Vulnerabilities

Rocket.Chat in versions before 8.8.0, 8.7.1, 8.6.2, 8.5.3, 8.4.6, 8.3.8, 8.2.8, 8.1.8, and 7.10.15 has a REST API endpoint POST /api/v1/livechat/visitor that accepts an unauthenticated, unsanitized name field for Livechat visitors. This name is stored raw and later rendered via dangerouslySetInnerHTML in the Omnichannel Queue side panel (InquireSidePanelItem.tsx), injecting a real, clickable HTML link - pointing to any attacker-controlled domain, with arbitrary social-engineering text - into the DOM of any agent viewing the queue.

Rocket.Chat's SAML SSO before versions 8.7.0, 8.6.1, 8.5.2, 8.4.5, 8.3.7, 8.2.7, 8.1.7, 8.0.8, and 7.10.14 verified XML signatures but did not bind the validated signature to samlp:Response / saml:Assertion. An attacker could submit a wrapped document carrying forged identity attributes alongside any valid signature made by the trusted IdP certificate, and log in as an arbitrary user.

gopacket, a widely used Go packet-processing library, miscomputes the data length of Diameter Attribute-Value Pairs (AVPs): it subtracts the fixed 12-byte AVP header size from the attacker-controlled Length field without validation, so a vendor-flagged AVP whose Length is smaller than the header causes an unsigned 32-bit underflow that drives an unbounded allocation of roughly 4 GiB. An attacker who can send crafted Diameter messages to a service using the affected decoder needs no authentication or user interaction, and two such messages in succession exhaust memory and OOM-kill the collector, yielding unauthenticated remote denial of service. Affected users are operators of Go applications built with gopacket 1.6.0 or earlier that parse Diameter traffic, such as telecom/AAA network collectors; gopacket usage that never decodes Diameter is not exposed. The flaw is fixed in version 1.6.1, and exploitation has not been observed in the wild: EPSS is 0.5% (42nd percentile), it is not in CISA KEV, and a public PoC exists via the GHSA-6r28-9ppf-4hj5 advisory.

Do: Upgrade gopacket to version 1.6.1 or later, and use govulncheck or go.mod inspection to confirm no remaining dependencies pull in versions at or below 1.6.0. For services that must parse untrusted Diameter traffic (typically TCP/SCTP port 3868), restrict and rate-limit access to those endpoints and set process/container memory limits to contain OOM impact until the patched library is deployed.

lakeFS through 1.83.0, fixed in commit 71a45ee, contains an authentication bypass vulnerability in the /setup_comm_prefs endpoint that allows unauthenticated attackers to overwrite operator metadata including email, name, and company after setup completion. Attackers can POST to this endpoint to modify security update preferences, disable security communications, and trigger falsified telemetry events using the legitimate installation ID.

A vulnerability in keras-team/keras version 3.12.0 allows an attacker to craft a malicious tar archive that bypasses the `filter_safe_tarinfos` validation in `keras/src/utils/file_utils.py`. Specifically, symlink entries are not subjected to the same `is_path_in_dir` validation as regular file entries, allowing symlinks to be created outside the intended extraction directory. This can lead to symlink-based file read, file overwrite, or directory escape attacks. The issue is particularly impactful on Python 3.10 and 3.11, where `filter_safe_tarinfos` is the sole defense against tar path traversal. This vulnerability is distinct from CVE-2025-12060 and other previously reported issues.

Improper Neutralization of Special Elements used in an SQL Command ('SQL Injection') vulnerability in Drupal Location Selector allows SQL Injection. This issue affects Location Selector versions: from 0.0.0 to 1.3.0.

CakePHP Authentication is an authentication plugin for CakePHP that can also be used in PSR-7 based applications. Prior to 2.11.1, 3.3.6, and 4.1.1, the getLoginRedirect() method contains a weakness to backslash bypasses that allows redirect targets with attacker-controlled hostnames through the redirect query string parameter. This issue is fixed in versions 2.11.1, 3.3.6, and 4.1.1.

Socket.IO enables bidirectional and low-latency communication for every platform. From 4.1.0 before 6.6.7, Engine.IO protocol v4 polling transport does not properly close the HTTP response for invalid binary POST requests with Content-Type: application/octet-stream, allowing an unauthenticated attacker to exhaust server-side connections and sockets. This issue is fixed in version 6.6.7.

A vulnerability in keras-team/keras version 3.14.0 allows for arbitrary code execution due to improper handling of deserialization in the `Lambda` layer. Specifically, the `_raise_for_lambda_deserialization()` function fails to enforce the safe-mode guard when `safe_mode` is set to `None`, which is the default value when `from_config()` is called outside of a `SafeModeScope` context. This logic error conflates `None` (unset/default-deny) with `False` (explicitly disabled), bypassing the guard and allowing attacker-controlled `marshal` bytecode to be deserialized. Affected call sites include `keras.layers.deserialize(config)`, `keras.models.clone_model(model)`, and any direct invocation of `Lambda.from_config(config)` without an enclosing `SafeModeScope(True)`. This vulnerability can be exploited to achieve arbitrary OS-level code execution in the context of the server or user process.

Improper handling of untrusted remote references in Snowflake CLI versions prior to 3.19 allowed server-side request forgery. The SQL statement reader's !source/!load directives could reference remote URLs that were retrieved at runtime without sufficient restriction on the request destination. By supplying crafted SQL content processed through a vulnerable command path, an attacker could cause the victim's environment to issue unintended outbound requests to internal or otherwise non-public network locations, and could cause remote SQL content to be retrieved and executed in the context of the victim user's session. Successful exploitation requires the victim to process attacker-controlled content through a vulnerable command path and is limited by the privileges available to that session and environment. The fix is available in Snowflake CLI version 3.19, which adds an option to disable remote URL retrieval.

A flaw was found in the libblkid library of util-linux. During nested partition probing, the BSD, Minix, Solaris x86, and UnixWare partition probers cache a raw pointer to a parent partition entry in a dynamically allocated array. When subsequent partition additions cause the array to be reallocated, this pointer becomes stale, leading to a heap use-after-free read. An attacker who can present a crafted block device image (for example, via USB insertion or a loop-mounted disk image) can trigger this flaw without user interaction, as libblkid is invoked automatically by udev/udisks as root on block-device hot-plug events. This could lead to limited information disclosure or denial of service.

Kestra is an open-source, event-driven orchestration platform. Prior to 1.0.45 and 1.3.21, the authentication filter for the REST API (@Filter("/api/v1/**")) treats any request whose path ends in /configs as the public instance-config endpoint and forwards it without a credential check. kestra addresses its resources by URL path segments that the caller chooses (/api/v1/{tenant}/flows/{namespace}, /api/v1/{tenant}/executions/{namespace}/{id}, /api/v1/{tenant}/namespaces/{namespace}/kv/{key}). An anonymous caller picks the literal configs as the final segment, and the request bypasses Basic-Auth entirely. Because the bypass reaches the flow-create and execution-trigger routes, an unauthenticated caller creates a flow containing a Shell or Process task and runs it. The task executes as root inside the kestra container. The official docker-compose.yml mounts /var/run/docker.sock, so root in the container reaches the host Docker daemon. This vulnerability is fixed in 1.0.45 and 1.3.21.

CVE-2026-49869 is a critical (CVSS 3.1: 10.0) authentication bypass in the open-source Kestra event-driven orchestration platform: the AuthenticationFilter exempts the public configuration endpoint from Basic Auth using a suffix match (request paths ending in "/configs") rather than an exact path comparison. As a result, an unauthenticated remote attacker can call any API URL whose final path segment is "configs", which is enough to create and execute arbitrary workflows without credentials. Because Kestra ships with script execution plugins (plugin-script-shell, plugin-script-python, etc.) enabled by default, this directly yields unauthenticated remote code execution as root inside the Kestra worker container. All Kestra OSS deployments prior to 1.0.45 and 1.3.21 are affected, with the greatest risk for instances whose API is reachable from the internet or untrusted networks. The flaw is being actively exploited: CISA added it to the KEV catalog on 2026-09-02 (EPSS 1.9%, percentile 79), and headlines covering the KEV batch report attackers deploying reverse shells and crypto miners.

Do: Upgrade to Kestra 1.0.45 (for the 1.0.x line) or 1.3.21 (for the 1.3.x line), or later releases that incorporate the fix. If immediate patching is not possible, restrict network access to the Kestra API (firewall/ACL or a reverse proxy that enforces authentication for all paths, including any ending in /configs). Given active exploitation, audit instances for signs of compromise such as unexplained workflows, reverse shells, or cryptomining processes in worker containers; federal agencies must patch per CISA BOD 26-04 timelines.

An out-of-bounds read vulnerability exists in dnsmasq's find_soa() function in src/rfc1035.c. When parsing NS section records, extract_name() is called with extrabytes=0, failing to validate that 10 additional bytes exist for fixed-length DNS record fields. A remote attacker controlling a DNS zone can exploit this via a crafted NXDOMAIN response to cause a 10-byte heap out-of-bounds read, potentially accessing stale data from prior transactions.

A heap-based buffer overflow was found in dnsmasq. When DNSSEC validation and query logging are both enabled, logging of DS or DNSKEY replies containing unsupported algorithm or digest types can cause dnsmasq to write past the end of an internal logging buffer. A remote attacker able to supply such a DNS response may crash the dnsmasq process, resulting in denial of service.

Rocket.Chat versions <8.5.1, 8.4.4, 8.3.6, 8.2.6, 8.1.6, 8.0.7, 7.13.9, 7.10.13 has an access control vulnerability in Livechat files. Protected file downloads at /file-upload/:fileId/:name authorize livechat access using rc_room_type=l with rc_rid+rc_token, but the authorization path does not verify that rc_rid matches the requested file's rid. Furthermore, :fileId is predictable via sequential MongoDB IDs, and :name can be anything, allowing unauthenticated discovery of all uploaded files.

Moby is an open source container framework. In Docker Engine prior to version 29.5.1, Docker Daemon versions 28.5.2 and prior, and Moby Daemon prior to version 2.0.0-beta.14, a race condition during docker cp mount setup allows a malicious container to redirect a bind mount target to an arbitrary host path, potentially overwriting host files or causing denial of service. This issue has been patched in Docker Engine version 29.5.1 and Moby Daemon version 2.0.0-beta.14.

Keras versions prior to 3.14.0 are vulnerable to a path traversal issue in the archive extraction utilities located in `keras/src/utils/file_utils.py`. The functions `filter_safe_tarinfos()` and `filter_safe_zipinfos()` validate archive member paths against the process current working directory (CWD) instead of the actual extraction destination. When the process runs with CWD set to `/`, which is common in Docker containers, CI/CD runners, and Jupyter environments, the validation boundary becomes the filesystem root, allowing traversal paths to bypass the security check. Additionally, the zip filter contains a bug that causes an `AttributeError` when a blocked entry is encountered, leading to incomplete extraction. Furthermore, Python 3.11 installations lack the `filter="data"` safety net, leaving them entirely reliant on the flawed CWD-based filter. Exploitation of this vulnerability can result in arbitrary file writes outside the intended extraction directory, enabling attackers to overwrite configuration files, inject malicious code, or corrupt machine learning datasets and pipelines.

A flaw was found in migration-planner. An authenticated user can exploit this vulnerability by sending a DELETE request to the /api/v1/sources route, which lacks proper authorization and filtering. This allows for the destruction of all customer data, including sources, agents, and assessments, leading to a critical loss of availability and integrity across the entire SaaS platform.

A vulnerability has been found in Streamlit up to 1.53.0. Impacted is an unknown function in the library lib/streamlit/runtime/caching/hashing.py of the component Palette Handler. Such manipulation leads to use of weak hash. Local access is required to approach this attack. The attack requires a high level of complexity. The exploitability is considered difficult. The exploit has been disclosed to the public and may be used. The pull request to fix this issue awaits acceptance.

The MLX inference backend in Docker Model Runner on macOS uses the MLX-LM library, which unconditionally imports and executes arbitrary Python files from model directories via the model_file configuration field in config.json. When a model's config.json specifies a model_file pointing to a Python file, MLX-LM uses importlib to load and execute it with no trust_remote_code gate or equivalent safety check. The MLX backend runs without sandboxing, resulting in arbitrary code execution on the Docker host as the Docker Desktop user. Any container on the Docker network can trigger this by calling the model-runner.docker.internal API to pull a malicious model from an attacker-controlled OCI registry and request inference.

GitBucket 4.23.1 contains an unauthenticated remote code execution vulnerability that allows attackers to execute arbitrary commands by exploiting weak secret token generation and insecure file upload functionality. Attackers can brute-force the Blowfish encryption key, upload a malicious JAR plugin via the git-lfs endpoint, and execute system commands through an exposed exploit endpoint.

Pocketbase is an open source web backend written in go. Prior to 0.22.42 and 0.37.4, in some situations, if an attacker knows the email address of the victim they can create and link an unverified PocketBase user in advance by authenticating with one of the OAuth2 app providers, e.g. "A". When the victim gets invited or decides to sign up to your app on their own with provider "B" (PocketBase OAuth2 auth requires to be with a different provider because we don't allow multiple OAuth2 accounts from the same provider to be associated to a single PocketBase user), the user created previously by the attacker will be autolinked, upgraded to "verified" and its old password reset. This vulnerability is fixed in 0.22.42 and 0.37.4.

The snorkel library thru v0.10.0 contains an insecure deserialization vulnerability (CWE-502) in the MultitaskClassifier.load() method of the MultitaskClassifier class. The method loads model weight files using torch.load() without enabling the security-restrictive weights_only=True parameter. This default behavior allows the deserialization of arbitrary Python objects via the Pickle module. A remote attacker can exploit this by providing a maliciously crafted model file, leading to arbitrary code execution on the victim's system when the file is loaded via the vulnerable method.

Pocket ID is an OIDC provider that allows users to authenticate with their passkeys to your services. Prior to 2.6.0, The createTokenFromRefreshToken function (oidc_service.go) validates the refresh token's cryptographic integrity but does not re-validate the user's current authorization state before issuing new tokens. This allows (1) the client to refresh the token indefinitely after authorization revocation, (2) the refresh token to continue to work after the account is disabled, and (3) the token to work after the client is removed from the group. This vulnerability is fixed in 2.6.0.

CVE-2026-45321 is a supply chain compromise in which 84 malicious versions across 42 @tanstack/* npm packages (including @tanstack/react-router, @tanstack/react-start, @tanstack/history, and related router/start packages) were published to the npm registry on 2026-05-11 between roughly 19:20 and 19:26 UTC, authenticated through TanStack's legitimate GitHub Actions OIDC trusted-publisher binding. The attacker chained three known weakness classes — a pull_request_target 'Pwn Request' misconfiguration, GitHub Actions cache poisoning across the fork-to-base trust boundary, and runtime memory extraction of the OIDC token from the Actions runner process — allowing publication under a trusted identity without modifying the publish workflow. Each affected package received exactly two malicious versions carrying credential-stealing malware, so developers, CI pipelines, or downstream builds that installed them could have npm, GitHub, and cloud credentials extracted; related reporting (FBI, StepSecurity) links the campaign to stolen cloud credentials and a self-spreading 'Mini Shai-Hulud' worm that also hit packages in other ecosystems such as Mistral AI and Guardrails AI. Exposure is limited to consumers who installed the two malicious versions published per package during the exposure window; other users of these widely deployed libraries were not affected by the malicious publishes. Exploitation is confirmed in the wild: the issue was added to CISA's Known Exploited Vulnerabilities catalog on 2026-05-27 with known ransomware use, and EPSS estimates a 2.3% probability of exploitation in the next 30 days (83rd percentile).

Do: Audit lockfiles and CI logs for the affected @tanstack/* packages' versions published during the 2026-05-11 ~19:20-19:26 UTC window; if found, reinstall from clean versions per the TanStack postmortem (tanstack.com/blog/npm-supply-chain-compromise-postmortem) and rotate exposed credentials (npm tokens, GitHub PATs/secrets, cloud keys), treating any cached CI artifacts as suspect. Apply mitigations per vendor instructions and CISA BOD 22-01 guidance for cloud services, or discontinue use if mitigations are unavailable, and remediate the underlying CI weaknesses (pull_request_target handling, Actions cache hygiene, and OIDC token handling) to prevent recurrence.

Kestra v1.3.3 and before is vulnerable to SQL Injection. The vulnerability occurs because user-controlled input from a GET parameter is directly concatenated into an SQL query without proper sanitization or parameterization. As a result, attackers can inject arbitrary SQL expressions into the database query.

Authentication Bypass vulnerability exists in Netmaker versions prior to 1.5.0. The VerifyHostToken function in logic/jwts.go fails to validate the JWT signature when verifying host tokens. An attacker can forge a JWT signed with any arbitrary key and use it to impersonate any host in the network, gaining access to sensitive information