out users and users without permission to access the control panel are unable to take advantage of this.

Stored XSS vulnerability in content titles allow authenticated users with content creation permissions to inject malicious JavaScript that executes when viewed by higher-privileged users. Malicious user must have an account with control panel access and content creation permissions. This vulnerability can be exploited to allow super admin accounts to be created.

A prototype pollution vulnerability exists in the the npm package set-in (>=2.0.1). Despite a previous fix that attempted to mitigate prototype pollution by checking whether user input contained a forbidden key, it is still possible to pollute Object.prototype via a crafted input using Array.prototype. This has been fixed in version 2.0.5.

A prototype pollution vulnerability exists in the the npm package set-in (>=2.0.1). Despite a previous fix that attempted to mitigate prototype pollution by checking whether user input contained a forbidden key, it is still possible to pollute Object.prototype via a crafted input using Array.prototype. This has been fixed in version 2.0.5.

Pion DTLS versions v1.0.0 through v3.1.0 use random nonce generation with AES GCM ciphers, which makes it easier for remote attackers to obtain the authentication key and spoof data by leveraging the reuse of a nonce in a session and a "forbidden attack".

Pion DTLS versions v1.0.0 through v3.0.10 use random nonce generation with AES GCM ciphers, which makes it easier for remote attackers to obtain the authentication key and spoof data by leveraging the reuse of a nonce in a session and a "forbidden attack".

Pion DTLS versions v1.0.0 through v3.1.0 use random nonce generation with AES GCM ciphers, which makes it easier for remote attackers to obtain the authentication key and spoof data by leveraging the reuse of a nonce in a session and a "forbidden attack".

Pion DTLS versions v1.0.0 through v3.0.10 use random nonce generation with AES GCM ciphers, which makes it easier for remote attackers to obtain the authentication key and spoof data by leveraging the reuse of a nonce in a session and a "forbidden attack".

Pion DTLS versions v1.0.0 through v3.1.0 use random nonce generation with AES GCM ciphers, which makes it easier for remote attackers to obtain the authentication key and spoof data by leveraging the reuse of a nonce in a session and a "forbidden attack".

Pion DTLS versions v1.0.0 through v3.0.10 use random nonce generation with AES GCM ciphers, which makes it easier for remote attackers to obtain the authentication key and spoof data by leveraging the reuse of a nonce in a session and a "forbidden attack".

An out-of-bounds write may be triggered when loading a specially crafted PSD image. Pillow >= 10.3.0 users are affected.

Phraseanet 4.0.3 contains a stored cross-site scripting vulnerability that allows authenticated users to inject malicious scripts through crafted file names during document uploads. Attackers can upload files with embedded SVG scripts that execute in the browser, potentially stealing cookies or redirecting users when the file is viewed.

nanotar through 0.2.0 has a path traversal vulnerability in parseTar() and parseTarGzip() that allows remote attackers to write arbitrary files outside the intended extraction directory via a crafted tar archive containing path traversal sequence.

Milvus exposes TCP port 9091 by default with two critical authentication bypass vulnerabilities: The /expr debug endpoint uses a weak, predictable default authentication token derived from etcd.rootPath (default: by-dev), enabling arbitrary expression evaluation. The full REST API (/api/v1/*) is registered on the metrics/management port without any authentication, allowing unauthenticated access to all business operations including data manipulation and credential management.

Milvus exposes TCP port 9091 by default with two critical authentication bypass vulnerabilities: The /expr debug endpoint uses a weak, predictable default authentication token derived from etcd.rootPath (default: by-dev), enabling arbitrary expression evaluation. The full REST API (/api/v1/*) is registered on the metrics/management port without any authentication, allowing unauthenticated access to all business operations including data manipulation and credential management.

Calls issued by the UI against /api/v1/ingestionPipelines leak JWTs used by ingestion-bot for certain services (Glue / Redshift / Postgres)

The ChatOpenAI.get_num_tokens_from_messages() method fetches arbitrary image_url values without validation when computing token counts for vision-enabled models. This allows attackers to trigger Server-Side Request Forgery (SSRF) attacks by providing malicious image URLs in user input.

Kimai 2 contains a persistent cross-site scripting vulnerability that allows attackers to inject malicious scripts into timesheet descriptions. Attackers can insert SVG-based XSS payloads in the description field to execute arbitrary JavaScript when the page is loaded and viewed by other users.

DiskCache (python-diskcache) through 5.6.3 uses Python pickle for serialization by default. An attacker with write access to the cache directory can achieve arbitrary code execution when a victim application reads from the cache.

ajv (Another JSON Schema Validator) through version 8.17.1 is vulnerable to Regular Expression Denial of Service (ReDoS) when the $data option is enabled. The pattern keyword accepts runtime data via JSON Pointer syntax ($data reference), which is passed directly to the JavaScript RegExp() constructor without validation. An attacker can inject a malicious regex pattern (e.g., &#34;^(a|a)*$&#34;) combined with crafted input to cause catastrophic backtracking. A 31-character payload causes approximately 44 …

The RecursiveUrlLoader class in @langchain/community is a web crawler that recursively follows links from a starting URL. Its preventOutside option (enabled by default) is intended to restrict crawling to the same site as the base URL. The implementation used String.startsWith() to compare URLs, which does not perform semantic URL validation. An attacker who controls content on a crawled page could include links to domains that share a string prefix with …

The RecursiveUrlLoader class in @langchain/community is a web crawler that recursively follows links from a starting URL. Its preventOutside option (enabled by default) is intended to restrict crawling to the same site as the base URL. The implementation used String.startsWith() to compare URLs, which does not perform semantic URL validation. An attacker who controls content on a crawled page could include links to domains that share a string prefix with …

The sign-package command in @rage-against-the-pixel/unity-cli logs sensitive credentials in plaintext when the –verbose flag is used. Command-line arguments including –email and –password are output via JSON.stringify without sanitization, exposing secrets to shell history, CI/CD logs, and log aggregation systems. Users who run sign-package with –verbose and credential arguments expose their Unity account passwords. This affects all versions prior to 1.8.2. The vulnerability requires explicit user action (using –verbose) but creates …

Microsoft is releasing this security advisory to provide information about a vulnerability in .NET 8.0, .NET 9.0, and .NET 10.0. This advisory also provides guidance on what developers can do to update their applications to remove this vulnerability. An attacker could exploit this vulnerability by crafting a malicious payload that bypasses the security checks in the affected System.Security.Cryptography.Cose versions, potentially leading to unauthorized access or data manipulation.

A flaw was found in Keycloak. When the logging format is configured to a verbose, user-supplied pattern (such as the pre-defined 'long' pattern), sensitive headers including Authorization and Cookie are disclosed to the logs in cleartext. An attacker with read access to the log files can extract these credentials (e.g., bearer tokens, session cookies) and use them to impersonate users, leading to a full account compromise. Patches are available, see: …

A vulnerability was discovered in go-git whereby data integrity values for .pack and .idx files were not properly verified. This resulted in go-git potentially consuming corrupted files, which would likely result in unexpected errors such as object not found. For context, clients fetch packfiles from upstream Git servers. Those files contain a checksum of their contents, so that clients can perform integrity checks before consuming it. The pack indexes (.idx) …

Description An authentication bypass vulnerability in FUXA allows an unauthenticated, remote attacker to execute arbitrary code on the server when the Node-RED plugin is enabled. This affects FUXA version 1.2.8 through version 1.2.10. This has been patched in FUXA version 1.2.11.

An authorization bypass vulnerability in the FUXA allows an unauthenticated, remote attacker to create and modify arbitrary schedulers, exposing connected ICS/SCADA environments to follow-on actions. This vulnerability affects FUXA version 1.2.8 through version 1.2.10. This has been patched in FUXA version 1.2.11.

A flaw in the path sanitization logic allows an authenticated attacker with administrative privileges to bypass directory traversal protections. By using nested traversal sequences (e.g., ….//), an attacker can write arbitrary files to the server filesystem, including sensitive directories like runtime/scripts. This leads to Remote Code Execution (RCE) when the server reloads the malicious scripts. It is a new vulnerability a patch bypass for the sanitization in the last release …

Unauthenticated access to Adminer UI

A case-sensitivity flaw in the password validation logic allows any authenticated user to change their password (or an admin to change any user's password) without providing the current password. By using Title Case field name "Password" instead of lowercase "password" in the API request, the current_password verification is completely bypassed. This enables account takeover if an attacker obtains a valid JWT token through XSS, session hijacking, or other means. CVSS …

An authenticated user can bypass the application's "Disallow" file path rules by modifying the request URL. By adding multiple slashes (e.g., //private/) to the path, the authorization check fails to match the rule, while the underlying filesystem resolves the path correctly, granting unauthorized access to restricted files.

The cookies property in emmett_core.http.wrappers.Request does not handle CookieError exceptions when parsing malformed Cookie headers. This allows unauthenticated attackers to trigger HTTP 500 errors and cause denial of service.

Impact It is possible to make a specially crafted request with a valid API token that leads to privilege escalation. Affected Versions: ≥= 0.27.19 Mitigation: Upgrade to a patched version: 1.5.13 and later (regular release) 1.4.2 (active LTS release) 1.0.14 (end-of-life LTS release) References The issue was reported by our Core engineer, Dmitrii Patsura (@ovr), in our internal Slack and was promptly patched in a recent update.

Impact It is possible to make the entire Cube API unavailable by submitting a specially crafted request to a Cube API endpoint. Affected Versions: >= 1.1.17 Mitigation: Upgrade to a patched version: 1.5.13 and later (regular release) 1.4.2 (active LTS release) References The issue was reported by our Core engineer, Dmitrii Patsura (@ovr), in our internal Slack and was promptly patched in a recent update.

The public_key_from_numbers (or EllipticCurvePublicNumbers.public_key()), EllipticCurvePublicNumbers.public_key(), load_der_public_key() and load_pem_public_key() functions do not verify that the point belongs to the expected prime-order subgroup of the curve. This missing validation allows an attacker to provide a public key point P from a small-order subgroup. This can lead to security issues in various situations, such as the most commonly used signature verification (ECDSA) and shared key negotiation (ECDH). When the victim computes the shared …

CASL Ability, versions 2.4.0 through 6.7.4, contains a prototype pollution vulnerability.

The cap-go/capacitor-native-biometric library was found to be subject to an authentication bypass as the current implementation of the onAuthenticationSucceeded() does not appear to handle a CryptoObject[^HackTricks1] [^SecuringBiometricAuthentication] as seen in the following code block starting from line 88 in AuthActivity.java: @Override public void onAuthenticationSucceeded( @NonNull BiometricPrompt.AuthenticationResult result ) { super.onAuthenticationSucceeded(result); finishActivity("success"); } As the current implementation only checks whether onAuthenticationSucceeded() was called and does not handle a CryptoObject the biometric …

Unsafe handling of request parameters in the RPC HTTP server results in command injection

Deserialization of untrusted data in Azure SDK allows an unauthorized attacker to execute code over a network.

Observable Timing Discrepancy vulnerability in Apache Shiro. This issue affects Apache Shiro: from 1., 2. before 2.0.7. Users are recommended to upgrade to version 2.0.7 or later, which fixes the issue. Prior to Shiro 2.0.7, code paths for non-existent vs. existing users are different enough, that a brute-force attack may be able to tell, by timing the requests only, determine if the request failed because of a non-existent user vs. …

Affected Products and Versions Apache Druid Affected Versions: 0.17.0 through 35.x (all versions prior to 36.0.0) Prerequisites: * druid-basic-security extension enabled LDAP authenticator configured Underlying LDAP server permits anonymous bind Vulnerability Description An authentication bypass vulnerability exists in Apache Druid when using the druid-basic-security extension with LDAP authentication. If the underlying LDAP server is configured to allow anonymous binds, an attacker can bypass authentication by providing an existing username with …

Versions of amphp/http-server prior to 3.4.4 for the 3.x release branch and prior to 2.1.10 for the 2.x release branch are vulnerable to the HTTP/2 "MadeYouReset" DoS attack described by CVE-2025-8671 and https://kb.cert.org/vuls/id/767506. In versions 3.4.4 and 2.1.10, stream reset protection has been refactored to account for the number of reset streams within a sliding time window. Note that your application must expose HTTP/2 connections directly to be affected by …

Adminer v5.4.1 has a version check mechanism where adminer.org sends signed version info via JavaScript postMessage, which the browser then POSTs to ?script=version. This endpoint lacks origin validation and accepts POST data from any source. An attacker can POST version[] parameter which PHP converts to an array. On next page load, openssl_verify() receives this array instead of string and throws TypeError, returning HTTP 500 to all users.

A sandbox escape vulnerability allows sandboxed code to mutate host built-in prototypes by laundering the isGlobal protection flag through array literal intermediaries. When a global prototype reference (e.g., Map.prototype, Set.prototype) is placed into an array and retrieved, the isGlobal taint is stripped, permitting direct prototype mutation from within the sandbox. This results in persistent host-side prototype pollution and may enable RCE in applications that use polluted properties in sensitive sinks …

Critical Sandbox Escape & Server Takeover: A critical security vulnerability exists in mcp-run-python due to a lack of isolation between the Python runtime (Pyodide) and the host JavaScript environment. The runPython and runPythonAsync functions execute Python code using Pyodide without restricting access to the JavaScript bridge. This allows any executed Python code—whether from a user or an AI model—to access the js module in Pyodide. Through this bridge, the Python …

Server-Side Request Forgery (SSRF): A security vulnerability exists in the mcp-run-python tool (specifically within the Pydantic-AI integration) due to an overly permissive Deno sandbox configuration. The tool configures the Deno runtime—which is intended to isolate the execution of untrusted Python code—with network permissions that include access to the host's loopback interface (localhost). Consequently, malicious Python code executed through the tool can bypass network isolation and send HTTP requests to internal …

FileStore maps cache keys to filenames using Unicode NFKD normalization and ord() substitution without separators, creating key collisions. When FileStore is used as response-cache backend, an unauthenticated remote attacker can trigger cache key collisions via crafted paths, causing one URL to serve cached responses of another (cache poisoning/mixup)

CORS origin validation can be bypassed because the allowed-origins allowlist is compiled into a regex without escaping metacharacters (notably .). An allowed origin like https://good.example can match https://goodXexample, resulting in Access-Control-Allow-Origin being set for an untrusted origin

AllowedHosts host validation can be bypassed because configured host patterns are turned into regular expressions without escaping regex metacharacters (notably .). A configured allowlist entry like example.com can match exampleXcom

The LangSmith SDK's distributed tracing feature is vulnerable to Server-Side Request Forgery via malicious HTTP headers. An attacker can inject arbitrary api_url values through the baggage header, causing the SDK to exfiltrate sensitive trace data to attacker-controlled endpoints.

The LangSmith SDK's distributed tracing feature is vulnerable to Server-Side Request Forgery via malicious HTTP headers. An attacker can inject arbitrary api_url values through the baggage header, causing the SDK to exfiltrate sensitive trace data to attacker-controlled endpoints.

A flaw was found in Keycloak. A vulnerability exists in the jwt-authorization-grant flow where the server fails to verify if an Identity Provider (IdP) is enabled before issuing tokens. The issuer lookup mechanism (lookupIdentityProviderFromIssuer) retrieves the IdP configuration but does not filter for isEnabled=false. If an administrator disables an IdP (e.g., due to a compromise or offboarding), an entity possessing that IdP's signing key can still generate valid JWT assertions …

A flaw was found in Keycloak. An attacker can exploit this vulnerability by modifying the organization ID and target email within a legitimate invitation token's JSON Web Token (JWT) payload. This lack of cryptographic signature verification allows the attacker to successfully self-register into an unauthorized organization, leading to unauthorized access.

A flaw was found in Keycloak. A significant Broken Access Control vulnerability exists in the UserManagedPermissionService (UMA Protection API). When updating or deleting a UMA policy associated with multiple resources, the authorization check only verifies the caller's ownership against the first resource in the policy's list. This allows a user (Owner A) who owns one resource (RA) to update a shared policy and modify authorization rules for other resources (e.g., …

Arbitrary Code Injection (Remote Code Execution & XSS): A critical security vulnerability affects all versions of the jsonpath package. The library relies on the static-eval module to evaluate JSON Path expressions but fails to properly sanitize or sandbox the input. This allows an attacker to inject arbitrary JavaScript code into the JSON Path expression. When the library evaluates this expression, the malicious code is executed. Node.js Environments: This leads to …

Fiber v2 contains an internal vendored copy of gofiber/utils, and its functions UUIDv4() and UUID() inherit the same critical weakness described in the upstream advisory. On Go versions prior to 1.24, the underlying crypto/rand implementation can return an error if secure randomness cannot be obtained. In such cases, these Fiber v2 UUID functions silently fall back to generating predictable values — the all-zero UUID 00000000-0000-0000-0000-000000000000. On Go 1.24+, the language …

Faraday's build_exclusive_url method (in lib/faraday/connection.rb) uses Ruby's URI#merge to combine the connection's base URL with a user-supplied path. Per RFC 3986, protocol-relative URLs (e.g. //evil.com/path) are treated as network-path references that override the base URL's host/authority component. This means that if any application passes user-controlled input to Faraday's get(), post(), build_url(), or other request methods, an attacker can supply a protocol-relative URL like //attacker.com/endpoint to redirect the request to an …

The save_images_Asset graphql mutation allows a user to give a url of an image to download. (Url must use a domain, not a raw IP.) Attacker sets up domain attacker.domain with an A record of something like 169.254.169.254 (special AWS metadata IP) Attacker invokes save_images_Asset with url: http://attacker.domain/latest/meta-data/iam/security-credentials and filename "foo.txt" Craft fetches sensitive information on attacker's behalf, and makes it available for download at /assets/images/foo.txt Normal checks to verify …

Type: Privilege Escalation (CWE-269) Affected: Craft CMS 5.x (likely affects 4.x and 3.x as well) Location: src/gql/resolvers/mutations/Asset.php lines 57-107

A stored XSS vulnerability exists in the Number field type settings. The Prefix and Suffix fields are rendered using the |md|raw Twig filter without proper escaping, allowing script execution when the Number field is displayed on users' profiles.

Stored XSS via Entry Type names. The name is not sanitized when displayed in the Entry Types list.

The saveAsset GraphQL mutation validates the initial URL hostname and resolved IP against a blocklist, but Guzzle follows HTTP redirects by default. An attacker can bypass all SSRF protections by hosting a redirect that points to cloud metadata endpoints or any internal IP addresses.

The saveAsset GraphQL mutation uses filter_var(…, FILTER_VALIDATE_IP) to block a specific list of IP addresses. However, alternative IP notations (hexadecimal, mixed) are not recognized by this function, allowing attackers to bypass the blocklist and access cloud metadata services.

The element-indexes/get-elements endpoint is vulnerable to SQL Injection via the criteria[orderBy] parameter (JSON body). The application fails to sanitize this input before using it in the database query. An attacker with Control Panel access can inject arbitrary SQL into the ORDER BY clause by omitting viewState[order] (or setting both to the same payload). [!NOTE] The ORDER BY clause executes per row. SLEEP(1) on 10 rows = 10s delay.

A Remote Code Execution (RCE) vulnerability exists in Craft CMS where the assembleLayoutFromPost() function in src/services/Fields.php fails to sanitize user-supplied configuration data before passing it to Craft::createObject(). This allows authenticated administrators to inject malicious Yii2 behavior configurations that execute arbitrary system commands on the server. This vulnerability represents an unpatched variant of the behavior injection vulnerability addressed in GHSA-255j-qw47-wjh5, affecting different endpoints through a separate code path.

The mergeConfig function in axios crashes with a TypeError when processing configuration objects containing proto as an own property. An attacker can trigger this by providing a malicious configuration object created via JSON.parse(), causing complete denial of service.

Authentication Bypass: A vulnerability exists in Apache Shiro that allows authentication bypass for static files when served from a case-insensitive filesystem (such as the default configuration on macOS or Windows). The issue arises when Shiro's URL filters are configured with lower-case rules (a common default), but the underlying operating system treats mixed-case filenames as identical. An attacker can access protected static resources by varying the capitalization of the filename in …

Exposure of Sensitive Information: An information disclosure vulnerability exists in the Apache Airflow UI that allows authenticated users to view Import Errors for DAGs they are not authorized to access. In affected versions, the Import Errors view does not correctly filter errors based on granular DAG permissions. This means a user with access to only DAG_A can view import errors generated by DAG_B, DAG_C, or system-level DAGs. These error logs …

Confidentiality Loss: Task logs often contain sensitive operational data, debugging information, or potentially leaked secrets (environment variables, connection strings) that should not be visible to all users with basic task access. Broken Access Control: This bypasses the intended security model for restricted user roles.

A vulnerability was found in r-huijts xcode-mcp-server up to f3419f00117aa9949e326f78cc940166c88f18cb. This affects the function registerXcodeTools of the file src/tools/xcode/index.ts of the component run_lldb. The manipulation of the argument args results in command injection. It is possible to launch the attack remotely. The exploit has been made public and could be used. This product takes the approach of rolling releases to provide continious delivery. Therefore, version details for affected and updated …

A vulnerability was determined in BurtTheCoder mcp-maigret up to 1.0.12. This affects an unknown part of the file src/index.ts of the component search_username. Executing a manipulation of the argument Username can lead to command injection. The attack may be launched remotely. Upgrading to version 1.0.13 is able to mitigate this issue. This patch is called b1ae073c4b3e789ab8de36dc6ca8111ae9399e7a. Upgrading the affected component is advised.

The DNS C2 listener accepts unauthenticated TOTP bootstrap messages and allocates server-side DNS sessions without validating OTP values, even when EnforceOTP is enabled. Because sessions are stored without a cleanup/expiry path in this flow, an unauthenticated remote actor can repeatedly create sessions and drive memory exhaustion.

What kind of vulnerability is it? Who is impacted? An Arbitrary File Write vulnerability has been identified in Microsoft's Semantic Kernel .NET SDK, specifically within the SessionsPythonPlugin. Developers who have built applications which include Microsoft's Semantic Kernel .NET SDK and are using the SessionsPythonPlugin

What kind of vulnerability is it? Who is impacted? An Arbitrary File Write vulnerability has been identified in Microsoft's Semantic Kernel .NET SDK, specifically within the SessionsPythonPlugin. Developers who have built applications which include Microsoft's Semantic Kernel .NET SDK and are using the SessionsPythonPlugin

If an attacker has the ability control configuration options passed to sceditor.create(), like emoticons, charset, etc. then it's possible for them to trigger an XSS attack due to lack of sanitisation of configuration options. Proof of concept: sceditor.create(textarea, { emoticons: { dropdown: { ':)': { url: 'x" onerror="window.__xss = true' } } } });

A Path Traversal vulnerability in the Pydantic AI web UI allows an attacker to serve arbitrary JavaScript in the context of the application by crafting a malicious URL. If a victim clicks the link or visits it via an iframe, attacker-controlled code executes in their browser, enabling theft of chat history and other client-side data. This vulnerability only affects applications that use: Agent.to_web to serve a chat interface clai web …

A Path Traversal vulnerability in the Pydantic AI web UI allows an attacker to serve arbitrary JavaScript in the context of the application by crafting a malicious URL. If a victim clicks the link or visits it via an iframe, attacker-controlled code executes in their browser, enabling theft of chat history and other client-side data. This vulnerability only affects applications that use: Agent.to_web to serve a chat interface clai web …

A Server-Side Request Forgery (SSRF) vulnerability exists in Pydantic AI's URL download functionality. When applications accept message history from untrusted sources, attackers can include malicious URLs that cause the server to make HTTP requests to internal network resources, potentially accessing internal services or cloud credentials. This vulnerability only affects applications that accept message history from external users, such as those using: Agent.to_web or clai web to serve a chat interface …

A Server-Side Request Forgery (SSRF) vulnerability exists in Pydantic AI's URL download functionality. When applications accept message history from untrusted sources, attackers can include malicious URLs that cause the server to make HTTP requests to internal network resources, potentially accessing internal services or cloud credentials. This vulnerability only affects applications that accept message history from external users, such as those using: Agent.to_web or clai web to serve a chat interface …

During a security audit conducted with Claude Opus 4.6 and GPT-5.3-Codex, we identified three specific ways to bypass the XSS (cross-site-scripting) protection built into Phlex. The first bypass could happen if user-provided attributes with string keys were splatted into HTML tag, e.g. div(**user_attributes). The second bypass could happen if user-provided tag names were passed to the tag method, e.g. tag(some_tag_name_from_user). The third bypass could happen if user’s links were passed …

A critical OS Command Injection vulnerability exists in the P7M (signed XML) file decoding functionality. An authenticated attacker can upload a ZIP file containing a .p7m file with a malicious filename to execute arbitrary system commands on the server.

Critical Time-Based Blind SQL Injection vulnerability affecting multiple search modules in OpenSTAManager v2.9.8 allows authenticated attackers to extract sensitive database contents including password hashes, customer data, and financial records through time-based Boolean inference attacks with amplified execution across 10+ modules. Status: ✅ Confirmed and tested on live instance (v2.9.8) Vulnerable Parameter: term (GET) Affected Endpoint: /ajax_search.php Affected Modules: Articoli, Ordini, DDT, Fatture, Preventivi, Anagrafiche, Impianti, Contratti, Automezzi, Interventi

Critical Time-Based Blind SQL Injection vulnerability in the article pricing module of OpenSTAManager v2.9.8 allows authenticated attackers to extract complete database contents including user credentials, customer data, and financial records through time-based Boolean inference attacks. Status: ✅ Confirmed and tested on live instance (v2.9.8) end demo.osmbusiness.it (v2.9.7) Vulnerable Parameter: idarticolo (GET) Affected Endpoint: /ajax_complete.php?op=getprezzi Affected Module: Articoli (Articles/Products)

Critical Error-Based SQL Injection vulnerability in the Scadenzario (Payment Schedule) bulk operations module of OpenSTAManager v2.9.8 allows authenticated attackers to extract complete database contents including user credentials, customer PII, and financial records through XML error messages. Status: ✅ Confirmed and tested on live instance (v2.9.8) Vulnerable Parameter: id_records[] (POST array) Affected Endpoint: /actions.php?id_module=18 (Scadenzario module) Attack Type: Error-Based SQL Injection (IN clause)

Critical Error-Based SQL Injection vulnerability in the Prima Nota (Journal Entry) module of OpenSTAManager v2.9.8 allows authenticated attackers to extract complete database contents including user credentials, customer PII, and financial records through XML error messages by injecting malicious SQL into URL parameters. Status: ✅ Confirmed and tested on live instance (v2.9.8) Vulnerable Parameters: id_documenti (GET parameters) Affected Endpoint: /modules/primanota/add.php Attack Type: Error-Based SQL Injection (IN clause)

An authenticated SQL Injection vulnerability in OpenSTAManager's Scadenzario (Payment Schedule) print template allows any authenticated user to extract sensitive data from the database, including admin credentials, customer information, and financial records. The vulnerability enables complete database read access through error-based SQL injection techniques.

A SQL Injection vulnerability exists in the ajax_select.php endpoint when handling the componenti operation. An authenticated attacker can inject malicious SQL code through the options[matricola] parameter.

Insufficient escaping of unicode characters in query log in Neo4j Enterprise and Community editions prior to 2026.01 can lead to XSS if the user opens the logs in a tool that treats them as HTML. There is no security impact on Neo4j products, but this advisory is released as a precaution to treat the logs as plain text if using versions prior to 2026.01. Proof of concept exploit: https://github.com/JoakimBulow/CVE-2026-1337

Disclosure of Salesforce OAuth bearer tokens used by the MCP.

Mattermost Confluence plugin version < 1.7.0 fails to properly escape user-controlled display names in HTML template rendering which allows authenticated Confluence users with malicious display names to execute arbitrary JavaScript in victim browsers via sending a specially crafted OAuth2 connection link that, when visited, renders the attacker's display name without proper sanitization. Mattermost Advisory ID: MMSA-2025-00557

A malformed or tampered-with LookupResources Cursor token can cause a panic in the SpiceDB process if it fails to parse. If an attacker were able to make requests to a SpiceDB instance, they could affect its availability.

The Keylime registrar does not enforce mutual TLS (mTLS) client certificate authentication since version 7.12.0. The registrar's TLS context is configured with ssl.CERT_OPTIONAL instead of ssl.CERT_REQUIRED, allowing any client to connect to protected API endpoints without presenting a valid client certificate. Who is impacted: All Keylime deployments running versions 7.12.0 through 7.13.0 Environments where the registrar HTTPS port (default 8891) is network-accessible to untrusted clients What an attacker can do: …

Gophish <= 0.12.1 is vulnerable to Incorrect Access Control. The administrative dashboard exposes each user’s long-lived API key directly inside the rendered HTML/JavaScript of the page on every login. This makes permanent API credentials accessible to any script running in the browser context.

Due to the insufficient patch for the https://github.com/gogs/gogs/security/advisories/GHSA-wj44-9vcg-wjq7, it's still possible to update files in the .git directory and achieve remote command execution.

Stored XSS via mermaid diagrams due to usage of vulnerable renderer library

A Path Traversal vulnerability exists in the updateWikiPage function of Gogs. The vulnerability allows an authenticated user with write access to a repository's wiki to delete arbitrary files on the server by manipulating the old_title parameter in the wiki editing form.

Contact OpenAI Security Research at outbounddisclosures@openai.com to engage on this report. See PDF report for easier reading.

The DELETE /api/v1/repos/:owner/:repo endpoint lacks necessary permission validation middleware. Consequently, any user with read access (including read-only collaborators) can delete the entire repository. This vulnerability stems from the API route configuration only utilizing the repoAssignment() middleware (which only verifies read access) without enforcing reqRepoOwner() or reqRepoAdmin().

Vulnerability Description In the endpoint: /username/reponame/settings/hooks/git/:name the :name parameter: Is URL-decoded by macaron routing, allowing decoded slashes (/) Is then passed directly to: git.Repository.Hook("custom_hooks", name) which internally resolves the path as: filepath.Join(repoPath, "custom_hooks", name) Because no path sanitization is applied, supplying ../ sequences allows access to arbitrary paths outside the repository. As a Result: GET: Arbitrary file contents are displayed in the hook edit page textarea (Local File Inclusion). POST: …

An authenticated user can cause a DOS attack. If one of the repo files is deleted before synchronization, it will cause the application to crash.

Duplicate Advisory This advisory has been withdrawn because it is a duplicate of GHSA-4jqp-9qjv-57m2. This link is maintained to preserve external references. Original Description A flaw was found in Keylime. The Keylime registrar, since version 7.12.0, does not enforce client-side Transport Layer Security (TLS) authentication. This authentication bypass vulnerability allows unauthenticated clients with network access to perform administrative operations, including listing agents, retrieving public Trusted Platform Module (TPM) data, and …

Versions 0.2.1 and 0.3.0 of client-certificate-auth contain an open redirect vulnerability. The middleware unconditionally redirects HTTP requests to HTTPS using the unvalidated Host header, allowing an attacker to redirect users to arbitrary domains.

Claude Code failed to properly validate commands using piped sed operations with the echo command, allowing attackers to bypass file write restrictions. This vulnerability enabled writing to sensitive directories like the .claude folder and paths outside the project scope. Exploiting this required the ability to execute commands through Claude Code with the "accept edits" feature enabled. Users on standard Claude Code auto-update received this fix automatically. Users performing manual updates …

Claude Code failed to properly validate directory changes when combined with write operations to protected folders. By using the cd command to navigate into sensitive directories like .claude, it was possible to bypass write protection and create or modify files without user confirmation. Reliably exploiting this required the ability to add untrusted content into a Claude Code context window. Users on standard Claude Code auto-update received this fix automatically. Users …

Claude Code's bubblewrap sandboxing mechanism failed to properly protect the .claude/settings.json configuration file when it did not exist at startup. While the parent directory was mounted as writable and .claude/settings.local.json was explicitly protected with read-only constraints, settings.json was not protected if it was missing. This allowed malicious code running inside the sandbox to create this file and inject persistent hooks (such as SessionStart commands) that would execute with host privileges …

Claude Code failed to strictly enforce deny rules configured in settings.json when accessing files through symbolic links. If a user explicitly denied Claude Code access to a file (such as /etc/passwd) and Claude Code had access to a symbolic link pointing to that file, it was possible for Claude Code to read the restricted file through the symlink without triggering deny rule enforcement. Users on standard Claude Code auto-update received …

When using P256 certificates (which is not the default configuration), it is possible to evade a blocklist entry created against the fingerprint of a certificate by using ECDSA Signature Malleability to use a copy of the certificate with a different fingerprint. In order for this to affect a user or network, all of the following must be true: CURVE_P256 certificates are being used There are one or more entries on …

When using P256 certificates (which is not the default configuration), it is possible to evade a blocklist entry created against the fingerprint of a certificate by using ECDSA Signature Malleability to use a copy of the certificate with a different fingerprint. In order for this to affect a user or network, all of the following must be true: CURVE_P256 certificates are being used There are one or more entries on …

Antrea's network policy priority assignment system has a uint16 arithmetic overflow bug that causes incorrect OpenFlow priority calculations when handling a large numbers of policies with various priority values. This results in potentially incorrect traffic enforcement. If a user creates a large number of Antrea NetworkPolicies (ANP or ACNP) with varying priorities, some rules with lower logical priorities (higher numerical priority values) may take precedence over rules with higher logical …

A Denial of Service (DoS) vulnerability (CWE-400) exists in the multipart file handling logic of @adonisjs/bodyparser. When processing file uploads, the multipart parser may accumulate an unbounded amount of data in memory while attempting to detect file types, potentially leading to excessive memory consumption and process termination. This issue affects applications that accept multipart/form-data uploads using affected versions of @adonisjs/bodyparser.

A Prototype Pollution vulnerability (CWE-1321) in AdonisJS multipart form-data parsing may allow a remote attacker to manipulate object prototypes at runtime. This impacts @adonisjs/bodyparser through version 10.1.2 and 11.x prerelease versions prior to 11.0.0-next.8. This issue has been patched in @adonisjs/bodyparser versions 10.1.3 and 11.0.0-next.9

When passing a non-existing folder to the actix_files::Files::new() method causes the actix server to expose unexpected files.

A PyPI user account compromised by an attacker and was able to upload a malicious version (1.1.5.post1) of the dydx-v4-client package. This version contains a highly obfuscated multi-stage loader that ultimately executes malicious code on the host system. While the final payload is not visible because it is tucked away inside 100 layers of encoding, the structural design—specifically the use of recursive decompression followed by an exec() call is a …

It depended on the evm-units crate, which appeared to be attempting to steal cryptocurrency.

It appeared to be attempting to steal credentials from local files.

It appeared to be typosquatting existing crate polymarket-client-sdk (clients vs client) and attempting to steal credentials from local files. The malicious crate had 6 versions published on 2026-02-05 and had been downloaded only 59 times. There were no crates depending on this crate on crates.io. Polymarket thanks Socket.dev for detecting and reporting this to the crates.io team!

It depended on the sha-rust crate, which appeared to be attempting to steal credentials from local files.

It appeared to be attempting to steal cryptocurrency.

A GET request for a static file served by actix-files with an empty Range header triggers a panic. With panic = "abort", a remote user may crash the process on-demand.

Vulnerability Description The endpoint PUT /repos/:owner/:repo/contents/* does not require write permissions and allows access with read permission only via repoAssignment(). After passing the permission check, PutContents() invokes UpdateRepoFile(), which results in: Commit creation Execution of git push As a result, a token with read-only permission can be used to modify repository contents. Attack Prerequisites Possession of a valid access token Read permission on the target repository (public repository or collaborator …

When experiments.buildHttp is enabled, webpack’s HTTP(S) resolver (HttpUriPlugin) can be bypassed to fetch resources from hosts outside allowedUris by using crafted URLs that include userinfo (username:password@host). If allowedUris enforcement relies on a raw string prefix check (e.g., uri.startsWith(allowed)), a URL that looks allow-listed can pass validation while the actual network request is sent to a different authority/host after URL parsing. This is a policy/allow-list bypass that enables build-time SSRF behavior …

When experiments.buildHttp is enabled, webpack’s HTTP(S) resolver (HttpUriPlugin) enforces allowedUris only for the initial URL, but does not re-validate allowedUris after following HTTP 30x redirects. As a result, an import that appears restricted to a trusted allow-list can be redirected to HTTP(S) URLs outside the allow-list. This is a policy/allow-list bypass that enables build-time SSRF behavior (requests from the build machine to internal-only endpoints, depending on network access) and untrusted …

web2py versions 2.27.1-stable+timestamp.2023.11.16.08.03.57 and prior contain an Open Redirect vulnerability. If this vulnerability is exploited, the user may be redirected to an arbitrary website when accessing a specially crafted URL. As a result, the user may become a victim of a phishing attack.

This issue may lead to disclosure of PII of guest users (including names, addresses and phone numbers).

This issue may lead to disclosure of PII of guest users (including names, addresses and phone numbers).

A critical IDOR vulnerability exists in Spree Commerce's guest checkout flow that allows any guest user to bind arbitrary guest addresses to their order by manipulating address ID parameters. This enables unauthorized access to other guests' personally identifiable information (PII) including names, addresses and phone numbers. The vulnerability bypasses existing ownership validation checks and affects all guest checkout transactions.

A critical IDOR vulnerability exists in Spree Commerce's guest checkout flow that allows any guest user to bind arbitrary guest addresses to their order by manipulating address ID parameters. This enables unauthorized access to other guests' personally identifiable information (PII) including names, addresses and phone numbers. The vulnerability bypasses existing ownership validation checks and affects all guest checkout transactions.

When user-provided input is provided to any type that parses with the RFC 2822 format, a denial of service attack via stack exhaustion is possible. The attack relies on formally deprecated and rarely-used features that are part of the RFC 2822 format used in a malicious manner. Ordinary, non-malicious input will never encounter this scenario.

When user-provided input is provided to any type that parses with the RFC 2822 format, a Denial of Service attack via stack exhaustion is possible. The attack relies on formally deprecated and rarely-used features that are part of the RFC 2822 format used in a malicious manner. Ordinary, non-malicious input will never encounter this scenario.

A Path Traversal vulnerability in the website content subsystem lets an authenticated operator read arbitrary files on the Sliver server host. This is an authenticated Path Traversal / arbitrary file read issue, and it can expose credentials, configs, and keys.

A Path Traversal vulnerability in the website content subsystem lets an authenticated operator read arbitrary files on the Sliver server host. This is an authenticated Path Traversal / arbitrary file read issue, and it can expose credentials, configs, and keys.

The existing layers of security in enclave-vm are insufficient: The AST sanitization can be bypassed with dynamic property accesses, the hardening of the error objects does not cover the peculiar behavior or the vm module and the function constructor access prevention can be side-stepped by leveraging host object references.

The existing layers of security in enclave-vm are insufficient: The AST sanitization can be bypassed with dynamic property accesses, the hardening of the error objects does not cover the peculiar behavior or the vm module and the function constructor access prevention can be side-stepped by leveraging host object references.

It is possible to append to arbitrary files via /logger endpoint. Minimal privileges are required (read-only access). Tested on Qdrant 1.15.5

pgAdmin versions 9.11 are affected by a Restore restriction bypass via key disclosure vulnerability that occurs when running in server mode and performing restores from PLAIN-format dump files. An attacker with access to the pgAdmin web interface can observe an active restore operation, extract the \restrict key in real time, and race the restore process by overwriting the restore script with a payload that re-enables meta-commands using \unrestrict <key>. This …

A cross-collection Insecure Direct Object Reference (IDOR) vulnerability exists in the payload-preferences internal collection. In multi-auth collection environments using Postgres or SQLite with default serial/auto-increment IDs, authenticated users from one auth collection can read and delete preferences belonging to users in different auth collections when their numeric IDs collide. Users are affected if ALL of these are true: Multiple auth collections configured (e.g., admins + customers) Postgres or SQLite database …

OpenFGA v1.8.5 to v1.11.2 ( openfga-0.2.22 <= Helm chart <= openfga-0.2.51, v.1.8.5 <= docker <= v.1.11.2) are vulnerable to improper policy enforcement when certain Check calls are executed.

A security issue was discovered in Reva based products that enables a malicious user to bypass the scope validation of a public link, allowing it to access resources outside the scope of a public link.

A security issue was discovered in Reva that enables a malicious user to bypass the scope validation of a public link. That allows it to access resources outside the scope of a public link. OpenCloud uses Reva as one of its core components and thus it is affected.

The ui.markdown() component uses the markdown2 library to convert markdown content to HTML, which is then rendered via innerHTML. By default, markdown2 allows raw HTML to pass through unchanged. This means that if an application renders user-controlled content through ui.markdown(), an attacker can inject malicious HTML containing JavaScript event handlers. Unlike other NiceGUI components that render HTML (ui.html(), ui.chat_message(), ui.interactive_image()), the ui.markdown() component does not provide or require a sanitize …

NiceGUI's FileUpload.name property exposes client-supplied filename metadata without sanitization, enabling path traversal when developers use the pattern UPLOAD_DIR / file.name. Malicious filenames containing ../ sequences allow attackers to write files outside intended directories, with potential for remote code execution through application file overwrites in vulnerable deployment patterns. This design creates a prevalent security footgun affecting applications following common community patterns. Note: Exploitation requires application code incorporating file.name into filesystem paths …

Cross-site Scripting vulnerability in the "/admin/order/abandoned" endpoint of Microweber 2.0.19. An attacker can manipulate the "orderDirection" parameter in a crafted URL and lure a user with admin privileges into visiting it, achieving JavaScript code execution in the victim's browser. The issue was reported to the developers and fixed in version 2.0.20.

There is a Cross-site Scripting vulnerability in the "/admin/category/create" endpoint of Microweber 2.0.19. An attacker can manipulate the "rel_id" parameter in a crafted URL and lure a user with admin privileges into visiting it, achieving JavaScript code execution in the victim's browser. The issue was reported to the developers and fixed in version 2.0.20.

An insecure default configuration in FUXA allows an unauthenticated, remote attacker to gain administrative access and execute arbitrary code on the server. This affects FUXA through version 1.2.9 when authentication is enabled, but the administrator JWT secret is not configured. This issue has been patched in FUXA version 1.2.10.

An insecure default configuration in FUXA allows an unauthenticated, remote attacker to gain administrative access and execute arbitrary code on the server. This affects FUXA through version 1.2.9 when authentication is enabled, but the administrator JWT secret is not configured. This issue has been patched in FUXA version 1.2.10.

Description A path traversal vulnerability in FUXA allows an unauthenticated, remote attacker to write arbitrary files to arbitrary locations on the server filesystem. This affects FUXA through version 1.2.9. This issue has been patched in FUXA version 1.2.10.

Description A path traversal vulnerability in FUXA allows an unauthenticated, remote attacker to write arbitrary files to arbitrary locations on the server filesystem. This affects FUXA through version 1.2.9. This issue has been patched in FUXA version 1.2.10.

An authentication bypass vulnerability in FUXA allows an unauthenticated, remote attacker to gain administrative access via the heartbeat refresh API and execute arbitrary code on the server. This affects FUXA through version 1.2.9 when authentication is enabled. This issue has been patched in FUXA version 1.2.10.

An authentication bypass vulnerability in FUXA allows an unauthenticated, remote attacker to gain administrative access via the heartbeat refresh API and execute arbitrary code on the server. This affects FUXA through version 1.2.9 when authentication is enabled. This issue has been patched in FUXA version 1.2.10.

Description An authorization bypass vulnerability in FUXA allows an unauthenticated, remote attacker to modify device tags via WebSockets. This affects FUXA through version 1.2.9. This issue has been patched in FUXA version 1.2.10.

Description An authorization bypass vulnerability in FUXA allows an unauthenticated, remote attacker to modify device tags via WebSockets. This affects FUXA through version 1.2.9. This issue has been patched in FUXA version 1.2.10.

An information disclosure vulnerability in FUXA allows an unauthenticated, remote attacker to retrieve sensitive administrative database credentials. This affects FUXA through version 1.2.9. This issue has been patched in FUXA version 1.2.10.

An information disclosure vulnerability in FUXA allows an unauthenticated, remote attacker to retrieve sensitive administrative database credentials. This affects FUXA through version 1.2.9. This issue has been patched in FUXA version 1.2.10.

Vulnerability in base Docker images (PHP, Go, and Alpine) not automatically propagating to FrankenPHP images. FrankenPHP's container images were previously built only when specific version tags were updated or when manual triggers were initiated. This meant that if an upstream base image (such as Alpine Linux or official PHP/Go images) received a security patch under an existing tag, the FrankenPHP image would remain on the older, vulnerable version of those …

When querying JSON or richText fields, user input was directly embedded into SQL without escaping, enabling blind SQL Injection attacks. An unauthenticated attacker could extract sensitive data (emails, password reset tokens) and achieve full account takeover without password cracking. Users are affected if ALL of these are true: Payload version < v3.73.0 Using a Drizzle-based database adapter (@payloadcms/drizzle as dependency): @payloadcms/db-postgres @payloadcms/db-vercel-postgres @payloadcms/db-sqlite @payloadcms/db-d1-sqlite At least one accessible collection that …

A sandbox escape vulnerabilities due to a mismatch between the key on which the validation is performed and the key used for accessing properties.

A sandbox escape is possible by shadowing hasOwnProperty on a sandbox object, which disables prototype whitelist enforcement in the property-access path. This permits direct access to proto and other blocked prototype properties, enabling host Object.prototype pollution and persistent cross-sandbox impact. The issue was reproducible on Node v23.9.0 using the project’s current build output. The bypass works with default Sandbox configuration and does not require custom globals or whitelists.

As Map is in SAFE_PROTOYPES, it's prototype can be obtained via Map.prototype. By overwriting Map.prototype.has the sandbox can be escaped.

The return values of functions aren't wrapped. Object.values/Object.entries can be used to get an Array containing the host's Function constructor, by using Array.prototype.at you can obtain the hosts Function constructor, which can be used to execute arbitrary code outside of the sandbox.

Affected versions of Winter CMS allow users with access to the CMS Asset Manager were able to upload SVGs without automatic sanitization. To actively exploit this security issue, an attacker would need access to the Backend with a user account with the following permission: cms.manage_assets The Winter CMS maintainers strongly recommend that the cms.manage_assets permission only be reserved to trusted administrators and developers in general.

The following security vulnerability was identified in jsPDF versions <=3.0.4: Local File Inclusion/Path Traversal.

The following security vulnerability was identified in jsPDF versions <=3.0.4: Local File Inclusion/Path Traversal.

Membership and confirmation tags may not be checked correctly due to a missing length check. Any tag that is shorter than the expected tag, but matches up to its length, as well as any empty tag is considered valid.

An unauthenticated local client could use the Gateway WebSocket API to write config via config.apply and set unsafe cliPath values that were later used for command discovery, enabling command injection as the gateway user.

An unauthenticated local client could use the Gateway WebSocket API to write config via config.apply and set unsafe cliPath values that were later used for command discovery, enabling command injection as the gateway user.

The isValidMedia() function in src/media/parse.ts allows arbitrary file paths including absolute paths, home directory paths, and directory traversal sequences. An agent can read any file on the system by outputting MEDIA:/path/to/file, exfiltrating sensitive data to the user/channel.

Neo4j Enterprise and Community editions versions prior to 2026.01.3 and 5.26.21 are vulnerable to a potential information disclosure by a user who has ability to access the local log files. The "obfuscate_literals" option in the query logs does not redact error information, exposing unredacted data in the query log when a customer writes a query that fails. It can allow a user with legitimate access to the local log files …

An XSS vulnerability in the frontend allows a malicious attacker to inject code through the comment metadata of a song to exfiltrate user credentials. An attacker's maliciously crafted song has to be added to Navidrome to exploit the vulnerability.

An XSS vulnerability in the frontend allows a malicious attacker to inject code through the comment metadata of a song to exfiltrate user credentials. An attacker's maliciously crafted song has to be added to Navidrome to exploit the vulnerability.

Authenticated users can crash the Navidrome server by supplying an excessively large size parameter to /rest/getCoverArt or to a shared-image URL (/share/img/<token>). When processing such requests, the server attempts to create an extremely large resized image, causing uncontrolled memory growth. This triggers the Linux OOM killer, terminates the Navidrome process, and results in a full service outage. If the system has sufficient memory and survives the allocation, Navidrome then writes …

Authenticated users can crash the Navidrome server by supplying an excessively large size parameter to /rest/getCoverArt or to a shared-image URL (/share/img/<token>). When processing such requests, the server attempts to create an extremely large resized image, causing uncontrolled memory growth. This triggers the Linux OOM killer, terminates the Navidrome process, and results in a full service outage. If the system has sufficient memory and survives the allocation, Navidrome then writes …

The use of Buffer.allocUnsafe() and Buffer.allocUnsafeSlow() in the task runner allowed untrusted code to allocate uninitialized memory. Such uninitialized buffers could contain residual data from within the same Node.js process (for example, data from prior requests, tasks, secrets, or tokens), resulting in potential information disclosure. Only authenticated users are able to execute code through Task Runners. This issue affected any deployment in which both of the following conditions were met: …

A vulnerability in the file access controls allows authenticated users with permission to create or modify workflows to read sensitive files from the n8n host system. This can be exploited to obtain critical configuration data and user credentials, leading to complete account takeover of any user on the instance.

A Cross-site Scripting (XSS) vulnerability has been identified in the handling of webhook responses and related HTTP endpoints. Under certain conditions, the Content Security Policy (CSP) sandbox protection intended to isolate HTML responses may not be applied correctly. An authenticated user with permission to create or modify workflows could abuse this to execute malicious scripts with same-origin privileges when other users interact with the crafted workflow. This could lead to …

A vulnerability in the HTTP Request node's credential domain validation allowed an authenticated attacker to send requests with credentials to unintended domains, potentially leading to credential exfiltration. This only might affect user who have credentials that use wildcard domain patterns (e.g., *.example.com) in the "Allowed domains" setting.